the incidental acquisition of english vocabulary...
TRANSCRIPT
Academiejaar 2014-2015
THE INCIDENTAL ACQUISITION OF ENGLISH
VOCABULARY THROUGH SUBTITLING:
The Role of Vocabulary Input
Onderzoekstaak voorgelegd tot het bekomen van de graad van
Master in de Taal- en Letterkunde: Engels - Duits
Heleen Cools
Promotor: Dr. Kristof Baten
2
3
4
Acknowledgements
Writing this paper has been a difficult and time-consuming task, which would not have been
possible without the help of several people. First, I would like to thank Inneke De Regge. I
would not even have been able to analyse a single test, if it had not been for her help in the
organisation of my experiments. In addition, I am grateful to the Emmaüsinstituut Aalter for
letting me conduct my research in their school.
Special thanks go out to my supervisor dr. Kristof Baten, who agreed to supervise my
research on such short notice. I also thank Liv Persson, who helped me to define the research
questions of this paper.
5
Table of Contents
1 Introduction ....................................................................................................... 8
2 Theoretical Background .................................................................................. 12
2.1 The Importance of English .................................................................................... 12
2.1.1 English in the World ....................................................................................... 12
2.1.3 English in Flanders ......................................................................................... 13
2.2 Second Language Acquisition ............................................................................... 14
2.2.1 Definitions ...................................................................................................... 14
2.2.2 Factors influencing Second Language Acquisition ........................................ 16
2.2.3 Central Theories in the field of SLA .............................................................. 17
2.3 Vocabulary Acquisition ......................................................................................... 20
2.3.1 Stages of Vocabulary Acquisition .................................................................. 20
2.3.2 The Role of Vocabulary Input in SLA ........................................................... 22
2.4 Multimedia ............................................................................................................. 25
2.4.1 Definition ........................................................................................................ 25
2.4.2 The Influence of Multimedia on Second Language Acquisition .................... 25
2.5 Subtitling ............................................................................................................... 27
2.5.1 Definitions ...................................................................................................... 27
2.5.2 Processing Subtitled Media ............................................................................ 29
2.5.3 The Advantages of Subtitling ......................................................................... 29
2.5.4 The Influence of Subtitling on Second Language Acquisition ....................... 32
3 Methodology ..................................................................................................... 35
3.1 Educational System in Flanders ............................................................................ 35
3.2 Participants ............................................................................................................ 38
3.3 Materials ................................................................................................................ 39
6
3.3.1 Questionnaire .................................................................................................. 39
3.3.2 Film Fragment ................................................................................................ 39
3.3.3 Vocabulary Test .............................................................................................. 40
3.4 Procedure ............................................................................................................... 41
3.5 Data Processing ..................................................................................................... 43
3.6 Hypotheses ............................................................................................................. 43
4 Results and Discussion .................................................................................... 46
4.1 Different Types of Subtitling ................................................................................. 46
4.2 Questionnaire ......................................................................................................... 49
4.3 Influence of Personal Background ........................................................................ 57
4.3.1 Gender ............................................................................................................ 57
4.3.2 Exposure to English ........................................................................................ 59
4.3.3 Affective Factors ............................................................................................ 65
4.4 Vocabulary Input ................................................................................................... 68
4.4.1 General Overview ........................................................................................... 68
4.4.2 Cognacy .......................................................................................................... 70
4.4.3 Frequency in the Film Fragment .................................................................... 73
4.4.4 Word Frequency ............................................................................................. 76
4.4.5 Visual Represtentation of the Vocabulary Input ............................................ 80
5 Conclusion ....................................................................................................... 83
6 References ....................................................................................................... 86
7 Appendices ....................................................................................................... 91
7.1 Appendix 1: Questionnaire .................................................................................... 91
7.2 Appendix 2: Transcript of the 15-minute Film Fragment ..................................... 96
7.3 Appendix 3: Words in the Vocabulary Test .......................................................... 98
7.4 Appendix 4: Test Results per Participant .............................................................. 99
7
7.4.1 Group 1: E-D .................................................................................................. 99
7.4.2 Group 2: E-E ................................................................................................. 100
7.4.3 Group 3: E .................................................................................................... 101
7.4.4 Group 4: D .................................................................................................... 102
7.5 Test Results per Lexical Item .............................................................................. 103
(Introduction – Conclusion: 25 330 words)
8
1 Introduction
English is one of the most, if not the most, important foreign languages in Belgium and more
specifically in Flanders. The English proficiency of the Flemish inhabitants is generally quite
good. For example, the European Survey on Language Competences (ESLC) has shown that
Flemish youngsters are among the top three in Europe when it comes to reading, listening and
writing in English. The standard educational system in Flanders provides its students with
eight years of French instruction and five years of English instruction, but recent studies have
shown that Flemish youngsters generally obtain better results for English than for French
(Declercq et al. 2012). This would be remarkable, considering French is an official language
in Belgium, if it were not for the omnipresence of English in the daily lives of Flemish
inhabitants. English has obtained a noticeable presence and oftentimes even a dominant
position in various facets of Flemish society. One of those facets is the entertainment industry.
The rise of Hollywood as the entertainment centre of the Western world has provided the
English language with an unmatched dominance as lingua franca of films and television
programmes.
Instead of dubbing English-spoken media, small countries such as Belgium and others
like Denmark, Finland, Greece, Luxembourg, the Netherlands, Portugal and Sweden
(Koolstra, Peeters and Spinhof, 2002) prefer subtitling to make the content of those English
media easily accessible and comprehensible. So from a very young age onwards, Flemish
children are exposed to English-spoken television programmes and films. It is reasonable to
assume that this exposure to English during leisure activities could influence the acquisition
of English. Numerous scholars have investigated this possible influence and have found that
subtitling does indeed have a positive effect on the acquisition of several facets of the
English language, such as vocabulary (e.g. d’Ydewalle & Pavakanun, 1995, 1996, 1997;
Pavakanun & d’Ydewalle, 1992; d’Ydewalle & Van de Poel, 1999) and grammar (e.g. Van
Lommel, Laenen & d’Ydewalle, 2006). Focussing on incidental foreign vocabulary
acquisition through subtitling, very little research has been done one the influence of the
nature of the vocabulary input of the subtitled media. Studies on the acquisition of foreign
vocabulary through subtitling often fail to include the vocabulary list used in the described
vocabulary test (e.g. d’Ydewalle & Van de Poel 1999, Koolstra & Beentjes 1999) or fail to
motivate the choice for the lexical items that are used in the test (e.g. Kuppens 2007).
Secondly, the way this vocabulary input is presented to a participant through subtitling has
not yet been studied in detail. There are several types of subtitling (e.g. English audio with
9
English subtitles or English audio with Dutch subtitles), which may influence the acquisition
of English vocabulary differently.
Therefore, the nature of vocabulary input is a very interesting and worthwhile research
topic. Relevant questions on this topic are ‘Does the frequency of a word in the subtitled
fragment influence the acquisition of that word?’ or ‘Are cognate words acquired more
easily than non-cognate word?’ Regarding the different types of subtitling, the most
interesting question is ‘Which type of subtitling leads to the highest acquisition of English
vocabulary?’ Such questions have not yet been studied or answered in the current literature
on second language acquisition through subtitling. This paper investigates those questions
and constructs an answer to them. Since second language acquisition is subject to a large
variety of influencing factors, this paper naturally also considers possible other influences on
the acquisition of foreign vocabulary. As factors like gender, motivation, exposure to English
etc. all influence an individual’s language acquisition process, the personal background of
the participants will be taken into account. This design will allow a better investigation of the
true influence of vocabulary input in subtitling and exclude a limited or false interpretation
of the collected data.
The research questions of this paper can be subdivided into three main questions,
consisting of several sub questions.
(1) Do the students incidentally acquire English vocabulary through watching an
English spoken film fragment? Which type of subtitling results in the largest number of
acquired English words (Dutch subtitles, English subtitles or no subtitles)? The fragment
without subtitles, the fragment with English subtitles or the fragment with Dutch subtitles?
(2) Does the student’s background influence his/her incidental acquisition of English
vocabulary? Is there an influence of gender, of the frequency of exposure to English
through multimedia such as television, gaming, the Internet and music, of personal
preference regarding subtitling/dubbing, of the student’s personal motivation to learn/acquire
English, of the student’s personal opinion on the importance of English and of the estimated
personal proficiency of English, compared to fellow students?
10
(3) Does the nature of the vocabulary input provided in the film fragment influence the
acquisition of specific words? More specifically, is there a difference in the acquisition of
cognate vs. noncognate lexical items? Does the frequency of the word within the fragment
influence the acquisition of that word? Does the general frequency of a word influence the
acquisition of that word? And does it make a difference if auditory verbal input is
simultaneously accompanied with its visual representation?
To answer these questions, the research consisted of three main parts. First, the
participants were asked to fill out a questionnaire regarding their personal background.
Second, they were asked to watch a film fragment of fifteen minutes. And the third part of
the research consisted of a vocabulary test, which was taken directly after watching the film
fragment. The vocabulary test consisted of fifty lexical items. These items were presented to
the participants in the format of a listening exercise. Each word was played out loud in
English and the students were asked to write down the Dutch translation of the word. In
order to test the influence of subtitling on the incidental acquisition of English vocabulary,
the participants were divided into four groups. The first group watched an English film
fragment with Dutch subtitles; the second watched the same fragment with English subtitles.
The third group watched the fragment without subtitles and the fourth group watched the
same film fragment, but with a Dutch soundtrack and no subtitles. This last group served as
the control group, since the participants of that group were not exposed to any of the English
audio provided to the other three groups.
Although the acquisition of a second language includes grammar, morphology,
phonology and numerous cultural aspects, this paper only focuses on vocabulary acquisition.
This is done for practical purposes, since the size of this research does not allow an
investigation of any further aspects of SLA. Secondly, the focus on vocabulary acquisition
enables a detailed investigation of this one facet of SLA. According to d’Ydewalle and Van
de Poel (1999), “words, particularly nouns, are indeed the easiest building blocks in
acquiring a new language; accordingly, it is not too surprising to find the first signs of
language acquisition in the vocabulary test” (1999: 240). As this paper investigates the
acquisition of English by children who have not yet received formal instruction, the possible
influence of subtitles on SLA will probably be most noticeable when the aspect of
vocabulary acquisition is studied.
11
The structure of this paper is traditional; the theoretical background in chapter 2 presents
an extensive overview of theoretical aspects and literature on second language acquisition,
multimedia and subtitling. In chapter 3, the applied methodology (including the selection of
participants, materials, data and data analysis) will be discussed. The results and discussion
of those results will be given in chapter 5 and the conclusion of this paper can be found in
chapter 6. The conclusion attempts to give conclusive answers to the proposed research
questions and it also includes suggestions for further research.
12
2 Theoretical Background
This chapter will give an overview of the relevant theoretical aspects and literature on the
subject of this research paper. The structure of the chapter will develop from broader to more
specific topics. It starts with the importance of English, followed by a review of the relevant
concepts of the field ‘Second Language Acquisition’. That review is largely based on the
works of Gass & Selinker (2008) Krashen (1987) and Ellis (2002). Next, a specific
influencing factor on the process of SLA, multimedia, will be discussed. Paivio’s (2014)
Dual Coding Theory forms the basis of the theoretical framework on multimedia. In addition,
the influence of subtitling as a specific element of multimedia will be discussed and
elaborated on. The articles of Koolstra & Beentjes (1999), d’Ydewalle and Van de Poel
(1999), Vanderplank (1988) and Danan (2004) provide the main sources of information for
this part of the literature review. And finally, several aspects of the influence of vocabulary
input on SLA will be discussed.
2.1 The Importance of English
2.1.1 English in the World
Why study the English language? “Because it’s important. The dominant role of English as a
world language forces it upon our attention in a way that no language has ever done before.”
(Crystal 2003: 3). English has become the chief means of communication between nations in
economics, politics and culture. As portrayed by Kachru (1992), the distribution of English
around the globe can be visualized by three concentric circles, each representing different
manners of language acquisition, types of spreads and language use (see Figure 1).
Crystal (2003) explains that the inner circle
represents those countries, where English is a
primary language. The outer circle incorporates
“earlier phases of the spread of English in non-native
settings” (Crystal 2003: 107), where English has
become an important part of the country’s
institutions and acts like an essential second
language in a multicultural setting, but English is not
the mother tongue of most inhabitants.
Figure 1. The Concentric Circles of English after Kachru (1992).
13
Nations in the expanding circle acknowledge English an important international language,
without a history of colonization, like those countries of the inner circle. In the countries of
the expanding circle, including Belgium, English is taught as a foreign language.
The motivations for the acquisition of English are similar in most countries. Crystal
(2003) defined six different kinds of motivation. A first kind is historical; “because of the
legacy of British or American imperialism, the country’s main institutions may carry out
their proceedings in English.” (Crystal 2003: 106). Secondly, English may be considered a
neutral means of communication between different nations, enabling international political
conversation. A third reason is economic. Companies who wish to develop international
markets prefer the use of English, mainly because of the USA’s dominant economic position
in international trade. Next, practical reasons influence the use of English. International air
traffic control, international maritime, policing and emergency services use English as their
primary language of communication. Furthermore, English is also the chief language of
international and academic conferences and it is the leading language of tourism. The fifth
kind of motivation is based on intellectual reasons. Most scientific, technological and
academic information is formulated in English. Nowadays, English performs a similar role to
Latin in Western Europe for over a thousand years. Finally, there is also the entertainment
reason. Popular music and culture of this century mainly use English as a means of
expression.
2.1.3 English in Flanders Belgium is part of the expanding circle for English. However, the language situation of
Belgium is often considered as quite complex. The country has three official languages;
Dutch, French and German. Each of these languages is used in one of the four language
regions of Belgium. In the northern region, Flanders, Dutch is the dominant language;
French is predominantly spoken in the southern region, Wallonia. In addition, there is the
bilingual region in the middle of the country, Brussels and a German-speaking region in the
eastern part of Wallonia. As reported by Goethals (1997), Flanders is a monolingual Dutch
region. Therefore, English is considered a foreign language in Flanders.
Even though English has this official status of ‘foreign language’ in Flanders, it is not
treated as such. Flemish children receive official instruction in French in primary school, at
the age of 11. But otherwise, French is relatively absent from the daily lives of the Flemish.
English on the other hand, is the dominant language of many work and leisure related media,
such as television programmes, movies, video games, the Internet etc. According to Xu and
14
Van de Poel (2011: 274) “in Flanders English has indeed transcended its traditional role as a
foreign language and is closely tied up with the global phenomenon of English as a lingua
franca or international language.” The global popularity of English and its international
status give the language a specific attractiveness. A study conducted by Sonia Houthuys
(2011) indicates that Flemish students would be more motivated to study English than
Walloon students. Her research indicated that 88,9% of her Flemish participants, who have
not yet received formal instruction, are motivated to study English, whereas 70,8% of the
Walloon participants declared to be motivated.
2.2 Second Language Acquisition
2.2.1 Definitions As stated in Gass & Selinker second language acquisition (SLA) is a relatively young field
of study that has expanded and developed significantly in the past 40-45 years. They
describe the study of SLA as “the study of the acquisition of a non-primary language; that is,
the acquisition of a language beyond the native language” (Gass & Selinker 2008: 1).
Additionally, Ellis explains that “the term ‘second’ can refer to any language that is learned
subsequent to the mother tongue. Thus, it can refer to the learning of a third or fourth
language” (Ellis 2002: 3). Gass & Selinker add to this definition that “L2 can refer to any
language learned after learning the L1, regardless of whether it is the second, third, fourth or
fifth language” (Gass & Selinker 2008: 7). The study of SLA involves various aspects and a
variety of terminology is used to describe SLA studies. The following paragraphs will
elaborate on the most important terms and concepts within SLA.
A distinction has to be made between the concepts ‘acquisition’ and ‘learning’. In his
Acquisition-Learning Distinction Hypothesis, Krashen (1984) claims that language learners
have two distinct and independent ways of developing competence in a second language
(L2). The first way is language acquisition, which is a subconscious process. In non-
technical terms, this is often referred to as ‘picking-up’ a language. Krashen & Terrell (1983:
26) state that subjects are generally unaware of language rules or the acquired linguistic
competence. They indicate that “we acquire when we focus on what is being said, rather than
how it is said” (Krashen & Terrell 1983: 19). The concept of language learning is used to
refer to a conscious process. “Language learning as conceived by Krashen consists in the
internalization of explicit rules under conscious control. Here, there can be no invariant
acquisition order beyond what is predetermined by the process of instruction” (Klein 1986:
28). This leads to a distinction between two types of knowledge: conscious vs. subconscious
15
knowledge (the learned vs. the acquired system). Krashen & Terrell (1983: 26) state that
conscious knowledge is explicit knowledge whereas subconscious knowledge is implicit.
Within the field of SLA, there are scholars who question this distinction between
acquisition and learning. For example, Watson-Gegeo & Nielsen question the assumption
that acquisition is a natural process, which only occurs in non-instructed settings. They argue
that “although classrooms involve a distinct discourse register that may not be as rich as
other contexts in a student’s life, they are not inherently ‘unnatural’” (Watson-Gegeo &
Nielsen 2003: 157). Ellis has pointed out another critique regarding this distinction. He
underlines the difficulty of indicating whether learners dispose of acquired or learned
knowledge (Ellis 1994: 14). However, this distinction is useful in research with participants
that have not yet received formal instruction of a foreign language – English in this case.
The distinction between ‘subconscious’ acquisition and ‘conscious’ learning may not be
as clear-cut in reality as it is represented in Krashen’s work. For example, learners may
detect personal progression in learning a language, they can become aware of their
acquisition process to a certain extent. And on the other hand, learning may happen without
conscious awareness on the part of the learner (Ellis 1994: 362). Therefore, it becomes
relevant to make a further distinction between two types of language learning: incidental and
intentional learning. The difference between the two is particularly relevant for the
discussion of vocabulary acquisition (Ellis & Fotos 1999: 36).
The difference between these two concepts is explained by Hulstijn (2001: 271):
“incidental vocabulary learning refers to the learning of vocabulary as the by-product of any
activity not explicitly geared to vocabulary learning, with intentional vocabulary learning
referring to any activity aiming at committing lexical information to memory.” De Bot et al.
also point out the importance of instruction when it comes to this opposition. “When a
person reads for pleasure and doesn’t bother to look up a word he or she doesn’t know in a
dictionary, but a few pages later realizes what that word means, then incidental learning is
said to have taken place. If a teacher instructs a student to take a text and read it and find out
the meanings of unknown words, then it becomes an intentional learning activity” (de Bot et
al. 2005: 10). According to Krashen (1984), younger children would be better at incidental
language acquisition, whereas older children are better at intentional language learning, due
to their larger cognitive abilities.
16
Ellis & Fotos (1999: 36) add that not even the most scrupulous learner is able to learn all
of his vocabulary intentionally. Some lexical items will inevitably be learned incidentally.
Hulstijn (2001: 275) goes as far as to say that “the question of whether they process lexical
information without or with the intent to commit it to memory” is irrelevant.
2.2.2 Factors influencing Second Language Acquisition The study of SLA has shown that several factors influence the speed and quality of L2
acquisition. Those most discussed are attitude & motivation, age, gender, intelligence &
aptitude, anxiety & self-confidence and exposure to the L2. In the next paragraphs, the most
relevant factors for the study in this paper will be discussed and explained.
Anxiety & Self-Confidence
For a successful L2 acquisition, anxiety should be as low as possible, whereas the learner’s
self-confidence would ideally be high. These factors (affective factors cf. 2.2.3 Central
Theories in the Field of SLA – The Affective Filter Hypothesis) can influence the ease with
which a learner acquires a second language. As Krashen puts it: “There appears to be a
consistent relationship between various forms of anxiety and language proficiency in all
situations, formal and informal. Anxiety level may thus be a very potent influence on the
affective filter” (Krashen 1981: 29). Anxiety can be described as a mental block and its
occurrence is closely related to the evaluation of performance in a social and educational
setting (Horwitz et al. 1986: 125). Certain types of classroom situations may promote
language anxiety, for example if a learner could be exposed to negative evaluations.
Motivation and Attitude
A second, social-psychological factor is the motivation, combined with the attitude of the
learner. According to Skehan (1989), motivation seems to be the second strongest predictor
of success, whereas aptitude would be the first. However, it has proven difficult to produce a
consistent definition of the motivational factor. Gardner (2001) has constructed a basic
model of the role of aptitude and motivation in second language learning, which illustrates
the interconnectedness of several factors. Motivation as it is defined by Ellis “involves the
attitudes and affective states that influence the degree of effort that learners make to learn an
L2” (Ellis 2002: 75). According to Baker & Jones (1998: 645) it primarily influences the
rapidness and final proficiency of the L2, while the sequence or order of acquisition remains
rather unaffected.
17
Furthermore, Ellis (2002) identified four different kinds of motivations. First, there is
instrumental motivation. This kind inspires the learner to acquire an L2 for functional
purposes – to pass an exam, to get a certain job. The second kind is integrative motivation.
Learners may choose to learn an L2 because of an interest in the culture or the people
represented by that language. Thirdly, there is resultative motivation. It is possible that the
motivation of a learner resulted out of learning. In getting more familiar with the target
language and culture, the learner may be more inclined to learn more about that language.
The final variety of motivation is intrinsic. According to this view, “motivation involves the
arousal and maintenance of curiosity and can ebb and flow as a result of such factors as
learners’ particular interests and the extent to which they feel personally involved in learning
activities” (Ellis 2002: 76). Ellis believes that these four kinds of motivation should be seen
as complementary rather than as distinct and oppositional.
Exposure to the L2
For a successful L2 acquisition, exposure to the L2 in a naturalistic setting is of vital
importance. Kuppens (2007) conducted research on the influence of popular English media
on the vocabulary knowledge of Flemish children. Those children frequently exposed to
English television programmes and movies have a greater active knowledge of the English
lexicon than those who have not been. According to Kuppens, incidental learning through
media is beneficial because there is no excessive time or effort required to do it. Learning
occurs in a leisurely, non-instructed setting, which seems to be the key to success. According
to Kuppens, English teachers notice a considerable difference between pupils who have been
exposed to English via radio, television and music and those students who have been kept
away from these media by their parents.
2.2.3 Central Theories in the field of SLA
Krashen’s Input Hypothesis
The fundamental thought of Krashen’s Input Hypothesis is that learners acquire a first or
second language through comprehensible input. More specifically, this hypothesis focuses
on input that is slightly beyond the learner’s competence (Krashen & Terrell 1983: 32).
Baker & Jones understand the notion of ‘comprehensible input’ as follows: “Krashen
proposes that language acquisition occurs when the student is exposed to sources of
comprehensible input (oral or written) which are slightly above the level of the learner’s
current ability” (Baker & Jones 1998: 649). If the learner’s current level of proficiency is
18
represented by the letter i, the difficulty level of comprehensible input is represented as
‘i + 1’. As put by Krashen (1984: 21) “a necessary condition to move from stage ‘i’ to stage
‘i + 1’ is that the acquirer understands input that contains ‘i + 1’, where ‘understand’ means
that the acquirer is focussed on the meaning and not on the form of the message. We acquire
[…] only when we understand language that contains structure that is a ‘little beyond’ where
we are now.”
In Krashen’s theory, the comprehensibility of input depends on context, the acquirer’s
grammatical system or his/her extra-linguistic knowledge. “We use more than our linguistic
competence to help us understand. We also use context, our knowledge of the world, our
extra-linguistic information to help us understand a language directed at us” (Krashen 1984:
21). The optimal input would be “(1) sufficient in quantity, (2) given in a non-threatening
atmosphere, (3) both attended to and understood by the language learner, and (4) at an
appropriate level (just a little beyond the learner’s current linguistic competence level)”
(Ondarra 1997: 3). Once the learner has been exposed to comprehensible input, he/she will
acquire language structures in a natural way.
Krashen’s Comprehensible Input Hypothesis has been criticized by several SLA scholars.
As Sharwood Smith (1986, quoted in Ellis 1994) puts it, learning and comprehension are two
different processes and there might not be an obvious connection between the two (Ellis
1994: 27). Smith’s view does not deny the possible influence of comprehension in a learning
process, but he attributes less importance to its influence in a language acquisition process.
Additionally, Ellis underlines the importance of learner output. In the input hypothesis, the
learner would gain language competence merely trough comprehensible input (Ellis 1994:
27). However, the importance of learner output is considered a fundamental aspect of
language acquisition. Swain’s (quoted in Ellis 1994) ‘comprehensible output hypothesis’ is
based on this aspect of the language learning process. This hypothesis articulates that
learners can only learn and apply certain grammatical elements of a language through pushed
output. When learners are required to pay attention to the correct form, they will learn to
produce a foreign language correctly through trial and error (Ellis, 1994: 27). Krashen
himself does not agree with that concept. As he puts it, a foreign language can be acquired
up to a very high level without the occurrence of actual language production in the learning
process (Krashen 1998: 177). Additionally, he believes that forced output production could
possibly harm the learning process, because it could result in anxiety ( Krashen 1998: 179).
Further research by Ellis (1994: 283) shows that pushed output does not necessarily lead to
19
language acquisition, but it might lead to a better control over grammatical features that have
already been acquired.
Krashen’s Affective Filter Hypothesis
There are several factors connected to the learner, which can influence the speed and quality
of second language acquisition. Krashen’s Affective Filter Hypothesis deals with affective
factors influencing the process, such as anxiety, self-confidence and motivation. As the name
of the hypothesis suggests, these affective factors can form a filter, which prevents
successful language acquisition. Gass phrases this mechanism as follows: “If the filter is up,
input is prevented from passing through; if input is prevented from passing through, there
can be no acquisition. If, on the other hand, the filter is down or low, and if the input is
comprehensible, the input will reach the acquisition device and acquisition will take place”
(Gass 2011: 82). So learners with high self-confidence, a high motivation and little to no
anxiety will have a low or non-existent affective filter. According to Krashen “these learners
will be more open to input and acquire language more easily and deeply. […] However,
input remains the primary source of language acquisition, these variables only affect how the
acquisition of the second language develops” (Krashen 1987: 31-32). Figure 5 is a visual
representation of this hypothesis, in which L.A.D. stands for Language Acquisition Device.
Figure 5. Krashen’s Affective Filter Hypothesis (Krashen 1987: 82)
20
2.3 Vocabulary Acquisition
2.3.1 Stages of Vocabulary Acquisition A specific aspect of SLA is the acquisition of vocabulary. There are numerous reasons to
believe that the lexicon is important in second language acquisition. “In fact, the lexicon may
be the most important language component for learners” (Gass & Selinker 2008:449).
According to Levelt’s Lexical Hypothesis (1989) an item in the mental lexicon – a lexical
entry – has an entry-specific internal structure. Those lexical entries are considered to consist
of semantic, syntactic, morphological and formal (phonological and orthographic)
information. These four different types of lexical information are supposedly represented in
the two constituents that construct a lexical entry: the lemma, consisting of semantics and
syntax and the lexeme, built out of the morphology and the form (see Figure 2). When a
child is instructed in the orthography of its mother tongue (L1), it is able to extract the
semantic, syntactic and morphological information of a lexical entry simultaneously while
getting acquainted with the orthographical form of a word
Figure 2. The internal structure of the lexical entry
(adapted from Levelt 1989).
Figure 3. Jiang’s (2000) model of lexical
development.
21
Jiang (2000) applied Levelt’s Lexical Hypothesis to construct a conceptual framework within
which the findings of L2 vocabulary studies can be interpreted and discussed (Figure 3). At
the core of this developmental framework are two specific learning conditions, which
distinguish L1 vocabulary acquisition from L2 vocabulary acquisition. First, there is a
poverty of contextualized input in classroom second language learning, which renders
difficulties to extract and generate semantic, syntactic and morphological specifications
about a word and incorporate those specifications into the lexical entry of a word. Secondly,
when a L2 learner is confronted with the vocabulary of the L2, he is already in the
possession of an established conceptual and semantic system belonging to the mother tongue.
The L2 learner tends to rely on this L1 system when learning new words in the L2. Because
the semantic information of an L2 word can be acquired through the translation of its L1
equivalent, the learner may be less inclined to pay attention to the contextual indications for
meaning extraction.
Because of those two learning conditions, the vocabulary acquisition of an L2 follows a
specific path, consisting of three stages (see Figure 3). In L1 development, the main goals of
vocabulary acquisition are to understand and acquire the meaning as well as the other
properties of the word. In classroom L2 acquisition, the objective of vocabulary acquisition
is primarily to memorize the word. Whereas L1 words are learned as both semantic and
formal entities, L2 words are acquired predominantly as formal entities, because of the
presence of the L1 translation, which provides semantic information.
In the first stage of lexical development, the word association stage, L2 lexical entries
are considered lexical items without lemmas (Jiang 2000: 51). At this stage, the use and
recognition of L2 words involves the translation to the L1 equivalent, which provides the
necessary semantic and syntactic information to assist comprehension. At this stage, there
are no connections between the L2 form and the conceptual system of the L2. The learner
still relies on his L1 conceptual system. When the learner’s experience in L2 increases, the
associations between the L2 word and its L1 translation grow stronger. Continued exposure
to the L2 gives rise to significant changes in the representation and processing of the L2
word. The semantic and syntactic specifications of the L1 lemma are transferred to the empty
L2 lemma space. Consequently, the L2 form is now directly connected with the L1 syntax
and semantics, so it is directly connected with the conceptual system without necessarily
involving translation. This stage is called the L1 lemma mediation stage because the use of
L2 words is mediated by the lemmas of their L1 translations. The L2 item still does not
22
contain any morphological information, however. In the third and final stage of lexical
development, the full integration stage, all of the word’s specifications are represented in the
mental lexicon. The L1 system is no longer involved in the processing of L2 words, because
strong lexical links have developed between the L2 system and the conceptual system.
2.3.2 The Role of Vocabulary Input in SLA With the discussion of Krashen’s comprehensible input hypothesis (cf. 2.2.3 Central
Theories in the Field of SLA), the importance of input quality has already been highlighted.
There are several facets to language input e.g. frequency, form, salience, function, contextual
diversity etc. The aspects of L2 input relevant to this research will be discussed in the
following paragraphs.
Frequency
The frequency of words in L2 input is a strong predictor of SLA (Ellis & Collins 2009). The
predictive power of frequency effects relies on Zipfian distributions (Zipf, 1935). Zipf’s law
describes “how the highest frequency words account for the most linguistic tokens” (Ellis &
Collins 2009: 331). So the more frequent a word is in a specific language, the more often that
word occurs in the language input. “The frequency of any word is inversely proportional to
its rank in the frequency table. Thus the most frequent word will occur approximately twice
as often as the second most frequent word, three times as often as the third most frequent
word” (Ellis & Collins: 331). The most frequent word in English is the, which accounts for
roughly 7% of all word utterances. The second most frequent word is of, which – in
accordance with Zipf’s law, accounts for 3,5%.
According to a study by Mitz, Newport & Bever, Zipf’s law “indicates that, although
learners may be exposed to thousands of different words, the words they are regularly
exposed to will be a much smaller set of highly frequent words, which signifies a decrease in
input variability. Because learners hear and read the same sets of words consistently, these
words are learned and produced more quickly” (Crossley et al. 2013: 730).
Cognates
Kuppens (2007) investigated the influence of multimedia exposure on the acquisition of
English vocabulary. In her research, 374 participants were asked to complete a questionnaire
about their exposure to English-spoken media. All of the participants were in the sixth grade
of primary school (aged between 11 and 12) and had not yet received formal instruction in
23
English. Secondly, the participants had to take a vocabulary test. The children were shown
pictures and, if possible, they had to provide the correct English word to describe the picture.
The participants who reported a higher exposure to English-spoken media also scored higher
on the vocabulary test. But Kuppens immediately emphasizes the high number of cognates
(223 words) in the body of correctly answered questions (417 words in total). More than half
of the correct answers were cognates, which would indicate that the acquisition of cognates
is easier than the acquisition of noncognates. Kuppens describes cognates as “English words
that are related to their Dutch equivalent” (own translation Kuppens 2007: 331). But a more
specific definition of the term ‘cognate’ is desirable. Traditionally, cognates are defined as
being two words that share a source and are orthographically or phonologically similar
across two languages (Duñabeitia, Perea & Carreiras, 2010). Studies on L2 performance
have shown that cognates are translated faster and more accurately than noncognates, both in
forward and backward direction (e.g., De Groot, 1992; 1993; De Groot & Keijzer, 2000;
Ellis & Beaton, 1993; Hall, 2002; Kroll & Stewart, 1994; Sanchez-Casas, Davis, & Garcıa-
Albea, 1992; Sanchez- Casas & Garcıa-Albea, 2005). For instance, Sanchez-Casas et al.
suggest that cognates share the same lexical representation, whereas noncognates are
represented as separate entries in lexical memory. In accordance with Jiang’s (2000) model
of vocabulary acquisition, the acquisition of a noncognate would be more demanding – and
therefore more difficult – than the acquisition of a cognate. The three phases of vocabulary
acquisition – word association, lemma mediation and full integration – would be reduced to
one phase: full integration. So in theory, it would be easier to acquire cognate vocabulary.
A study by Rogers et al. (2015) investigated the acquisition of cognate vocabulary vs.
noncognate vocabulary in 30 Japanese learners of English as a second language at university.
In addition, they also studied the effect of word frequency on vocabulary acquisition. The
results of the study clearly indicated that frequent cognates are most easily acquired.
However, the research also indicated that it was more difficult for the students to use the
cognates appropriately in context, this did not prove as difficult for noncognates.
24
Visual Representation of the Vocabulary Input
Paivio’s Dual Coding Theory (explained in 2.4.2 The Influence of Mulitmedia on Second
Language Acquisition) suggests a better retention of information when visual and verbal
input are combined. In accordance with this theory, the acquisition of vocabulary input that
is accompanied by a visual image would be acquired and retained more easily. Evidence for
this assumption can be found in a study by Plass et al. (2003). They investigated whether
students’ reading comprehension and vocabulary acquisition would benefit from verbal
annotations, visual annotations or both when listening to a 762-word German story. The
participants, a group of 152 college students in the US, were asked to listen to the
multimedia version of Heinrich Böll’s “Anekdote zur Senkung der Arbeitsmoral” (created by
Chun & Plass 1997). Thirty-five of the words in this text were either accompanied by a
written translation, a visual image or both. Afterwards, the students completed a vocabulary
test about the annotated words in the text. The participants scored highest when both types of
annotations were present, followed by verbal only, then visual only. Students acquired the
smallest number of words when there was no type of annotation.
25
2.4 Multimedia
2.4.1 Definition “If one word had to be chosen to describe the nature of text in the twenty-first century, that
would be multimodality” (Ghia 2012: 1). This century has been highly influenced by the
omnipresence of multimedia, English being the dominant language of those media. An
increasing number of people gain access to various sources of information, including
television, social media and the Internet in general, radio, games etc. This increasing access
and the exposure to all forms of media has been and continues to be a very important tool for
language learning. As defined by the Concise Oxford English Dictionary (COED), the term
media indicates “the main means of mass communication (especially television, radio, and
newspapers) regarded collectively” (COED, 2008: 886). Nowadays, the Internet is also
considered a means of mass communication. When the term multimedia is used, a more
technical definition is provided. Multimedia refers to “using more than one medium of
expression or communication” (COED, 2008: 938). This definition includes subtitled
television programmes and films, since these combine different modes of communication.
Television and film provide a pictorial mode, combined with a spoken verbal mode. If
subtitles are added to a programme of film, a third mode of communication is added: the
written verbal mode.
2.4.2 The Influence of Multimedia on Second Language Acquisition An important theory regarding the influence of multimedia on second language acquisition is
the Dual Coding Theory as described by Paivio (2014). This theory claims that cognition is
made up of two subsystems: a verbal and a non-verbal system. More specifically, there is a
“verbal system for storing and processing linguistic information such as printed words and a
non-verbal system for storing and processing spatial information and mental imagery such as
static and dynamic pictures or objects” (Ahangari & Adbollahpour 2010: 3). Paivio uses the
terms logogen and imagen to refer respectively to “verbal and nonverbal representational
units that generate consciously experienced mental words and images and can function
unconsciously to mediate cognitive performance” (Paivio 2014: 146). Paivio underlines that
these two subsystems can work cooperatively as well as independently, since they are
connected to each other as well as to the sensory system. As represented by the lines in
Figure 6, pathways connect units within the same subsystem and across the different
subsystems. According to this theory, if an individual receives both verbal and visual input
regarding the same information, that information will be stored in the brain in more than one
26
way. Consequently, that information would be retained more successfully. Therefore,
according to the Dual Coding Theory, specific multimedia that combine visual with verbal
input on a specific topic, would allow a better retention of the input concerning this topic.
Figure 6. Paivio’s Dual Coding Theory (Paivio 2014)
Mayer and Anderson (1992) have applied the Dual Coding Theory in their research on
the effect of multimedia learning. In one of their tests, they investigated the possible
influence of a combination of verbal and nonverbal input, compared to verbal input only.
Their participants were college students; one group of students was shown a narrated
animation about how a bicycle pump or car brakes work. The other group only got a verbal
explanation of these mechanisms. The group of students who were shown the animation
combined with the verbal explanation scored higher on a problem-solving test than their
fellow students who had only hear the verbal explanation. This research indicated that a
deeper learning occurred after viewing and listening than after only listening.
Research by Lippens (2010), examining the influence and presence of English in the
lives of Flemish youngsters before formal instruction, has indicated that twelve year old
children are mainly exposed to English ‘through popular culture, especially through
watching English television programmes, playing English games, and listening to English
music.’ (Lippens, 2010: 143). In her study, Lippens asked 145 children, to fill out an
extensive questionnaire about their exposure to English. The results of that questionnaire
showed that “the majority of the informants spend up to one hour a day watching English
television programmes that are subtitled, and they do so every day of the week” (Lippens
27
2010: 141). Her findings also showed that the majority of the participants enjoy watching
English films and television programmes.
Kuppens (2007) has examined the correlation between exposure to English media and
the acquisition of English vocabulary. She wanted to examine the long term effects of
exposure to English media on the acquisition of L2 vocabulary. Her participants were asked
to fill out a questionnaire and to take a vocabulary test. By means of multiple choice
questions, she asked about the frequency of exposure to English media (including subtitled
and non-subtitled television and films, radio, websites, online chatrooms and videogames)
and she also asked about the participants’ attitudes towards English, their active use of
English, how well they think their English knowledge is and what their mother tongue is.
The questionnaire was followed by an oral and individual vocabulary test. The analysis of
the test results indicated that the most influential factor was the watching of subtitled
television, but only if a participant watches subtitled media several times a week. So this
study confirms the long term effects of exposure to English media and more specifically,
exposure to English subtitled television and films. Listening to English radio or playing
English videogames did not significantly influence the scores of the vocabulary test.
2.5 Subtitling In Flanders, subtitling is a common practice. The majority of English-spoken television
programmes or films broadcast in Flanders is subtitled, whereas the French-speaking region
of Belgium prefers dubbing as a translation practice (EACEA 2009). There is a quite simple
explanation for this tendency: “smaller countries import a large number of television
programs from abroad. The imported programs are generally either dubbed or subtitled in the
local language” (d’Ydewalle 2002: 59). And as Koolstra & Beentjes (1999: 52) put it,
“subtitling is preferred to dubbing in small countries because of its relatively low cost.”
Since the French-speaking population of the world is considerably larger than the Dutch-
speaking population, the costs for dubbing constitute a smaller inconvenience.
2.5.1 Definitions The term ‘subtitling’ is an overarching concept, which can refer to multiple types of
subtitling. Gottlieb (2004: 220) defines subtitling as “diasemiotic translation in polysemiotic
media [...], in the form of one or more lines of written text in sync with the original dialogue.”
The term diasemiotic refers to the two discourse channels that are converted; speech is
transformed into written text. With polysemiotic media, the multitude of information
28
channels is defined; subtitled media provide a visual, an auditory and a textual information
channel. As d’Ydewalle (2002: 59) puts it, there are “three different input channels: the
visual image, the soundtrack (including foreign voices) and the subtitles (a translation of the
voices). The text lines of the subtitles should, ideally, be completely overlapping with the
translated information of the soundtrack.”
This standard definition is applicable to all three types of subtitling: standard subtitling,
reversed subtitling and intralingual subtitling or captioning (Koolstra & Beentjes 1999: 53).
Standard subtitling is the most commonly used type and is also referred to as interlingual
subtitling. The standard type uses a foreign soundtrack with subtitles in the viewer’s native
language; this is also the most common type in Flanders. Intralingual subtitles provide the
viewer with a written translation of auditory (spoken) input. This type of subtitling is often
used by the Deaf or hearing-impaired, but it can also be used in language teaching settings
(Koolstra & Beentjes 1999: 53). A third type, reversed subtitling, provides subtitles in a
foreign language combined with a soundtrack in the native language. Reversed subtitling is
quite rare and this paper mainly focuses on the differences between intralingual, interlingual
and subtitles, therefore reversed subtitling will not be further elaborated on.
Gottlieb (1997) describes subtitles from both a linguistic as well as a technical
perspective. On a linguistic level, there are intralingual and interlingual subtitles. The first
cross over from one oral into written discourse, therefore he calls intralingual subtitles
vertical. Interlingual subtitles on the other hand cross over from oral into written discourse,
but they also cross language boundaries. Because interlingual subtitles have two dimensions
(i.e. language and discourse crossing), he calls them diagonal. On a technical level, Gottlieb
differentiates between the concepts open and closed subtitles. Open subtitles come with the
original media file and are not optional, whereas closed subtitles are to chosen by the viewer
(e.g. teletext). Since the rise of the Internet and DVDs (which have known a decline in
popularity in the past five years), the distinction between open and closed subtitles may no
longer be relevant. DVD menus offer the possibility of switching between subtitles and
English-spoken films and programmes on the Internet nearly always come with the option of
subtitles. Netflix for example, provides a legal platform for streaming films and television
series with the option of subtitling.
29
2.5.2 Processing Subtitled Media Watching subtitled programmes and films provides a viewer with three different channels of
information that need to be processed. D’Ydewalle & De Bruycker (2007: 196) state these
channels are “three different but overlapping sources of information: the visual image, the
soundtrack in a foreign language; and the subtitles in the native language”. Watching,
listening and reading occur simultaneously, as d’Ydewalle & De Bruycker put it, they
overlap.
Research has shown that viewers tend to start reading subtitles from the moment they
appear on screen. D’Ydewalle & Gielen (1992) found that the starting point of subtitle
reading is not correlated with the starting point of the soundtrack or other contextual factors.
The reading of subtitles seems like an automatic, compulsory action that is performed people
who are familiar with the practice of subtitling and those who are not. Even more, Van
Lommel, Laenen & d’Ydewalle (2006) have found that this also occurs when intralingual
subtitles are added to a soundtrack in the mother tongue. The research by d’Ydewalle &
Gielen (1992: 425) has shown that the processing of subtitled media is cognitively effective.
They described their findings as follows: “When people watch television, the distribution of
attention between different channels of information turns out to be an effortless process.
Viewers seem to have developed a strategy that allows them to process these channels
without problems and in which reading the subtitles occupies a major place’’. A more recent
study by Perego et al. (2010) also found that the quality of subtitle segmentation does not
influence the cognitive efficiency of subtitle processing. The same study also included an
eye-movement experiment, which concludes that a significant number of eye fixations was
devoted to the subtitled area. They also observed eye-movements from the subtitled area to
visual scenes and vice versa. Therefore, they believe that viewers are able to process both
sources of visual information (the visual scenes and the subtitles). Van Lommel, Laenen &
d’Ydewalle (2006) argue that processing multimedia information including subtitles can
occur successfully only from the age of ten.
2.5.3 The Advantages of Subtitling Opinions regarding the advantages or disadvantages of subtitling differ among scholar. Like
Zarei (2009), some consider subtitles to be disturbing whereas others, for example
Vanderplanck (1988) have a different view on the topic. He states: “far from being a
distraction and a source of laziness, subtitles might have a potential value in helping the
30
learning process by providing learners with the key to massive quantities of authentic and
comprehensible language input” (Vanderplanck 1988: 272-273).
Because of these different views on the possible advantages of subtitling, views which
are sometimes based on prejudice, Koolstra et al. (2002) have published an article regarding
the (dis)advantages of subtitling and dubbing. Their arguments in favour of subtitling are
based on the research of Bruls and Kerkman (1989), Luyken et al. (1991) and Spinhof &
Peeters (1999). In addition to the literature, Koolstra et al. have determined “whether or not
support can be found in empirical research” (Koolstra et al. 2002: 327). Their review of the
advantages and disadvantages has been classified into three subjects: 1) information
processing 2) aesthetics and 3) learning effects. As this paper focuses on the influence of
subtitling on the acquisition of second language vocabulary, arguments based on aesthetics
will not be discussed.
Information Processing
According to Koolstra et al. (2002), viewers can process subtitled media quite well. Subtitles
will nearly always provide a condensed version of the spoken text, due to the ‘six-second
rule’. “The longest possible subtitle of two lines containing a total of 64 characters
(including spaces, is shown on screen for six seconds” (Koolstra et al. 2002: 328). The
average speech rate is a little more than two words per second, whereas the average
presentation rate of subtitles is about two words per second. Therefore, the text presented in
subtitles has to be a condensed version of the spoken text. But condensation does not
necessarily imply information loss, because “experienced subtitlers are capable of producing
translations that are of equal value to the original information and condensation will not lead
to loss of information” (Koolstra et al. 2002: 328). Aside from condensation, the assumption
that subtitles might distract the viewer’s attention from the images on screen has been
considered a disadvantage of subtitling. This assumption has never been proven, however
and an eye-movement-registration experiment by Gielen (1988 cited in Koolstra et al. 2002)
indicates that viewers apply a specific ‘viewing strategy’, when watching subtitled media.
“The eye is focused primarily on the area just above the subtitles, a tactic that makes it easy
to monitor the most important events in the picture and read the subtitles almost
simultaneously” (Koolstra et al. 2002: 331). So it is reasonable to assume that viewers are
perfectly capable of watching the visual images on screen while reading subtitles, without a
loss of information. Gielen’s experiment proved that subtitles do not distract the viewer’s
31
attention from the screen. It is not necessarily a fact, however that watching subtitled media
is as efficient as watching dubbed media, when it comes to information processing. A study
by Koolstra et al. (1999) investigated the efficiency of information processing when
watching a subtitled programme. It became clear that efficiency increases with age (and
therefore also with reading ability). In any case, as early research by Vanderplank (1988)
already indicated, subtitles often facilitates the processing of fast speech and unfamiliar
accents: “Subjects also reported that sub-titles made fast, authentic speech and unfamiliar
accents (regional and American) much easier to tune in to and to follow. Some mentioned
that they had never been able to follow some of the programmes watched because of the
accents. All reported that comedy programmes, plots, relationships, and characterization
were much easier to understand and appreciate with sub-titles” (Vanderplank 1988: 275).
Learning Effects
Aside from entertainment purposes, subtitled media could also lead to unintentional learning.
Koolstra et al. (1997) investigated the influence of watching subtitled media on the
development of reading skills. Their three-year panel study followed more than 1000
primary school children. The results of the study confirmed the assumption that children who
often watch subtitled programmes develop better reading and decoding skills. Another
unintentional learning effect of watching subtitled media is the vocabulary acquisition of the
viewer’s own language. Viewers are exposed to verbal input, which is often accompanied
with the visual representation of those words. “While listening to spoken texts on television
‘fast mapping’ occurs: on the basis of only one exposure, partial understanding of the
mapping of a new word can lead to reorganization and expansion of vocabulary knowledge”
(Carey 1978, cited in Koolstra et al. 2002: 341). The third – and for the purpose of this paper,
the most interesting – learning effect is the acquisition of foreign languages. Aside from
second language vocabulary acquisition, which will be discussed further in this chapter
(2.5.4 The Influence of Subtitling on the Process of SLA), watching subtitled media also
results in a better pronunciation of a foreign language. According to Koolstra et al. the
question of whether listening to foreign languages spoken on television facilitates
pronunciation has not been studied yet. But they believe that “frequently watching subtitled
television programmes may be responsible for the well-known phenomenon of Dutch and
Flemish children being able to pronounce English or American words perfectly – even
‘slang’ ” (Koolstra et al. 2002: 343). Of course, all of the discussed learning effects depend
on the quality of the translation provided in the subtitles.
32
2.5.4 The Influence of Subtitling on Second Language Acquisition One of the major advantages of subtitling would be the acquisition of a foreign language.
The influence of subtitling on the acquisition of a (foreign) language is a well-studied topic
within the field of SLA. The acquisition of a foreign language through watching subtitled
media depends on several factors. A European study on the use of subtitling states that there
are different variables that influence the degree of effectiveness of SLA through subtitling:
1. “Being accustomed to subtitling (students accustomed to subtitling develop
learning strategies more quickly than those accustomed to dubbing)
2. The learner's level (depending on the learner's level, either intralinguistic or
interlinguistic subtitling will be more appropriate)
3. The objectives of the teaching (intralinguistic subtitling is better suited to
learning grammar and spelling if the learner is not a beginner, whereas
interlinguistic subtitling is more useful for building vocabulary)
4. Proximity between languages (learning through subtitling seems to be more
effective for languages whose written form is identical to that of the learner's
mother tongue).”
(EACEA 2009: 16)
In research by d’Ydewalle and Van de Poel (1999), evidence for a sensitive language-
acquisition period was sought. In their experimental set-up, they compared the influence of
subtitling on the acquisition of foreign vocabulary. Their participants, 327 children in the
third, fourth, fifth and sixth grade of primary school (resp. aged 9, 10, 11 and 12) were
shown a ten minute long still-motion movie. There were four different versions of the film:
an interlingual version (Dutch soundtrack, French subtitles) and a reversed version (French
soundtrack, Dutch subtitles) and an interlingual and reversed version with Danish audio
and/or subtitles. The participants showed a noticeable acquisition of Danish vocabulary in
both the interlingual and reversed subtitling version. But their findings did not provide
evidence for the sensitive period hypothesis; there was a clear acquisition of Danish
vocabulary in all groups. Interestingly, in the French version of the experiment there was no
or nearly no acquisition of French vocabulary. A possible explanation for this phenomenon
could be the larger proximity between Dutch and Danish, which are Germanic languages,
whereas French is a Romance language (cf. variable 4, P. 32). In reference to previous
research, d’Ydewalle and Van de Poel also point out that children benefit more from
interlingual subtitles when it comes to vocabulary acquisition, but adults acquire more
33
foreign vocabulary from reversed subtitling (d’Ydewalle & Pavakanun 1996, 1997).
According to Danan (2004), this difference could be due to the (un)familiarity with subtitling
(cf. variable 1, P. 32). If a viewer is unfamiliar with subtitles, he/she has not yet developed
“strategies to process subtitles efficiently and derive the most benefits from them” (Danan
2004: 74). Research by Vanderplank (1988) has shown similar findings when it comes to
familiarity to subtitles: “The more familiar subjects were with watching programmes sub-
titled in their mother tongue, as in the case of a Danish subject, the more rapidly they
adapted to English subtitles and developed strategies for using them to best effect”
(Vanderplank 1988: 276).
Koolstra & Beentjes (1999) investigated if standard subtitling results in higher
vocabulary acquisition compared to television with the same English soundtrack, but without
subtitling. Their participants, 246 in total, were divided into two age groups; the first group
was aged between nine and ten, the second between eleven and twelve – this group had
already received formal instruction in English. These age groups were again subdivided into
three groups, who watched the same fifteen-minute long television fragment, but under
different circumstances. One group watched the programme with standard subtitling, the
second watched without subtitles and the third (control) group got to see the fragment with
Dutch instead of English audio. After the viewing of the fragment, the participants
completed a multiple-choice vocabulary test, which asked the Dutch translation of thirty-five
English target words, used in the programme. The group that watched the fragment with
standard subtitling scored significantly higher on the vocabulary test than the other two
groups, but the group who watched the non-subtitled fragment still acquired more
vocabulary than the control group. The oldest age group scored significantly higher in all
situations than the younger group. This finding is quite different from the research by
d’Ydewalle and Van de Poel (1999), in whose research the same age differences did not
prove to be a differentiating factor for vocabulary acquisition. So the difference between the
age groups in Koolstra and Beentjes’ research is most likely due to reception of formal
instruction by the oldest group (the oldest group in d’Ydewalle and Van de Poel’s research
had not received instruction in Danish). So one could assume that formal instruction is a
more important differentiator than age.
Van Lommel and Laenen & d’Ydewalle (2006) have done research on the incidental
acquisition of foreign grammar through subtitling. They used reversed subtitling to find out
if the grammatical rules of Esperanto would be acquired through watching a movie. The 107
34
participants were asked to take a grammar test after watching the movie. Contrary to the
research on the incidental acquisition of vocabulary, the results of this study did not indicate
any significant grammar acquisition through subtitling.
35
3 Methodology The focus of this research is the possible influence of subtitling on the acquisition of English
vocabulary. An element that is specifically focussed on, is the potential role of vocabulary
input. In order to examine the three research questions described in the introduction, as well
as their several subquestions, a threefold design proves most useful. The participants,
materials and procedure of this setup will be discussed in the following paragraphs.
The methodology of this research is quite similar to most studies on subtitling, although
one element is deliberately different. In the literature on subtitling, most scholars opt for a
written vocabulary test. In such tests, a list of English words are written on an answer sheet
and the participants are asked to provide the correct translation of those words in their
mother tongue. This paper focuses on the possible effects of Dutch, English and no subtitles,
so the participant group exposed to English subtitles are advantaged when they are given a
written vocabulary test. Therefore, an auditory vocabulary test has been chosen in this
research.
3.1 Educational System in Flanders In this research, all participants follow the same level and type of education. This paragraph
will present a short overview of the Flemish educational system, in order to situate the type
of education followed by the participants. The information used in this paragraph originates
from De Ro (2008). The three communities in Belgium, the Flemish, Walloon and German-
speaking community, provide their own educational system. In Belgium, education is
compulsory between the ages of six and eighteen. However, most children start nursery
school at the age of 2.5. After three years of nursery school, primary education starts around
the age of six. Within mainstream education, there are no different levels of primary
education. When primary school is finished at the age of twelve, the child receives a
certificate and the child’s teacher usually gives the child and its parents a non-binding advice
on which type of secondary education the child should best follow.
Mainstream secondary education consists of three stages, which each cover two years
(see Figure 7). The first stage has been divided into streams; A-stream and B-stream. A-
stream education provides children with broad general knowledge and it is followed by
children who did not experience any major difficulties during primary school. B-stream, also
referred to as pre-vocational training, is intended for those children who did experience
difficulties during primary school. This type of education provides children with education
36
that is adapted to their specific needs. After the first two years of secondary education,
Flemish children have to choose a more specific type of education; General Secondary
Education (ASO), Secondary Arts Education (KSO), Technical Secondary Education (TSO)
or Vocational Secondary Education (BSO). Students who have followed A-stream education
in the first stage of secondary education can choose any of these types, but those who have
followed pre-vocational training can only choose vocational training. ASO continues to
provide children with a very broad general knowledge and it does not focus on any specific
occupation. This type of education aims to prepare students for higher education. As De Ro
states “ASO provides a very firm foundation for passing on to tertiary education and that is
why most pupils choose to continue studying after ASO” (De Ro 2008: 32). About 40% of
Flemish pupils choose to follow ASO. Secondary Arts Education (KSO) combines broad
general knowledge with active arts practice. Students who follow this type of education can
choose to follow tertiary education after graduation or they can choose a profession. The
number of students in this type of education is very small; just below 2% of all Flemish
pupils choose KSO. Technical Secondary Education provides a mixture of practical and
theoretical education. There is a strong emphasis on technical knowledge and TSO students
can either pass on to tertiary education or choose a profession after graduation. Around 30%
of Flemish pupils are TSO students. Students who prefer practical education and do not
intend to follow higher education usually follow vocational training (BSO). In BSO, students
are trained in a specific profession. Around 27% of Flemish students can be found is
vocational training. BSO students only receive a diploma of secondary education after
following a seventh year of vocational training. Any student with a diploma of secondary
education is allowed to follow higher education. Which type of education was followed, does
not make a difference. But as indicated, ASO is specifically prepares students for higher
education whereas BSO trains students for a specific profession. Next to mainstream
education, there is of course also special education. But this paper focuses on ASO students
in mainstream education, which makes a discussion of special education less relevant.
37
Figure 7: The Structure of the Flemish Educational System (De Ro 2008: 21)
38
3.2 Participants For this research, it was important that the participants were old enough to be exposed to
English-spoken media for quite some time, but had not yet received formal instruction in
English. This second condition was important, because the results of this research would not
be interpretable if some of the words in the vocabulary test had already been taught to the
children in a classroom setting. In Belgium, the instruction of English in usually starts in the
second year of secondary education, when the students are between 13 and 14 years old.
Therefore, the participants of this research were in the first year of secondary education –
aged between 12 and 13 – so the time the participants could have been exposed to English-
spoken media was as long as possible. The group of participants would ideally be as
homogeneous as possible, when it comes to age, intelligence and mother tongue. The
participants came from the same school; Emmaüsinstituut Aalter. Aalter is a town between
Bruges and Ghent, with almost 20.000 inhabitants, but only 441 inhabitants are not Belgian1.
So chances are, the vast majority of participants would speak Dutch as a mother tongue,
which was a prerequisite for this study, as the possible influence of Dutch subtitles is one of
the studied subjects of this paper. Secondly, the students all received the same type and
subtype of secondary education: A-stream – Algemene Vorming (General Education).
The total number of participants was 78, but the results from two participants were left
out. One girl indicated that English is her native language and one boy was only 10 years old,
so he has not been exposed to English media for as long as the other participants. The rest of
the participants, 76 in total, were all aged between 12 and 13, their year of birth was 2002.
From these participants, 41 were female and 35 were male. All participants together cover
four classgroups, each group contained precisely 19 effective participants. These four groups
cover the four situations needed for this research; three groups watched an English-spoken
film fragment with Dutch, English or no subtitles and the fourth group watched the same
fragment with a Dutch soundtrack and no subtitles (the control group). As mentioned, all
effective participants speak Dutch as their native language and two participants indicated
that they had been raised bilingually, one in Dutch and Albanian, the other in Dutch and
French. None of the participants indicated to have a learning disability (e.g. dyslexia) and
1 Numbers retrieved from the official website http://www.aalter.be/file_uploads/5229.pdf
39
none of them followed an adjusted educational programme2. Furthermore, none of the
participants has received formal instruction in English outside of school and none of the
individuals had followed a language camp for English during holidays. So it is safe to state
that this group of participants is very homogeneous and therefore, the individual results of
the participants can be compared to one another.
3.3 Materials
3.3.1 Questionnaire The design of this research is threefold. First of all, the students were asked to fill out a
questionnaire, in order to obtain the relevant information regarding their personal
background (see Appendix 1). This questionnaire contained 13 questions in total. A first
category of questions asked for personal information, including age, gender, mother tongue,
other languages known and possible history of English instruction. If there had been major
differences between the participants when it comes to these factors, the interpretation of the
test results could not be conclusive. The second category asked about exposure to English
media and subtitled media. The participants were asked how often they watch English
programmes and films. They had to indicate whether they do this with Dutch subtitles. They
were also asked which type of programmes and films they prefer and which type they watch
most often (Dutch, English with Dutch subtitles, English with English subtitles or English
without subtitles). Furthermore, the students were asked how often they are exposed to
English through the Internet, games and music and if there are any other ways in which they
are exposed to English (e.g. friends or family, travelling, family etc.). The last category
investigated the student’s individual attitude towards English, including questions about
motivation, importance of English and the student’s self-confidence when it comes to his/her
own knowledge of English. The information obtained by means of this questionnaire is used
to determine the potential influence of any of these factors on the results of the vocabulary
test.
3.3.2 Film Fragment Secondly, all of the participants were shown the same film fragment. The first group (E-D)
watched it with English audio and Dutch subtitles, which is the type of subtitling most
2 The M-decree in Flemish Education assures that children with specific educational needs (such as
children with a learning disability, physical or mental disability) are able to receive standard education, if reasonable adjustments to the learning environment and educational programme of the child can be made. http://www.ond.vlaanderen.be/specifieke-onderwijsbehoeften/beleid/M-decreet/
40
common in Belgium. Group 2 (E-E) watched the fragment in English with English subtitles
(interlingual subtitling), the third group (E) watched the fragment without subtitling and the
control group (D) watched it in Dutch without subtitling. The film fragment was part of the
Disney Nature film Chimpanzee. This particular film was chosen for several reasons. First of
all, the film targets an audience between 10 and 14 years old. Therefore, the content and
language are appropriate for participants between 12 and 13 years old. Secondly, this target
audience assures comprehensible input (cf. 2.2.6 Central Theories in the Field of SLA:
Krashen’s Comprehensible Input Hypothesis). Third, this is a nature documentary film,
which facilitates the research of a specific aspect of vocabulary input: a visual representation.
Nature documentaries do not include dialogue or speech by characters. A narrator talks the
audience through the visual input of the documentary, often describing scenes or specific
phenomena, animals or actions. The film is narrated by Tim Allen, an American actor and
narrator. Carice Van Houten is the narrator of the Dutch version of the film. The students
were shown the first fifteen minutes of the film, in which 772 words are spoken. A transcript
of the used fragment can be found in Appendix 2. The film shows the story of Oscar, a baby
chimpanzee. His actions, habitat, family etc. are described during the documentary in
language that is easy to comprehend. The story was accessible to children and the speed of
the film was high enough not to become boring, but slow enough that it was easy to follow.
3.3.3 Vocabulary Test
The third part of the research consisted of a vocabulary test, which included 50 lexical items
in total (see Appendix 3). All of these words were part of the 15-minute film fragment and
they were selected for a number of reasons. First of all, three categories were chosen; nouns,
verbs and adjectives. The test included 16 nouns (N), 16 verbs (V) and 18 adjectives (A).
Secondly, a further division was made. Half of the words of the test were cognates, the other
half were noncognates. Next to this first subdivision, a second one was made; half of the
words were frequently used in the film fragment, the other half was only used once. If a
word is labelled as frequently used, this indicates that the specific word has been repeated
three times or more within the 15-minute fragment.
All of the participants, with exception of the control group, watched the film fragment
with the same English audio, but different subtitling situations. Therefore, the modality of
the vocabulary test had to be auditory. This way, the effect of the subtitling on vocabulary
acquisition could best be determined. A written presentation of the English words would
favour the group that had viewed the fragment with English subtitling. This group got to
41
watch the entire fragment, while the written version of every spoken word was shown in the
subtitles. Hence, a traditional translation exercise (where e.g. ‘dog’ is printed on the test
sheet and ‘hond’ has to be written down as the answer), would put the other groups at a
disadvantage. Therefore, the 50 lexical items were read out loud and the participants had to
write down the Dutch translation of the word they had just heard. For each of the words, an
audio file was created. Each file was integrated a PowerPoint slide, which also mentioned
the question number and the word category (N,V or A). As the film fragment was narrated
with an American accent, it was necessary that the audio files also played the words in an
American accent. Therefore, the audio files were downloaded from the Google Translate
database.3 The voice in these files is male and speaks with an American accent, which is as
close to the narrator’s voice as possible.
3.4 Procedure The procedure has been the same in each of the four different class groups. Because of
practical reasons, the researcher was allowed a period of 50 minutes with each of the groups.
Each part of the test, resp. questionnaire, film fragment and vocabulary test, took some 15
minutes. When entering the classroom, the students were told that I am a student at Ghent
University. I informed them that I would write a master-paper about children between 12 and
13, but no further information about the purpose of my visit was given. This way, the
participants would not be influenced and they would not be inclined to pay special attention
to certain aspects of the test.
The first phase of the research was the questionnaire. Each student was given a copy and
was asked to fill out his/her first name and class group. No last names were asked, in order to
take away some of the anxiety some students may have experienced. The participants were
reassured that none of their names would be used in my paper and the answers on their
questionnaires would be processed anonymously. The students were asked to tick off the
box next to the answer that best suited their personal situation. They were also informed that
only one answer could be given, unless indicated otherwise.
After the collection the completed questionnaires, the students got to watch the 15-
minute film fragment. For this second phase of the research, the participants were asked to
be quiet during the film and to enjoy it. In all classrooms, the film was projected on a white
3 This database is accesible through the URL https://ssl.gstatic.com/dictionary/static/sounds/de/0/boy.mp3
This URL links to a downloadable MP3 file (in this case, for the word ‘boy’). To access the MP3 file for any type of word, the word at the end of the URL, before the .mp3 (here ‘boy’) has to be changed.
42
projection screen, while the room was darkened. Each of the classrooms had an audio system
installed. There were four speakers in the room, one in each corner. The first frame of the
film was the illuminated Disney castle, to which the students reacted very enthusiastically.
Each of the four groups appeared very eager to watch the film. All groups were able to watch
the film fragment quietly and attentively. Interestingly, the E-E and E groups seemed less
amused with the content of the film than the E-D and D group. This phenomenon could be
explained by the fact that these groups had no input in their native language and therefore,
comprehension was more difficult. But even during scenes without any spoken language (e.g.
in one scene the chimpanzees were moving quite rhythmically, this scene was accompanied
with rhythmic music and therefore it seemed like the chimpanzees were dancing to the
music) there was less laughter and enjoyment in the E-E and E groups.
For the third part of the research, the vocabulary test, each student was handed an answer
sheet with 50 empty answer slots. The design of the test was explained thoroughly before the
test started. The students were told that they would have to give the Dutch translation of 50
English words. If they did not know the correct translation, they could leave the answer slot
open or give a description of the word in Dutch. They were also told that the test would
include nouns, verbs and adjectives. In order to ascertain that the students understood the
difference between these categories (N,V or A), I asked the group to give me a definition of
these concepts. The understanding of this difference between nouns, verbs and adjectives
was important, because some lexical items used in the test could be interpreted multiple
ways (e.g. a battle vs. to battle, live vs. to live). During the test, I told the students when one
category had been finished and when the next began. After the explanation of the test, the
participants were also asked to remain quiet during the test as not to give away answers to
their classmates. As mentioned, this was an auditory vocabulary test. Each of the audio files
was integrated in a PowerPoint slide, which also showed the participants the question
number and word category. Each file was played twice and if necessary – when a student
indicated that he/she had not entirely understood or heard the word – it was played once
more. After the files were played, the students got 15 seconds to fill in the correct answer.
When the students had finished the test, each file was played again, with a quicker
succession, so any unanswered questions could be answered if possible. This entire process
took between 15 and 20 minutes.
After the test, the students got the opportunity to ask question, express concerns or
discuss the test with their peers. When the 50 minutes had past, the students were asked not
43
to discuss the test or its purpose with pupils from other groups, as not to give those groups an
advantage at the test. All of the groups indicated during and after the test, that they found the
test quite difficult. During the test, some students – most often girls – shifted in their seats or
sighed deeply, when they did not know the answer to a question immediately. And after the
test, a number of students came up to me and told me that their results would probably be
bad, because they are not competent in English.
3.5 Data Processing The answers to the questionnaire and the results of the vocabulary test were processed with
Microsoft Excel 2011. The statistical calculations were done with StatPlus, a statistical
software program, designed to be work with data in Excel spread sheets. In order to visualise
the results of this research, tables and graphs, designed in Excel were made. The correction
of the vocabulary test was done manually, before the results of each question could be
entered in Excel. The answer to a question was either right or wrong, no half points were
given. No marks were deducted for spelling mistakes in the Dutch translations.4 If a student
was not able to give the exact translation of the lexical item in the test, but he/she provided a
correct description of the item, the answer was considered correct. One example of such a
description was ‘het tegengestelde van vriend’ (the opposite of friend), when the word
‘Enemy’ was given. The anonymous individual results of each participant can be found in
Appendix 4.
3.6 Hypotheses After the overview of the relevant literature and the methodology, it is possible to formulate
some hypotheses for my research questions and the sub-questions. The first question is
concerned with the possible influence of subtitling on the acquisition of foreign language
vocabulary. Based on the literature on this topic, it is very likely that there will indeed be
vocabulary acquisition by the participants of this research. The research by d’Ydewalle and
Van de Poel (1999) and Koolstra and Beentjes (1999) indicates that participants acquire
foreign vocabulary through both interlingual and intralingual subtitling, but intralingual
subtitling results in more vocabulary acquisition than interlingual subtitling. Therefore, the
hypothesis for the first research question is, that all groups (E-D, E-E, E) will acquire
vocabulary after watching the subtitled film fragment and the first group (E-D) will acquire 4 Quite remarkably, a very large number of participants (nearly one third) had difficulties with the spelling
of certain Dutch words. Especially words with ‘ij’-‘ei’ or ‘d’-‘t’ were spelled incorrectly. In one case, the student’s confusion about the spelling of ‘leider’ (Eng. Leader) was very remarkable. He wrote five different variations of the word, before he decided on an (incorrect) spelling.
44
more vocabulary than the second (E-E), which will again acquire more vocabulary than the
third group (E).
The second research question involves the influence of a participant’s personal
background on his/her results on the vocabulary test. The literature on SLA suggests that
several personal factors may influence the quality and quantity of vocabulary acquisition.
The participants of this study are all born in the same year, so age will not be a deciding
factor. All participants follow the same level of secondary education, which excludes
intelligence as a major influence. Although a certain level of education does not guarantee an
equal intelligence among participants, it is safe to assume that there will be no large
differences in intelligence. Another factor, self-confidence and anxiety may have influenced
the student’s results. Two of the questions in the questionnaire (Q. 12 & 13) focus on this
factor, so a possible influence will be investigated. The literature suggests that students with
a low affective filter (i.e. high self-confidence and low anxiety) should score higher than
those with a high affective filter, if Krashen’s Affective Filter Hypothesis is followed. A
factor closely related to self-confidence and anxiety is motivation. Once again, a high
motivation would result in higher test results. And of course, exposure to the L2 is
considered to be a very influential factor, especially with participants that have not received
any formal instruction. The larger the exposure to the L2, the better a participant should
score. So the hypothesis for the second research question assumes that a low affective filter,
a high motivation and a large exposure to the L2 will result in higher test scores. The actual
results will have to determine which of these factors influence the acquisition of L2
vocabulary the most. Whether L2 exposure through certain media would be more influential
than exposure through other media, has not yet been studied elaborately. But Kuppens’
(2007) research indicates that listening to English music has an effect on the acquisition of
noncognates.
The third research question is concerned with the influence of specific input qualities. First
of all, a lexical item’s absolute frequency (i.e. how often the word occurs within the English
language) should influence the acquisition of the word. Items with a high absolute frequency
will be acquired more easily than less frequent words. Secondly, the influence of a word’s
relative frequency (i.e. how often the word occurs within the film fragment) has not been
studied or tested yet. The hypothesis on this influencing factor is: the higher the relative
frequency, the better the test score. Third, the difference in acquisition of cognates and
noncognates will be discussed. Kuppens (2007) research shows that cognates are acquired
45
easily, whereas noncognates are barely acquired after watching subtitled television. Finally,
the influence of visual representations of a lexical item will be investigated. If Paivio’s
(2014) Dual Coding Theory is applied, one can assume that visual representation of words,
combined with an auditory and written version of the word, should result in a higher
acquisition of that word. In light of the relevant literature, the hypothesis on this final
research question suggests that cognate, highly frequent – both absolute and relative – and
visually represented words will be acquired by most students.
46
4 Results and Discussion 4.1 Different Types of Subtitling The first research question of this paper investigates the potential influence of subtitling on
second language acquisition. According to the hypothesis formulated for the first research
question, the E-D group should score highest, followed by the E-E group and the E group
should score lowest of the three. In any case, the control group (D) should score lower than
any of the other groups; otherwise there would be no influence of the English audio or
subtitling whatsoever. A basic overview of the average test score per group indicates that the
first hypothesis could be confirmed (see Graph 1).
Graph 1. The Average Test Score per Group in Percentage.
Group 1 has an average test score of 72,84%, which is 11,7% higher than the control group.
This average corresponds with an average of 36,4 out of 50 questions answered correctly.
The best scoring participant in this group made only one mistake, he scored 49/50 (98%). No
other participant in any other group managed to score this high. The lowest score in group 1
was 22/50 (44%) (see Graph 2). With a minimum of 22 and a maximum of 49, the standard
deviation (s.d.) in group 1 is 8,58. Group 2 scores slightly lower than group one, with an
average score of 69,9%. The second group’s score is 8,8% higher than the control group. The
maximum score was 46/50 (92%) and the minimum score was 18/50 (36%) with a s.d. of
9,61. The third group, which watched the film fragment without subtitles scored 67,5%
averagely, which is still 6,4% higher than group 4. The maximum score in group 3 was 44/50
(88%), the minimum score 16/50 (32%) and the s.d. 7,46. Finally, group 4 has an average
72,84 69,89 67,47 61,05
0
10
20
30
40
50
60
70
80
90
100
Group 1: E-‐D Group 2: E-‐E Group 3: E Group 4: D
47
score of 61,10. The maximum score was 45/50 (90%), the minimum score was 13/50 (26%)
and the s.d. 8.78. Graph 2. gives a clear overview of these results.
Graph 2: Overview of the Vocabulary Test Results in Percentage
The visual representation in this graph clearly indicates that the expectations expressed in the
hypothesis for the first research question are met. The highest, average and lowest test scores
gradually decline from group 1 to group 4. The absolute highest score was obtained by a
participant in group 1, the absolute lowest score by a participant in group 4. In order to
determine the statistical significance of these data, a one-way ANOVA (Analysis of
Variance) test was carried out between groups 1, 2 and 3 (see Table 1). Group 4 was left out
of this test, since the first three groups only differ when it comes to subtitling input, whereas
group 4 also differs from the others when it comes to auditory input.
Analysis of Variance (One-‐Way) Groups Sample size Sum Mean Variance
Group 1: E-‐D 19 13,84 0,72842 0,02948 Group 2: E-‐E 19 13,28 0,69895 0,03695 Group 3: E 19 12,82 0,67474 0,02226
ANOVA Source of Variation SS df MS F p-‐level F crit
Between Groups 0,02747 2 0,01373 0,46451 0,63093 3,16825 Within Groups 1,59651 54 0,02956
Total 1,62397 56
Table 1: One-way ANOVA of General Score vs. Subtitling Type.
72,84 69,89 67,47 61,05
44,00 36,00
32,00 26,00
98,00 92,00
88,00 90,00
0,00
10,00
20,00
30,00
40,00
50,00
60,00
70,00
80,00
90,00
100,00
Group 1: E-‐D Group 2: E-‐E Group 3: E Group 4: D
Average Score
Lowest Score
Highest Score
48
This one-way ANOVA shows that, although there may be a difference in mean (average)
between the three groups, there is no significant statistical difference: p > .05. Based on
these results, there are no real statistical differences between the three different types of
subtitling (Dutch, English and no subtitling).
There may not be any statistically significant difference between the three types of
subtitling, there could still be a significant difference between the three groups who watched
the film fragment with English audio and the control group who watched it with Dutch audio.
A one-way ANOVA of English versus Dutch audio proves that this difference is in fact
significant: p < .05 (see Table 2).
Analysis of Variance (One-‐Way) Summary
Groups Sample size Sum Mean Variance Group 1-‐3: English Audio 57 3.994, 70,07018 289,99499
Group 4: Dutch Audio 19 1.158, 60,94737 311,7193 ANOVA
Source of Variation SS df MS F p-‐level F crit Between Groups 1.185,96491 1 1.185,96491 4,01642 0,04872 3,97023 Within Groups 21.850,66667 74 295,27928
Total 23.036,63158 75
Table 2: One-way ANOVA of English vs. Dutch Audio
In any case, these data represent the influence of subtitling on the acquisition of all fifty
lexical items of the vocabulary test. The control group had an average score of 61,10%,
which is quite high. The large number of cognate lexical items in the vocabulary test (25 out
of 50) may account for this high score. It is assumable that different types of subtitling may
still prove a statistically significant variable when it comes to the acquisition of second
language vocabulary. Next to the difference between cognates and noncognates, there are
several other factors involved. For any of these other factors, the subtitling type may prove to
be a statistically significant differentiator. Therefore, the discussion of the following two
research questions will still make a distinction between the different types of subtitling. First
of all, because the potential influence of subtitling in different circumstances could still be
determined. Secondly because the results of the four different groups cannot only be
discussed as one sample, since each of the groups made the test after watching a different
type of subtitled media.
49
Answering the first research question, it became clear from the test results that there is
L2 vocabulary acquisition in all three subtitling situations (Dutch, English and no subtitles).
However, the statistical significance between the difference types of subtitling is not high
enough to make definitive conclusions so far.. The results did not prove to be statistically
significant, although the average scores of each test group indicate an affirmation of the
hypothesis formulated about this research questions.
4.2 Questionnaire In this paragraph, the answers to the questionnaire will be discussed. Since the questionnaire
was completed before the film fragment was watched, every participant completed it under
the same circumstances. Therefore, all answers will first be discussed as a whole, without a
subdivision based on the type of fragment a group watched. This discussion will give general
insights on 12-year olds’ exposure to English, preference of subtitling vs. dubbing, self-
confidence etc.
The first set of questions asked about some of the student’s personal information. The
answers to these questions were mainly used to ensure that the group of participants was as
homogeneous as possible. As mentioned in paragraph 3.2 (Participants), there were 41
female and 35 male participants, who were all born in the same year. Dutch was the mother
tongue of all participants and none of them had followed any English language courses or
camps. The second set of questions (Q1-10) gives insight in the participants’ exposure to
English media and their personal preferences regarding subtitling or dubbing.
The first question simply asked if the participants often watch television and films.
Regular exposure to media in general and English media is a prerequisite for this study.
Fortunately, the answers to this first question confirmed this assumption. 6 out of the 76
participants indicated that they do not watch televisions or films on a regular basis. 92,11%
of all participants regularly watch multimedia like film and television. This answer is
completely in line with the expectations. The second question asked the students how often
they watch English-spoken television programmes and films. This question includes both
subtitled and regular English media. The students could indicate their answer on a five-point
scale, ranging from ‘Never’ to ‘Every Day’.
50
Graph 3: Answers to Question 1 and Question 2.
The answers to question two indicate that English films and programmes are quite
popular among 12-year olds. 50% watches them several times a week and 17,11% even
watches them every day. Almost 20% of the participants watch English films and
programmes a few times a month. Roughly 14% watched these types of media less
frequently, 11,43% only watches them a few times a year and 2,63% never watches English
films and television programmes at all (see Graph 3).
The third question investigates whether or not the participants are familiar with
interlingual subtitling and if they use it often. To the question ‘If you watch English
programmes and films, do you watch them with Dutch subtitles?’ there was no single
participant who answered ‘Never’. Most students always watch with Dutch subtitles
(60,52%) or they regularly do so (39,47%). So it is safe to say that all participants are
accustomed to watching subtitled media (see Graph 4).
Graph 4: Answers to question 3 in percentage.
92,11%
7,89% 2,63%
11,43% 19,74%
50,00%
17,11%
0% 10% 20% 30% 40% 50% 60% 70% 80% 90% 100%
Q.1 Do you oKen watch Television and
Films?
Yes No Q.2 How oKen do you watch
English Programmes and Films?
Never A Few Times a Year
A Few Times a Month
A Few Times a Week
Every Day
0%
39,47%
60,53%
0%
10%
20%
30%
40%
50%
60%
70%
Never Regularly Always
51
The fourth and fifth question are concerned with the participants’ most watched and
most preferred type of subtitling. The four different possible answers (see Graph 5) to this
question coincide with the four different subtitling situations of this paper (i.e. English-
Dutch, English-English, English without subtitles or dubbing). The largest number of
participants watches English media with Dutch subtitles, 50% of them do so. The preference
for Dutch subtitles is even higher; 56,58% prefer Dutch subtitles over the other three
possibilities. Next to interlingual subtitling, dubbing also seems a popular mode of
translation. It is most watched by 43,42% of all participants and a slightly lower number
(38,16%) prefers dubbed programmes and films over media with English soundtracks. These
two options are most popular and their popularity is probably due to the fact that English
media without subtitles or with English subtitles are hardly available on mainstream
television channels. A minority of 5,26% mostly watches English-spoken media without
subtitles and an even smaller group (1,32%) watches English-spoken films and television
programmes with intralingual subtitles. 2,63% of the participants (which coincides with only
2 participants) prefer English-spoken media without subtitles and an equal number prefer
intralingual subtitling. Once again, these results do not prove to be surprising. Mainstream
television channels most often provide subtitled English media, which accounts for the
largest number of answers in this category. And since none of the participants have yet
received formal instruction, it would prove quite difficult for most of them to watch these
types of media without any form of Dutch input, either in subtitling or in dubbing.
Graph 5: Answers to questions 5 and 6 in percentage.
43,42% 50,00%
5,26% 1,32%
38,16%
56,58%
2,63% 2,63% 0%
10%
20%
30%
40%
50%
60%
70%
80%
90%
100%
Dubbed Dutch SubYtles No SubYtles English SubYtles
Most Watched
Preference
52
Question six asked the participants which music they listened to most often. This
question was included because Kuppens (2007) research indicated that exposure to English
music had a significant effect on the acquisition of noncognate English words (p<.01)
(Kuppens 2007: 332). The students could choose between three answers: Dutch music,
French music or English music. Looking at the current music charts, English music seems
most popular in Flanders, so it is expected that most participants will listen to English music
most often.
Graph 6: Answers to Question 6 in Percentage
As expected, a very large majority of the 12 and 13-year old participants listens to
English music. Only two participants (2,63%) listens to French music most often and one
participant (1,32%) listens to Dutch music (see Graph 6). This participant also indicated that
she does not watch television often, she never watches English television or films and she
prefers dubbed media over other types (like subtitled media). This very limited exposure to
English media will very likely influence the results of her vocabulary test. As only three
participants do not listen to English music (or rarely do so), it will become impossible to
compare the test results of this study to those of Kuppens (2007).
96,05
2,63 1,32
0 10 20 30 40 50 60 70 80 90 100
English Music French Music Dutch Music
53
Question seven deals with a type of exposure to English that is becoming increasingly
important. The Internet has become a major source of information and entertainment for
younger generations. The participants were asked how often they encounter English on the
Internet. Again, they had to indicate an answer on a five-point scale, ranging from ‘Never’ to
‘Every Day’. In addition to this question, the participants were also asked on which websites
they encounter English. This was the only question where multiple answers could be given.
Some very popular types websites were already provided on the answer sheet, but there was
also the possibility to add other types of websites.
Graph 7: Answers to Question 7 and 8.
As expected, more than half of the participants encounter English on the Internet on a
weekly (31,58%) or daily (27,63%) basis. These numbers indicate that, at the age of 12 or 13,
children encounter English more often on the Internet than through television or films. About
a fifth (21,05%) of all participants gets in contact with the English language through the
Internet a few times a month, around 15% only a few times a year and almost 4% of all
participants, which corresponds with three participants, never do so (see Graph 7). The
answers to this question, combined with the answers to question 2 confirm the statement that
Flemish children are very frequently exposed to English through various media. When asked
which websites with English content they visit most often, a very large group of participants,
3,95% 15,79% 21,05%
31,58% 27,63%
53,95%
89,47%
28,95% 28,95%
3,95% 5,26%
0% 10% 20% 30% 40% 50% 60% 70% 80% 90%
100%
Q7. How
oKe
n do
you
en
coun
ter E
nglish on
Never
A Few Tim
es a Year
A Few Tim
es a M
onth
A Few Tim
es a W
eek
Every Da
y
Q8. W
ebsites
Facebo
ok
YouTub
e
Gaming Web
site
Blog
Web
shop
54
almost 90% indicated they visit YouTube5 on a regular basis. About half of the participants
said that they encounter English via Facebook. However, the amount of English input via
Facebook is probably quite low, as Facebook offers its content in the language of the
individual user. An equal amount of participants gets in contact with English through gaming
websites and the social medium Instagram (both 28,95%). Blogs and webshops provide a
source of English input for resp. 3,95% and 5,26% of participants (see Graph 7).
The ninth question of exposure to English asks about the participants’ exposure to
English through videogames. Again, participants could answer on a five-point scale how
often they play English videogames, ranging from ‘Never’ to ‘Every Day’. Only a small
amount of videogames has a Dutch version, so in Flanders, most games are played in English.
Contrary to popular belief, playing videogames is (no longer) a form of entertainment for
boys only. The answers to this question indicate that girls also spend a reasonable amount of
time playing English videogames. More than half of the girls (22 out of 41) play videogames
at least a few times a month or more often. The vast majority of boys (30 out of 35)
frequently plays English videogames (ranging from a few times a month to every day) (see
Graph 8).
Graph 8: Answers to Question 9.
5 During YouTube’s 10-year existence, the website has become an increasingly popular platform for
entertainment, especially among teenagers. ‘YouTubers’ are people who post videos on YouTube on a regular basis and have made YouTube their main source of income. The vast majority of YouTubers speak English as a native language and very recently, YouTube have installed a subtitling function on the website, which makes English content more easily accessible for viewers from all over the world. Very specifically, YouTubers who make videos about the videogame ‘Minecraft’ have a huge amount of subscribers and views per video.
10,53%
21,05%
26,32%
32,89%
9,21%
0%
5%
10%
15%
20%
25%
30%
35%
40%
45%
50%
Never A Few Times a Year A Few Times a Month
A Few Times a Week
Every Day
55
The final three questions of the questionnaire explore the participants’ motivations for
learning English and their self-confidence about their own English knowledge, compared to
their peers. Question eleven asks if the participant thinks it is important to learn English.
Whether the answer was positive or negative, it had to be motivated by a short sentence.
There was only one participant who did not believe it is important to learn English. He stated
“Dutch is my mother tongue and it is difficult enough as it is. I do not need to learn yet
another language.” All other participants indicated that they did believe the learning of
English is important. They gave quite similar motivations. About 90% of the participants
think English it is important to learn English, because it is a language used all over the world
and they can use English when travelling and meeting foreign people. The other 10% believe
English is important when watching English-spoken television and films or when playing
English videogames.
Question twelve asked the participants how well they think their English knowledge is,
compared to their peers. The five-point scale of answers ranged between ‘Very Bad’ and
‘Very Good’. As the answers to this question prove, Flemish children appear quite confident
about their personal knowledge of English, compared to their peers. More than 40% believes
his/her knowledge is good. About one third thinks that he/she has an average knowledge of
English, whereas 17,11% thinks his/her knowledge is bad and just below 10% believes it is
very bad. Only one participant is very confident about his knowledge (see Graph 9).
Graph 9: Answers to Question 12.
9,21% 17,11%
31,58%
40,79%
1,32% 0%
10%
20%
30%
40%
50%
60%
70%
80%
90%
100%
Very Bad Bad Average Good Very Good
56
The last question asked the participants whether or not they think English is a difficult
language. The answers to this question divide the group in two. Just above half of the group
– 39 participants – think English is in fact a difficult language, just below half of the group –
37 participants – think it is not (see Graph 10). The participants that answered this question
positively gave motivations like ‘English has a difficult pronunciation, it is hard to write
English words correctly and the grammar is different from Dutch grammar’. None of the
participants mentioned vocabulary as a specifically difficult aspect of learning English.
Those participants that answered the question negatively motivated their answers as follows:
‘English has a lot of words that are similar to Dutch words. I already know some English
from television and gaming.’ One participant motivated his answer with the following
expression: ‘English is and cannot be a difficult language. If it were difficult, there would not
be as many people who are capable of speaking it. But since everyone is able to speak
English, it must be an easy language’.
Graph 10: Answers to Question 13.
51,3% 48,7%
0%
10%
20%
30%
40%
50%
60%
70%
80%
90%
100%
Yes No
57
4.3 Influence of Personal Background The literature on SLA and subtitling suggests that there are a number of factors that possibly
influence second language acquisition. The participants’ answers to the questionnaire will
give insights into the influence of each individual’s background on the acquisition of the
English vocabulary presented to them in the film fragment. Not all questions included in the
questionnaire will be included in this section. For example, the exposure to a specific type of
music will not be discussed, as the answers to this question indicated that only three
participants do not listen to English music most often. Certain other questions (Q. 1, 3, 6, 8)
will also not be included in this discussion. The main purpose of these questions was to give
insight into the habits and attitudes of the participants as a group of 12 to 13-year olds.
4.3.1 Gender
First, the difference in test scores between boys and girls will be discussed. The group of
participants consisted of 41 girls and 35 boys. If the results between the boys’ and the girls’
test scores are compared without making a distinction between the four different groups (E-
D, E-E, E, D), there is a large difference between boys and girls. The 35 boys score an
average of 73,31%, whereas the 41 girls have an average score of 63,12%. This is a
difference of almost 10%. An ANOVA indicates that this difference is statistically
significant: p < .05 (see Table 3).
Analysis of Variance (One-‐Way) Groups Sample size Sum Mean Variance
Boys 35 25,66 0,73314 0,02807 Girls 41 25,88 0,63122 0,02862
ANOVA Source of Variation SS df MS F p-‐level F crit
Between Groups 0,19615 1 0,19615 6,91457 0,01039 3,97023 Within Groups 2,09919 74 0,02837
Total 2,29534 75
Table 3: One Way ANOVA of Gender vs. Test Score.
When the boys’ results are compared to the girls’ across the different subtitling
situations, the results are similar. In every group, the boys score higher than the girls. In the
first three groups, the difference is even larger than 10% (see Graph 11). This graph also
shows the influence of the different subtitling situations on the average test scores of the
boys and girls. Again, the visual representation of these results confirms the difference
between the different subtitling situations. The graph shows a gradual decline from Group E-
D to Group D. This decline is visible in the average group scores, which has already been
discussed, but it is also present in the average group scores of the boys and the girls.
58
Graph 11: Overview of the boys’ and girls’ average test scores compared to the group averages (in %).
The study by Kuppens (2007) also showed a difference in average scores between boys
and girls. She attributes this difference to the amount of videogames played by boys. The
answers to question 9 showed that boys do play videogames more often than girls, but girls
still play them frequently. On the other hand, the questionnaire did not ask which types of
games the participants played. So it is possible that the difference between boys and girls in
this study originates from the input (i.e. the words in the film fragment). The fifty lexical
items of the test were all included in the fragment from Chimpanzee. At a certain point in the
fragment, a battle between two rival groups is shown and described. 20% of the words in the
vocabulary test (10 out of 50) come from this specific scene. These 10 lexical items are
battle, victory, enemy, territory, rival, leader, threat, escape, catch and safe. These are all
words that are often used in certain videogames and in some genres of television, which are
specifically marketed towards boys. This could explain the higher scores among the boys.
However, this is only a possible explanation, not a conclusive one. Since the questionnaire
did not include questions about which games participants play and which genres of television
and films they watch, there are no data to support this assumption.
78,00 76,44
74,29
64,22
72,84 69,89
67,47
61,05
67,11 64,00 63,50
58,20
50,00
55,00
60,00
65,00
70,00
75,00
80,00
85,00
90,00
95,00
100,00
Group E-‐D Group E-‐E Group E Group D
Boys
Average
Girls
59
4.3.2 Exposure to English
Television and Films
According to the literature (e.g. Kuppens 2007), higher exposure to English television and
films leads to a higher knowledge of English vocabulary. The data of this study confirm this
notion. The average score of participants who watch English television or films daily is more
than 35% higher than the average score of participants who never do so. Participants who
rarely watch English media (never or a few times a year) score below 50% on average.
Those who watch English television and films several times a month score above 60% on
average and those who watch English media on a regular basis (more than once a week)
score above 70% (see Graph 12). A one-way ANOVA (see Table 4) shows that these results
are highly significant: p = .00001. These numbers underline the huge impact of exposure to
the L2 on the acquisition of L2 vocabulary. The results confirm the hypothesis about the
second research question, which stated that a high exposure to the L2 would lead to a higher
score on the vocabulary test.
Graph 12: Average Test Score in Correlation with Exposure to English Television and Films
Analysis of Variance (One-‐Way) Groups Sample size Sum Mean Variance
Never 2 0,82 0,41 0,0002 A Few Times a Year 8 3,78 0,4725 0,01954 A Few Times a Month 15 9,18 0,612 0,02067 A Few Times a Week 38 27,68 0,72842 0,02267 Every Day 13 10,08 0,77538 0,02194
ANOVA Source of Variation SS df MS F p-‐level F crit
Between Groups 0,76672 4 0,19168 8,90304 0,00001 2,50076 Within Groups 1,52862 71 0,02153
Total 2,29534 75
Table 4: One-Way ANOVA of Test Score vs. Exposure to English Television and Films.
41,00% 47,71%
61,20% 72,84% 77,54%
0% 10% 20% 30% 40% 50% 60% 70% 80% 90% 100%
Never A Few Times a Year A Few Times a Month
A Few Times a Week
Every Day
60
Subtitling, no Subtitles or Dubbing
Next, it becomes possible to find out which type of media exposure leads to the highest
vocabulary acquisition. In question 5, the participants were asked which type of media (i.e.
dubbed, with Dutch subtitles, without subtitles or with English subtitles) they watched most
often. It would be expected that those participants exposed to dubbed media most often
would have a disadvantage over those who are exposed to English-spoken media more often.
The average test scores in correlation with the type of media most watched by the
participants confirms this assumption. Participants most exposed to dubbed media have an
average test score of 58,69%, whereas those exposed to subtitled media averagely score resp.
72,37% and 74% for Dutch subtitling and English subtitling. The two participants who watch
English media without subtitles most often score 96% on average (see Graph 13). Watching
English media without subtitles at the age of 12 indicates that these participants are very
familiar with English and feel quite comfortable about their English knowledge. Their
answers to other questions in the questionnaire support this assumption. Both participants
think their personal knowledge of English, compared to their peers is good and they are
exposed to English-spoken media on a daily basis. So once again, the higher the input of
English audio, the larger the knowledge of English vocabulary. A one-way ANOVA
confirms the statistical significance of the correlation between test score an exposure to a
specific type of media: p<.001 (see Table 5). The results from participants who watch
without subtitles or with English subtitles cannot be considered truly representative, as only
four participants – two for each group – belong to these groups.
61
Graph 13: Average Test Score in Correlation with Exposure to a Certain Type of Media.
Analysis of Variance (One-‐Way) Groups Sample size Sum Mean Variance
Dubbed 29 17,02 0,5869 0,03104 Dutch Subtitles 43 31,12 0,72372 0,02117 No Subtitles 2 1,92 0,96 0,0008 English Subtitles 2 1,48 0,74 0,0392
ANOVA Source of Variation SS df MS F p-‐level F crit
Between Groups 0,49732 3 0,16577 6,63817 0,00051 2,73181 Within Groups 1,79803 72 0,02497
Total 2,29534 75
Table 5: One-Way ANOVA of Test Score vs. Exposure to Media Type.
58,69%
72,37%
96,00%
74,00%
0%
10%
20%
30%
40%
50%
60%
70%
80%
90%
100%
Dubbed Dutch SubYtles No SubYtles English SubYtles
62
The Internet
The exposure of the participants to English on the Internet is the next possible influence on
the participants’ test scores. The answers to question 7 show that more than half of the group
encounters English on the Internet more than once a week. This makes the Internet a
valuable source of input. Again, the more exposure to English – here through the Internet –
the better the participants score on the vocabulary test. However, when the numbers from
Graph 14 are compared to those in Graph 12, it becomes clear that exposure to English
through the Internet is not as influential as the exposure through English television and films.
Between those participants that never encounter English through the Internet and those that
encounter it every day, there is a difference of around 15% in average scores. When the
influence of English films and television was discussed, this difference was larger than 35%.
As the one-way ANOVA shows, the influence of the exposure to English through the
Internet is not statistically significant: p>.05 (see Table 6).
Graph 14: The Correlation between the Average Test Scores and the Exposure to English through the Internet.
Analysis of Variance (One-‐Way) Groups Sample size Sum Mean Variance
Never 3 1,68 0,56 0,09 A Few Times a Year 12 6,96 0,58 0,02713 A Few Times a Month 16 10,74 0,67125 0,02523 A Few Times a Week 24 17,14 0,71417 0,0309 Every Day 21 15,02 0,71524 0,02548
ANOVA Source of Variation SS df MS F p-‐level F crit
Between Groups 0,21826 4 0,05456 1,86517 0,12606 2,50076 Within Groups 2,07708 71 0,02925
Total 2,29534 75
Table 6: One-way ANOVA of Average Test Score vs. Exposure to English through the Internet
56,00% 58,00%
67,13% 71,42%
71,52%
50%
60%
70%
80%
90%
100%
Never A Few Times a Year A Few Times a Month
A Few Times a Week
Every Day
63
Videogames
A fourth type of exposure to English happens while playing videogames. The answers to the
questionnaire revealed that more than 40% of all participants play videogames more than
once a week. Participants who never play videogames score an average of 53,14%. Those
who play a few times a year or a few times a month score almost exactly the same average,
resp. 63,25% and 63,10%. So there is a 10% difference between the participants who never
play games and those who sporadically do so. Participants who play games a few times a
week and participants who play every day score resp. 74,92% and 78,29% on average. This
is a difference of more than 10% with those who play only every now and then for those who
play weekly and a difference of 15% for those who play daily (see Graph 15). A one-way
ANOVA indicates that these results are statistically significant: p < .05 (see Table 7). These
data indicate that an increase in exposure to English through videogames results in a higher
score on the vocabulary test.
Graph 15: The Correlation between Average Test Scores and Exposure to English through Videogames.
Analysis of Variance (One-‐Way) Groups Sample size Sum Mean Variance
Never 8 4,6 0,575 0,04454 A Few Times a Year 16 10,12 0,6325 0,02079 A Few Times a Month 20 12,62 0,631 0,03375 A Few Times a Week 25 18,73 0,7492 0,02428 Every Day 7 5,48 0,78286 0,01312
ANOVA Source of Variation SS df MS F p-‐level F crit
Between Groups 0,36587 4 0,09147 3,37115 0,0139 2,50076 Within Groups 1,92641 71 0,02713
Total 2,29228 75
Table 7: One-Way ANOVA of Average Test Score vs. Exposure to English through Videogames.
53,14%
63,25% 63,10%
74,92% 78,29%
50% 55% 60% 65% 70% 75% 80% 85% 90% 95% 100%
Never A Few Times a Year A Few Times a Month
A Few Times a Week
Every Day
64
When it comes to the influence of exposure to English through various types of media,
the data from this study agree with the literature; the higher the exposure to English, the
higher the average test score. However, not every type of media is equally influential. First
of all, the exposure to English through English-spoken films and television appears to be
highly influential. The more frequently a participant watches this type of media, the higher
his/her test score was. Secondly, the way in which an English-spoken programme or film is
made accessible to an audience is also influential. As expected, those who watch dubbed
versions of a film or programme score lower than those who do not, since these participants
are not exposed to any form of English input. Between the different types of subtitling
(interlingual, intralingual or no subtitling), there is also a difference. Exposure to either
English or Dutch subtitling resulted in almost equal scores, whereas watching without any
type of subtitles led to an average score of 96%. Exposure to English through the Internet did
not prove to be an influencing factor. There are differences in average scores between
participants who are frequently exposed to English on the Internet and those who are not.
But these differences did not prove to be statistically significant. Exposure to English
videogames on the other hand does prove influential. So in short: exposure to English
through subtitled (interlingual or intralingual) or non-subtitled English television and
programmes and English videogames influence the average scores of the vocabulary test.
The first part of the hypothesis on research question two has been confirmed. A high
exposure to English through various media (television, films and videogames) result in a
higher test score. The second part of this hypothesis will be answered in the following
paragraph, which deals with affective factors.
65
4.3.3 Affective Factors
Self-confidence about Personal Proficiency
The last two questions in the questionnaire dealt with affective factors. The hypothesis
regarding affective factors assumes that a low affective filter results in a higher score on the
vocabulary test. The twelfth question dealt with the students’ self-confidence regarding their
knowledge of English. The answers to this question indicated that a large part of the group –
over 40% - thinks his/her English is good to very good. The data indicate that these
participants were able to estimate their own level of knowledge quite aptly. The one
participant that believes his English is very good, scored 90% one the test. This is more than
double the score of the average of those who believe their English is very bad (42,86%).
Remarkably, this one participant who is particularly confident about his English also
achieved the highest score on the vocabulary test, which is 98%. The participants who
believe their English is bad to very bad averagely score below the group average, which is
67,82%. Those who believe their knowledge is average, score very close to the group
average (70,08% vs. 67,82%) and those who believe their English is good to very good score
above this average. A one-way ANOVA shows that the statistical significance of this factor
is very high: p ≈ 0 (see Table 8).
Graph 16: The Correlation between Self-Confidence and Vocabulary Test Score.
42,86%
55,23%
70,08% 76,26%
90,00%
0%
10%
20%
30%
40%
50%
60%
70%
80%
90%
100%
Very Bad Bad Average Good Very Good
66
Analysis of Variance (One-‐Way) Groups Sample size Sum Mean Variance
Very Bad 7 3, 0,42857 0,01078 Bad 13 7,18 0,55231 0,01617 Average 24 16,82 0,70083 0,01908 Good 31 23,64 0,76258 0,02245 Very Good 1 0,9 0,9
ANOVA Source of Variation SS df MS F p-‐level F crit
Between Groups 0,92445 4 0,23111 11,96954 1,73049679697E-‐07 2,50076 Within Groups 1,37089 71 0,01931
Total 2,29534 75 Table 8: One-Way ANOVA of Average Test Score vs. Self-Confidence about English.
This correlation between self-confidence and test score is not only applicable to the
group as a whole. When the test scores and the answers to question 11 are compared within
the different subtitling groups, this trend is also visible. Again, those who estimate their
knowledge to be bad or very bad score below average, those who think it is good or very
good score above average. Within the different groups, there is a slight difference between
the group average and the score of those participants who believe their knowledge is average.
There is also a noticeable incline in test scores, which correlates with an incline in self-
confidence.
Graph 17: The Correlation between Self-confidence and Average Test Score per Subtitling Group.
0%
10%
20%
30%
40%
50%
60%
70%
80%
90%
100%
Group E-‐D Group E-‐E Group E Group D
Very Bad
Bad
Group Average
Average
Good
Very Good
67
Difficulty of English
A second affective factor can be found in the answers to question 13. The answers to this
question indicated that roughly half of the group thinks English is a difficult language, the
other half does not think so. Presumably, there will be a difference in test score between the
group who thinks English is difficult and the other half of the group. However, the test
results indicate that there is only a minor difference between the two groups. The group that
believes English is a difficult language has an average score that is only 1,1% lower than the
group that thinks English is not difficult. Contrary to expectations, this factor does not
influence the results of the vocabulary test.
Graph 18: The Correlation between the Answers to Question 12 and the Average Test Scores.
These data lead to the conclusion that self-confidence about one’s personal proficiency
is an influential affective filter in this study, whereas the belief whether or not English is a
difficult language does not influence the results of the vocabulary test. The second part of the
hypothesis on research question 2 stated that a low affective filter results in a higher score on
the vocabulary test. This hypothesis applies to one affective factor; self-confidence.
67,28% 68,38%
0%
10%
20%
30%
40%
50%
60%
70%
80%
90%
100%
Difficult Not Difficult
68
4.4 Vocabulary Input The previous paragraphs discussed average scores, based on groups of participants. This
paragraph deals with average scores, based on groups of lexical items. It will discuss how
certain characteristics of the words in the vocabulary test may or may not influence the
acquisition of those words. Each of the characteristics will first be discussed for the group as
a whole (i.e. no distinction will be made between the different types of subtitling). Once a
general tendency has been established for the entire group, the results from group 1, 2 and 3
will be compared with those from the control group in order to determine the possible
influence of each type of subtitling. The results from the control group will be used as a
reference point, so the effects of subtitling on the acquisition of these words can be
determined, depending on the word characteristic discussed.
4.4.1 General Overview The vocabulary test consisted of 50 lexical items, which were all part of the 15-minute film
fragment. The selection of the words depended on several parameters. First of all, there had
to be a fairly equal number of nouns, verbs and adjectives. Second, half of the words had to
be cognate, the other half noncognate. A third parameter was frequency, half of the words
were used three times or more during the film fragment, the other half was only used once.
After the selection of the vocabulary items, based on these parameters, each word belongs to
three different categories (Word Class, Cognacy, Frequency). When these three categories
are combined, there are twelve different subcategories the words can belong to. Table 9
gives an overview of the words used in the vocabulary test, in the categories they belong to.
Out of the 50 words, there were six words known by all participants. These words were: 1.
Day - 3. World - 4. Time - 22. Give - 34. Special. These words are all cognates and with the
exception of 22. Give, they are all used frequently in the fragment. This could already
indicate that cognate, frequent words are most easily acquired. The following paragraphs in
this chapter will give examine these tendencies more thoroughly. On the other side of the
spectrum, there are five words that were translated correctly by less than 20% of the entire
group: 10. Independence – 12. Threat – 20. Seek – 47. Bright – 50. Precious. All items, with
the exception of 20. Seek are noncognates. Three of them are frequent (10. Independence –
12. Threat – 20. Seek), two are non-frequent (47. Bright – 50. Precious). An overview of the
average score of each item (for the entire group and for the separate groups) can be found in
Appendix 5.
69
Noun Verb Adjective
Cognate Noncognate Cognate Noncognate Cognate Noncognate
Frequent 1. Day
2. Life
3. World
4. Time
5. Mother
9. Forest
10. Independence
11.Boy
12. Threat
17. Make
18. Live
19. Rain
20. Seek
25. Like
26. Know
27. Keep
28. Close
33. Rich
34. Special
35. Full
36. Far
41. Little
42. Safe
43. Real
44. Possible
Non-‐
Frequent
6. Territory
7. Rival
8. Leader
13. Enemy
14. Danger
15. Battle
16. Victory
21. Grow
22. Give
23. Find
24. Eat
29. Travel
30. Survive
31. Escape
32. Catch
37. Wild
38. Good
39. Deep
40. Light
45. Difficult
46. Small
47. Bright
48. Happy
49. Amazing
50. Precious
Table 9: Overview of the Lexical Items per Category.
70
4.4.2 Cognacy The vocabulary test, which included 50 words from the film fragment, consisted of 25
cognate words and 25 noncognates. The literature on vocabulary input indicates that
cognates are easier to acquire than noncognates. A comparison of the average scores of
noncognates with the averages of cognates reveals a significant difference between the two
categories. The average scores of the cognates is 80,05%, whereas the average of the
noncognates is 55,58%. So there is a difference of around 25% between cognates and
noncognates. An ANOVA indicates that these values are statistically significant: p < .01 (see
Table 10). These are averages for the entire group of participants, without making a
distinction between subtitling situations.
Graph 19: The Difference between Cognates and Noncognates in Percentage.
Analysis of Variance (One-‐Way) Groups Sample size Sum Mean Variance
Cognates 25 2.001,31724 80,05269 485,29901 Noncognates 25 1.389,47592 55,57904 809,2838
ANOVA Source of Variation SS df MS F p-‐level F crit
Between Groups 7.486,99591 1 7.486,99591 11,56665 0,00136 4,04265 Within Groups 31.069,98744 48 647,29141
Total 38.556,98336 49
Table 10: One Way ANOVA of Cognacy vs. General Score.
80,05
55,58
0
10
20
30
40
50
60
70
80
90
100
Cognates Noncognates
71
The effect of subtitling on the acquisition of cognates and noncognates has not yet been
discussed in the literature. As it turns out, subtitling influences the acquisition of both
cognates and noncognates. The E-D group has an average score of 85,95% for the cognates,
which is 11,21% higher than the control group. The E-E group and the E group scored
almost the same average for the cognates; resp. 81,05% and 81,47%, which is resp. 6,31%
and 6,73% higher than the control group (see Graph 20). An ANOVA between group E-D,
Group E-E and group E shows that there is no statistically significant difference between
these groups (p > .05) when it comes to the acquisition of cognates. And an ANOVA
between the first three groups (E-D, E-E and E) and the control group (D) also indicates that
there is no significant difference between the three test groups and the control group in the
acquisition of cognates (p>.05). Finally, ANOVAs between each subtitling group vs. the
control group (E-D vs. D, E-E vs. D and E vs. D) did not show any statistical significance (p
> .05).
For the noncognates, the E-D group scored an average of 62,74%, which is 15,37%
higher than the control group. With an average of 58,74, the E-E group scored 11,37%
higher than the control group. And the E group scored 6,1% higher than the control group
with the average score of 53,47%. But again, a one-way ANOVA between the three different
test groups shows no statistical significance (p > .05) and a one-way ANOVA of the three
test groups vs. the control group does not show any statistical significance either (p > .05).
And ANOVAs between each subtitling group vs. the control group did not show statistical
significance (p > .05).
Table 11: The Difference between the Average Scores of Cognates and Noncognates
Group E-‐D Group E-‐E Group E Group D
Cognates 85,95% 81,05% 81,47% 74,74%
Noncognates 62,74% 58,74% 53,47% 47,37%
Difference 23,21% 22,31% 28% 27,37%
72
The difference between the average score for cognates and noncognates across the
groups is 23,21% for the E-D group, 22,31% for the E-E group, 28% for the E group and
27,37% for the control group. So these differences are about the same for the two groups
who watched the fragment with subtitles and the two groups who watched it a form of
subtitling (see Table 11).
Graph 20: Averages Scores of Cognates and Noncognates across the Different Groups in Percentage.
In summary, there is a large difference in the acquisition of cognates and noncognates.
As expected, the acquisition of cognates is a lot easier than the acquisition of noncognates.
This difference between cognates and noncognates is present in each test group (E-D, E-E,
E) as well as the control group (D). In each group, a lot more cognate words were known
than noncognates. But if we want to find out which subtitling situation is most favourable for
acquiring cognates, there seems to be no statistical difference between the three subtitling
situations. The same situation applies for the acquisition of noncognates. This situation is
similar to the findings of 4.1 (Different Subtitling Situations). The numbers show that there
are differences between the test groups and the control group, as well as differences within
the different test groups. The visualization of these numbers (Graph 20) shows a decline in
average scores (both for cognates and noncognates) between group E-D and group D. But
after statistical analysis, these numbers do no prove to be statistically significant.
85,95 81,05 81,47
74,74
62,74 58,74
53,47 47,37
0
10
20
30
40
50
60
70
80
90
100
Group E-‐D Group E-‐E Group E Group D
Cognates
Noncognates
73
4.4.3 Frequency in the Film Fragment Half of the words used in the vocabulary test are frequent, the other half is non-frequent (see
Table 9). The hypothesis formulated for research question 3 states that frequent words
should be translated correctly more often than non-frequent items. A comparison of the
average scores for frequent and non-frequent items reveals that frequent items get a higher
score than non-frequent items. Words that are uttered more than three times during the film
fragment get an average score of 70,68%, those who are only used once score an average of
64,95%. This is a difference of 5,73%, which is nowhere near as big as the difference
between the cognates and noncognates. A one-way ANOVA also indicates that, although
there is a difference between frequent words and non-frequent ones, this difference is not
statistically significant (p > .05).
Graph 21: The Difference between Frequent Words and Non-frequent Words in Percentage.
Analysis of Variance (One-‐Way) Summary
Groups Sample size Sum Mean Variance Frequent 25 1.767,1075 70,6843 821,14638
Non-‐Frequent 25 1.623,68566 64,94743 768,25307 ANOVA Source of Variation SS df MS F p-‐level F crit
Between Groups 411,3965 1 411,3965 0,51768 0,47532 4,04265 Within Groups 38.145,58686 48 794,69973
Total 38.556,98336 49
Table 12: One-way ANOVA of Average Score vs. Frequency.
70,68 64,95
0
10
20
30
40
50
60
70
80
90
100
Frequent Non-‐Frequent
74
For the group as a whole, there is no significant difference in the acquisition of frequent
and non-frequent items. There are differences between the different groups (see Graph 22).
For the frequent words, the E-D group scores an average of 74,74%, 10,1% higher than the
control group. With an average score of 72,63, the E-E group scores 7,79% higher than the
control group and the E group scores 5,69% higher, with an average of 70,53%. So once
again, there is a gradual decline across the groups, the E-D group scores highest, followed by
the E-E group, then the E group and finally the D group. A one-way ANOVA shows no
statistical significance between the four groups in the acquisition of frequent items (p > .05).
An ANOVA between the three test groups proves that there is no statistically significant
difference in the acquisition of frequent items across these groups (p > .05). And ANOVAs
between each individual test group and the control group give similar results when it comes
to statistical significance (p > .05).
For non-frequent items, the E-D group scores 70,95%, 13,69% higher than the control
group, the E-E group scores an average of 67,16%, 9,9% higher and the E group scores
64,42% 7,16% higher than the control group. These results follow the same tendency as the
average results of the frequent items; there is again the gradual decline starting with group E-
D, followed by group E-E, then group E and finally the control group. And again, the three
different ANOVA tests (i.e. between the four different groups, between the test groups and
the control group and between each individual test group and the control group) show no
statistically significant difference between the different subtitling situations.
Group E-‐D Group E-‐E Group E Group D
Frequent 74,74% 72,63% 70,53% 64,84%
Non-‐Frequent 70,95% 67,16% 64,42% 57,26%
Difference 3,79% 5,47% 6,11% 7,58%
Table 13: The Difference in Average Scores between Frequent and Non-frequent Items.
75
The difference in average scores between frequent and non-frequent items in group E-D
is 3,79%, it is 5,47% in group E-E, 6,11% in group E and 7,58% in group D. So the
difference between frequent and non-frequent items increases (which is also visible on
Graph 22; the distance between the red and the blue line grows). As graph 22 indicates, the
non-frequent items are the source of this increase, as the decline of the red line (non-frequent
items) is steeper than the decline of the blue one (frequent items). A possible explanation for
this phenomenon can be found in the Dual Coding Theory. Each non-frequent item is only
used once in the fragment. So if this item is made available to the participant through several
channels (i.e. visual, auditory and written), the retention of this item will be more successful.
Since the input in group E-D provided the participants with English auditory input and Dutch
written input, the chance a word is retained – even if it is only used once – should be larger
than if the participant is provided with English auditory and written input (Group E-E), or
only English auditory input.
Graph 22: The Average Scores for Frequent and Non-frequent Items across the Different Groups in Percentage.
In summary, there is a slight difference in the acquisition of frequent and non-frequent
words, but it is not statistically significant. This difference is noticeable in the results of the
entire group, as well as in the results of the individual groups. As with the cognates and
noncognates, there is a gradual decline in average scores, both for frequent and non-frequent
items – ranging from group E-D to group D. But none of these differences prove to be
statistically significant. Interestingly, the difference in average scores between frequent items
and non-frequent words increases, it is smallest in group E-D and steadily increases across
group E-E, group E and group E.
74,74% 72,63%
70,53%
64,84% 70,95%
67,16% 64,42%
57,26% 50%
55%
60%
65%
70%
75%
80%
85%
90%
95%
100%
Group E-‐D Group E-‐E Group E Group D
Frequent
Non-‐Frequent
76
4.4.4 Word Frequency The previous paragraph dealt with the frequency of words within the 15-minute film
fragment. This one discusses word frequency in a more general sense, it discusses the
correlation between the frequency of a word in English (so not just in the film fragment, but
throughout the entire language) and the average score of the word. The frequency of a word
is indicated by a rank6 (see Appendix 5 for the rank of each word in the BNC). The literature
suggests that a higher rank results in a higher score. The rank of the words in Appendix 5
originates from a frequency list of all English words (the word ‘The’ gets the rank 1, because
it is the most frequent word in English, the word ‘Make’ gets rank 45). Table 13 gives a
limited frequency list of the 50 words used in the vocabulary test. The word ‘Make’ gets rank
‘1’ in this list, since it is the most frequent of the 50 words, the word ‘Rain’ (rank 4661 in the
BNC) gets rank ‘50’ in this frequency list, since it is the least frequent of the 50 words.
Rank Word Rank in the BNC Rank Word Rank in the BNC 1. Make 45 26. Close 649 2. Know 47 27. Seek 668 3. Time 52 28. Happy 747 4. Day 90 29. Far 975 5. Find 95 30. Deep 976 6. Give 98 31. Safe 1059 7. Good 110 32. Threat 1067 8. Life 114 33. Rich 1076 9. World 123 34. Travel 1082 10. Keep 156 35. Forest 1109 11. Small 203 36. Battle 1201 12. Like 208 37. Wild 1296 13. Live 210 38. Light 1306 14. Mother 229 39. Bright 1312 15. Little 256 40. Survive 1346 16. Real 305 41. Victory 1467 17. Grow 353 42. Danger 1551 18. Boy 382 43. Enemy 1601 19. Possible 459 44. Escape 2045 20. Leader 463 45. Territory 2192 21. Full 503 46. Independence 2305 22. Special 520 47. Amazing 2308 23. Eat 543 48. Precious 4020 24. Catch 586 49. Rival 4384 25. Difficult 608 50. Rain 4661
Table 13. Frequency List of the 50 Words in the Vocabulary Test.
6 The British National Corpus (BNC) has been used to determine the rank of each word. The corpus has
been accessed through the link http://corpus.byu.edu/bnc/
77
The discussion of paragraph 4.4.3 dealt with a binary opposition, a word was either
frequent or non-frequent in the film fragment. This paragraph deals with a spectrum, ranging
from more frequent to less frequent within the English language as a whole. The rank of each
word in the BNC was not a specific parameter for selecting the words in the vocabulary test.
So in order to discuss the correlation between word frequency in the BNC and the average
test score, the words have been divided into groups of ten words each. This division into
groups of ten has been made with regard to the rank of each word in the BNC. The words in
the group with rank 1-10 in Table 13 (Make – Keep) have a rank between 1 and 200 in the
BNC. Those in the group with rank 11-20 (Small – Leader) rank between 200 and 500 in the
BNC. The words with rank 21-30 (Full – Deep) rank between 500 and 1000 in the BNC. The
words with rank 31-40 (Safe – Survive) rank between 1000 and 1500 in the BNC and finally,
the words with rank 41-50 (Victory – Rain) rank between 1500 and 5000 in the BNC. In
Group 1-10, the range of ranks in the BNC is only 200, whereas it is 3500 in Group 41-50.
This makes sense, since words with a lower rank in the BNC are more frequent, so they are
more likely to occur in any type of text, including the film fragment used for this research
paper. The lower the rank of a word in the BNC, the less likely it is that such word occurred
in the film fragment.
Now, when the average scores of each rank group are compared, it becomes clear that
there is a correlation between word frequency and average test scores. Graph 23 gives an
overview of this correlation. As the graph indicates, the frequency of a word does in fact
influence how often it has been translated correctly in the vocabulary test. The words with
rank 1-10 in the frequency list score an average of 88,82%, those with rank 11-20 score an
average of 80%, those with rank 21-30 averagely score 69,87%, those with rank 31-40
averagely score 52,24% and finally, those with rank 41-50 averagely score 48,16%. So there
is a difference of more than 40% between the ten most frequent and the ten least frequent
words. A one-way ANOVA indicates that the differences between these five groups are
statistically significant: p < .01 (see Table 14).
78
Graph 23: The Correlation between Word Frequency in the BNC and Average Test Score in Groups of 10
Words (in %).
Analysis of Variance (One-‐Way) Summary
Groups Sample size Sum Mean Variance Rank 1-‐10 10 888,16053 88,81605 350,96649
Rank 11-‐20 10 799,99868 79,99987 160,74981 Rank 21-‐30 10 698,68526 69,86853 816,96161 Rank 31-‐40 10 522,36829 52,23683 893,7703 Rank 41-‐50 10 481,58039 48,15804 702,98582
ANOVA Source of Variation SS df MS F p-‐level F crit
Between Groups 12.228,07707 4 3.057,01927 5,2249 0,00152 2,57874 Within Groups 26.328,90628 45 585,08681
Total 38.556,98336 49
Table 14: One-Way ANOVA of Word Frequency in the BNC vs. Average Test Score.
The correlation between word frequency in the BNC and the average test score of the
words is also noticeable in the different test groups and the control group. As in every other
category discussed, the E-D group scored the highest average score for each rank group, the
E-E group was second, followed by the E group and the D group (the control group) was the
last of the four. The scores of all groups show the same decline; the more frequent a word is
(so the lower its rank), the higher the average test score for that word. Interestingly, graph 24
indicates that the differences between the different test groups are smaller for rank 1-10 and
they grow larger as the ranks increase. This indicates that different subtitling situations do
not make a large difference regarding the acquisition of frequent words, but they do make a
88,82
80
69,87
52,24 48,16
0
10
20
30
40
50
60
70
80
90
100
Rank 1-‐10 Rank 11-‐20 Rank 21-‐30 Rank 31-‐40 Rank 41-‐50
79
difference when it comes to the acquisition of less frequent ones. The difference in averages
scores between group E-D and group D is only 6,84% for the words in rank 1-10, whereas
this difference is 19,48% for rank 41-50 (see Table 15). The average scores of the three test
groups stay relatively close to one another for rank 1-30, but start to grow apart from rank
31-40 onwards.
Rank 1-‐10 Rank 11-‐20 Rank 21-‐30 Rank 31-‐40 Rank 41-‐50
Group E-‐D 91,58 84,21 72,63 58,42 57,37 Group E-‐E 90,53 82,11 71,58 56,84 51,05 Group E 88,42 81,58 68,42 51,58 46,32 Group D 84,74 72,11 66,84 42,11 37,89
Table 15: The Average Score in Percentage per Test Group for the Different Rank Groups.
Graph 24: The Correlation between the Frequency in the BNC and the Average Score per Test Group
in %.
In summary, there is a significant difference in the average scores of more frequent
words in the BNC and less frequent ones. This difference is noticeable for the group as a
whole as well as for each individual test group. Within the different test groups, the average
scores per rank group lay very close to one another for the most frequent words, but the less
frequent the words become, the larger the differences in average scores for each test group
grow. First of all, this indicates that exposure to English input influences the acquisition of
less frequent words. But more importantly, this also implies that the more accessible the
English input is, the better less frequent words are acquired.
30
40
50
60
70
80
90
100
Rank 1-‐10 Rank 11-‐20 Rank 21-‐30 Rank 31-‐40 Rank 41-‐50
E-‐D
E-‐E
E
D
80
4.4.5 Visual Represtentation of the Vocabulary Input Out of the 50 lexical items, 14 words were explicitly accompanied by their visual
representation in the film fragment. What is meant by ‘explicitly accompanied by a visual
representation’ is the following: if the narrator of the film fragment uses the word Forest,
there is a visual image of the forest at the exact time the narrator uses this word. For the
nouns, this is self-explanatory; the noun Forest is accompanied by an image of a forest. For
the verbs, this implies that the verb to rain is used to describe a scene in which it is raining
and for the adjectives, this means that the adjective small is used to describe a small
chimpanzee in the scene. These 14 words are per definition concretes, since abstract words
like Time cannot be represented by a single visual image.
Paivio’s Dual Coding Theory claims that there is a better retention of information when
multiple channels of input are combined. So according to this theory, the 14 words that are
explicitly accompanied by their visual representation should be translated correctly more
often. A comparison of the average scores of the 14 words with a visual representation and
the 36 words without it shows that there is a difference between these types of words. The 14
words with a visual representation score an average of 72,19%, those without visual
representation score 65,94%. This is a difference of a good 5%. However, a one-way
ANOVA indicates that this difference is not statistically significant: p > .05 (see Table 16).
Graph 25: Visual Representation of a Word vs. Average Test Score.
72,19% 65,94%
0%
10%
20%
30%
40%
50%
60%
70%
80%
90%
100%
Visual RepresentaYon No Visual RepresentaYon
81
Analysis of Variance (One-‐Way) Summary
Groups Sample size Sum Mean Variance Visual Representation 15 1.082,89579 72,19305 553,74102
No VR 35 2.307,89737 65,93992 893,94242 ANOVA
Source of Variation SS df MS F p-‐level F crit Between Groups 410,56688 1 410,56688 0,51662 0,47577 4,04265 Within Groups 38.146,41648 48 794,71701
Total 38.556,98336 49
Table 16: One-way ANOVA of Visual Representation vs. Average Test Score.
The same distinction is present in each of the test groups, but not in the control group see
Graph 26). In the control group, words without a visual representation score better than those
with visual representation. This is not surprising, since the control group watched the
fragment without English input. This group did not get the same combination of auditory and
visual input and therefore, the Dual Coding Theory does not apply for the control group.
Within the different test groups (E-D, E-E and E), there is a difference between words with a
visual representation and those without. Furthermore, the difference between words with a
visual representation and those without grows larger as the variety of input declines. In group
E-D, the difference is 3,12%. This group got exposed to three different forms of input: visual
information, English audio and Dutch subtitles. In group E-E, this difference is 9,78%.
Group E-E was exposed to visual information and English audio and subtitles. In group E,
the difference was 14,23% and this group was only exposed to visual information and
English audio.
Graph 26: The Correlation of a Visual Representation of a Word (VR) and the Average Score (in %) per
Group.
75 74,74 73,33 65,61
71,88 64,96
59,10 68,22
0
10
20
30
40
50
60
70
80
90
100
E-‐D E-‐E E D
VR
No VR
82
So there is a difference in the average scores of words with a visual representation and
those without it. But this difference only applies for the groups who were exposed to English
audio, which makes sense in light of the Dual Coding Theory. The average scores for words
with a visual representation is almost equal between the three control groups (resp. 75%,
74,74% and 73,33%), whereas the average scores for words without an explicit visual
representation grows smaller as the variety of input grows smaller (resp. 71,88%, 64,96%
and 59,10%).
In summary, the nature of vocabulary and its influence on vocabulary acquisition
through subtitling consists of four elements. First, there is the influence of cognacy. The
results from this study indicate that cognates are acquired far more easily than noncognates.
This proves to be the fact for the groups as a whole, as well as for each subtitling groups
individually. On the other hand, no significant differences in the acquisition of cognates and
noncognates were observed between the different control groups. None of the groups proved
to be significantly more successful in the acquisition of cognates than the other groups and
the same fact was observed for noncognates. Secondly, a difference in the acquisition of
frequent and non-frequent items (in the film fragment) has been found, but this difference
does not prove statistically significant. It was not significant for the group as a whole or for
the different groups individually. On the other hand, the difference in the scores for frequent
and non-frequent items grew larger between group 1 (E-D) and group 4 (D). The third
characteristic of vocabulary, frequency in the English language, proved to be very influential
for the group as a whole and for the different subtitling groups. There were no major
differences between the individual groups, but the observed difference did grow larger as the
frequency of the discussed words grew smaller. The final characteristic was the visual
representation of a word. There is a small difference in the acquisition of words with a visual
representation and those without, but this difference was not significant. So the influential
characteristic of the vocabulary input were cognacy and frequency in the English language.
The scores of the other two characteristics, frequency in the film fragment and visual
representation, were not different enough to be significant.
83
5 Conclusion This research paper consisted of three research questions, which all dealt with second
language vocabulary acquisition through subtitling. The first question focussed on different
types of subtitling and the effect of those types on the acquisition of English vocabulary. The
second dealt with the influence of the participant’s background on the acquisition of English
vocabulary and the third focussed on specific characteristics of the vocabulary input and the
acquisition of that input.
The hypothesis on the first research question stated that the groups exposed to an
English soundtrack (groups E-D, E-E and E) would all acquire vocabulary after watching the
subtitled film fragment and that the first group (E-D) would acquire more vocabulary than
the second group (E-E), which would in turn acquire more words than the third group (E).
This hypothesis, based on the research by e.g. d’Ydewalle & Van de Poel (1999) and
Koolstra & Beentjes (1999) has partially been proven. The first part of the hypothesis proved
to be correct: the groups exposed to English audio scored significantly higher than the
control group. The average score of the three test groups combined was 70,10%, which was
10% higher than the average score of the control group (60,95%). Secondly, the E-D group
did in fact score higher than the E-E group and the E-group, which showed that the second
part of the hypothesis was also correct. On the other hand, the difference between the
different types of subtitling did not prove to be statistically significant. Even though there
was no statistical significance, the difference between each of the four different test groups
was noticeable in the discussion of other variables in the test (gender, self-confidence,
cognacy, frequency in the film fragment, frequency in the English language and visual
representation).
For the second research question, the hypothesis stated that a high exposure to English
and a low affective filter would result in a higher score on the vocabulary test. The
discussion of the questionnaire showed that the participants of this research are very often
exposed to English, through various media. As expected, they often watch English television
and films and they do so with Dutch subtitles (either regularly or always). The largest part of
the group (56,58%) prefers Dutch subtitles when watching English media and half of the
group watches those media with Dutch subtitles. The discussion of the correlation between
the test results and the exposure to English television and films showed that exposure to
English films and television is a very influential factor in second language vocabulary
acquisition. There was a difference in average score of more than 35% between the group of
84
participants who watch English media daily and those who never do so. Secondly, the way in
which those media are watched also makes a difference. Those who most often watch those
media without subtitles score highest, followed by English subtitles, Dutch subtitles and
finally dubbed media. Next, the data showed that the participants are very often exposed to
English on the Internet and even though there are differences between participants who never
encounter English on the Internet and those who do so daily, these differences are not
significant. A type of exposure which did prove to be significant is videogames. The
majority of the groups plays English videogames several times a month and this form of
exposure proves to be quite influential on the average test scores. The vast majority of the
participants listen to English music most often, which made it impossible to discuss the
exposure to English music as a variable, since there was barely any variation in the answer to
question 6, although this would have been interesting for the discussion of cognacy, since
Kuppens (2007) observed the influence of English music on the acquisition of noncognates.
An unanticipated differentiator was gender. The average scores of the boys was quite a bit
higher (almost 10%) than the average score of the girls. In each of the test groups, this
difference was visible, it was between 11% and 12% for the groups who were exposed to
English audio and only 6% for the control group. Next to exposure to English, affective
factors were also the research topic of the second research question. The data showed that
self-confidence was a very influential factor in the process of SLA. Those who believe their
English is very good score more than twice as high as those who think their knowledge of
English is very bad. There is a clear correlation between self-confidence and test score,
although the question remains which is cause and which is effect. In any case, this difference
was again visible in each test group. A second affective factor, the estimated difficulty of
English did not prove to be an influential factor. So the answer to the second research
question is in line with the hypothesis on that question. A high exposure to English input
(through films, television and videogames) and a low affective filter (in the form of self-
esteem) as described by Krashen in his Affective Filter Hypothesis, result in a high score on
the vocabulary test. A variable which had not been included in the hypothesis was gender, as
it was not anticipated that there would be such a large difference in the scores of boys and
girls.
The final research question dealt with the nature of vocabulary input. Four
characteristics were discussed: cognacy, frequency in the film fragment, frequency in the
English language and visual representation. Two of those characteristics proved to be very
85
influential, the other two did not. In accordance with the literature (Kuppens 2007, Crossley
et al. 2013) cognacy and frequency in the English language influenced the average scores.
Frequency in the film fragment was not an influential characteristic, although the difference
in frequent and non-frequent items grew larger as the variety of input grew smaller (from
group 1: E-D to group 4:D). The visual representation of a word during the film fragment
was also not a significant differentiator. Based on the study by Plass et al. (2003) Paivio’s
Dual Coding Theory, the hypothesis for this aspect predicted that a visual representation
would make a difference. This did not prove to be the case in this study, but that does not
disprove the Dual Coding Theory, which claims that the combination of multiple channels of
input results in a better retention of information. The film fragment did provide multiple
channels of information and the variety of those channels had an influence on the average
test scores.
In summary, this study showed that watching English-spoken television (in either
subtitled or non-subtitled form) results in a higher score on a vocabulary test including words
from that film fragment. A high exposure to English (subtitled) media and a high level of self
esteem both lead to a higher score on the vocabulary test and cognate and frequent words in
English are acquired more easily than noncognate and less frequent ones. The exposure to
English through the Internet was not influential and the frequency of a word in the film
fragment as well as the visual representation of a word did also not prove to be influencing
factors.
In further research, it would be interesting to investigate whether the difference in
average scores between boys and girls in this study was a coincidence or if this is a tendency
for all 12-year olds. Next, the acquisition of English grammar through subtitling and the
possible influence of different types of subtitling would also be interesting. In any case, the
subjects studied in this paper could also be studied in a larger setting, with more participants
of different age groups.
86
6 References
Ahangari, Saeideh & Zeinab Abdollahpour. 2010. The effects of multimedia annotations on Iranian EFL learners’ L2 vocabulary learning. The Journal of Applied Linguistics 3(1). 1-19.
Baker, Colin Ronald, Sylvia Prys Jones. 1998. Encyclopedia of bilingualism and bilingual education. Clevedon: Multilingual Matters.
Bot, Kees de, Wander Lowie & Marjolijn Verspoor. 2005. Second language acquisition: An advanced resource book. London: Routledge.
Bruls, E. & E. Kerkman. 1988. Ondertiteling in beeld: Een communicatiewetenschappelijk en - historisch onderzoek naar ondertiteling in film en televisie. Amsterdam: University of Amsterdam.
Chimpanzee. 2012. Dir. Alastair Fothergill and Mark Linfield. Walt Disney Studios Motion Picture.
Concise Oxford English Dictionary. 11th ed. 2008���
Crossley, Scott A., Nicholas Subtirelu & Tom Salsbury. 2013. Frequency effects or context effects in second language word learning. Studies in Second Language Acquisition 35. 727-755.
Crystal, David. 2003. The Cambridge encyclopaedia of the English language. Cambridge: Cambridge University Press.
Danan, Martine. 2004. Captioning and subtitling: Undervalued manguage learning strategies. Meta: Journal des traducteurs / Meta: Translators’ Journal 49(1). 67-77.
Declercq, Karolien, Katrijn Denies & Rianne Janssen. 2012. Vlaamse vreemdetalenkennis in Europees perspectief: Balans van het ESCL-onderzoek. Kampenhout: Artoos.
De Ro, Jo (ed.). 2008. Education in Flanders. A broad view of the Flemish educational landscape. Haletra: Houthalen-Helchteren: Haletra.
Duñabeitia, Jon Adoni, Manuel Perea & Manuel Carreiras. 2010. Translation priming effects with simultaneous bilinguals. Experimental Psychology 57(2). 98-107.
d’Ydewalle, Géry. 2002. Foreign language acquisition by watching subtitled television programs. Journal of Foreign Language Education and Research 12. 59-77.
d’Ydewalle, Géry & Marijke Van de Poel. 1999. Incidental foreign-language acquisition by children watching subtitled television programs. Journal of Psycholinguistic Research 28(3). 227- 244.
87
d’Ydewalle, Géry & Ubolwanna Pavakanun. 1995. Acquisition of a second/foreign language by viewing a television program. In P. Winterhoff-Spurk (ed.), Psychology of media in Europe: The state of the art – perspectives for the future, 51–64. Opladen: Westdeutscher.
d’Ydewalle, Géry & Ubolwanna Pavakanun. 1996. Le sous-titrage `a la télévision facilite-t-il l’acquisition des langues [Does television subtitling facilitate language acquisition]? In Y. Gambier (ed.), Les transferts linguistiques dans les meédias audiovisuels, 217– 223. Lille: Presses Universitaires du Septentrion.
d’Ydewalle, Géry & Ubolwanna Pavakanun. 1997. Could enjoying a movie lead to language acquisition? In P. Winterhoff-Spurk & T. Van der Voort (eds.) New horizons in media psychology, 145–155. Opladen: Westdeutscher.
d’Ydewalle, Géry & Wim De Bruycker. 2007. Eye movements of children and adults while reading television subtitles. European Psychologist 12(3). 196-205.
D’Ydewalle, Géry & I. Gielen. 1992. Attention allocation with overlapping sound, image, and text. In K. Rayner (ed.) Eye Movements and Visual Cognition: Scene Perception and Reading, 415-427. New York: Springer-Verlag.
EACEA. 2009. Study on the use of subtitling. The potential of subtitling to encourage foreign language learning and improve the mastery of foreign languages. Paris: Media Consulting Group.
Ellis, Nick C. 2002. Frequency effects in language processing. A review with implications for theories of implicit and expilicit language acquisition. Studies in Second Language Acquisition 24. 143-188.
Ellis, Nick & Laura Collins. 2009. Input and Second Language Acquisition: The roles of frequency, form, and function. The Modern Language Journal 93(3). 329-336.
Ellis, Rod. 1994. The study of Second Language Acquisition. Oxford: Oxford University Press.
Ellis, Rod. 2002. Second Language Acquisition. Oxford: Oxford University Press.
Ellis, Rod & Sandra Fotos. 1999. Learning a second language through interaction. Amsterdam: Benjamins.
Gardner, Robert C. 2001. Integrative motivation and Second Language Acquisition. In Z. Dornyei and R. Schmidt (eds.), Motivation and Second Language Acquisition. Honolulu: University of Hawai.
Gass, Susan M. 2011. Input interaction and the second language learner. New York:
Routledge.
Gass, Susan M. & Larry Selinker. 2008. Second language acquisition: An introductory course. London: Routledge.
88
Ghia, Elisa. 2012. Subtitling matters: New perspectives on subtitling and foreign language learning. New York: Peter Lang.
Gielen, M. 1988. Perceptie en ondertitels: De paravofeale en perifere informatieverwerking van ondertitels [Perception and subtitles: The paravofeal and peripheral information processing of subtitles]. Leuven: University of Leuven.
Goethals, Michaël. 1997. English in Flanders (Belgium). World Englishes 16(1). 105-114.
Gottlieb, Henrik. 1997. Subtitles, translation & idioms. Copenhagen: University of Copenhagen.
Gottlieb, Henrik. 2004. Subtitles and international anglification. Nordic Journal of English Studies 3(2). 219-230.
Horwitz, Elaine K., Michael B. Horwitz & Joanne Cope. 1986. Foreign Language Classroom Anxiety. The Modern Language Journal 70(2). 125-132.
Houthuys, Siona. 2011. The differences between Flemish pupils acquiring English and Walloon pupils acquiring English before the input of formal instruction. Unpublished dissertation: University Ghent.
Hulstijn, Jan H. 2001. Intentional and incidental second language vocabulary learning: a reappraisal of elaboration, rehearsal and automaticity. In Robinson, Peter (ed.) Cognition and Second Language Instruction, 258-286. Cambridge: University Press.
Jiang, Nan. 2000. Lexical representation and development in a Second Language. Applied Linguistics 21(1). 47-77.
Kachru, Braj B. 1992. The Other tongue: English across cultures. Urbana: University of Illinois Press.
Klein, Wolfgang. 1986. Second Language Acquisition. Cambridge: University Press.
Koolstra, Cees M. & Johannes W.J. Beentjes. 1999. Children’s vocabulary acquisition in a foreign language through watching subtitled television programs at home. Educational Technology Research and Development 47(1). 51-60.
Koolstra, Cees M., T.H.A. van der Voort & Géry d’Ydewalle. 1999. Lengthening the presentation time of subtitles on television: Effects on children’s reading time and recognition. Communications 24. 407–22.
Koolstra, Cees M., Allerd L. Peeters & Herman Spinhof. 2002. The pros and cons of dubbing and subtitling. European Journal of Communication 17(3). 325-354.
Krashen, Stephen D. 1981. Second language acquisition and second language learning. Oxford: Pergamon.
Krashen, Stephen D. 1984. Principles and practice in Second Language Acquisition. Oxford: Pergamon.
Krashen, Stephen D. 1987. Principles and practice in Second Language Acquisition.
89
Hertfordshire: Prentice-Hall International.
Krashen, Stephen. 1998. Comprehensible output? System 26. 175-182.���
Krashen, Stephen D., & Tracy D. Terrell. 1983. The Natural Approach: Language Acquisition in the Classroom. Oxford: Pergamon Press.���
Kuppens, An. 2007. De invloed van het mediagebruik op de verwerving van Engelse woordenschat: een empirische studie bij Vlaamse jongeren. Tijdschrift voor communicatiewetenschap 35(4). 325-336.���
Levelt, Willem. 1989. Speaking: From intention to articulation. Cambridge: Bradford.
Lippens, Caroline. 2010. The incidental acquisition of English before the input of formal instruction: a contrastive analysis of TSO, B-stream & ASO in Flanders. Unpublished dissertation: Universiteit Gent.
Luyken, G., T. Herbst, J. Langham-Brown, H. Reid a H. Spinhof. 1991. Overcoming language barriers in television: Dubbing and subtitling for the European audience. Manchester: European Institute for the Media.���
Mayer, Richard E. & Richard B. Anderson. 1992. The instructive animation: Helping students build connections between words and pictures in multimedia learning. Journal of Educational Psychology 84. 444-452.
Ondarra, Kristi Jauregi. 1997. Collaborative negotiation of meaning: a longitudinal approach. Amsterdam: Rodopi.
Paivio, Allan. 2014. Intelligence, Dual Coding, and the brain. Intelligence 47. 141–158.
Pavakanun, Umbolwanna & Géry d’Ydewalle. 1992. Watching foreign television programs and language learning. In F. L. Engel, D. G. Bouwhuis, T. Bösser, & G. d’Ydewalle (eds.), Cognitive modelling and interactive environments in language learning, 193– 198. Berlin: Springer.
Plass, Jan L., Dorothy M. Chun, Richard E. Mayer & Detlev Leutner. 2003. Cognitive load in reading a foreign language text with multimedia aids and the influence of verbal and spatial abilities. Computers in Human Behavior 19(2). 221-243.
Perego, Elisa., Fabio Del Missie, Marco Porta, & Mauro Mosconi. 2010. The cognitive effectiveness of subtitle processing. Media Psychology 13(3). 243–272.
Rogers, James, Stuart Webb & Nakata Tatsuya. 2015. Do the cognacy characteristics of loanwords make them more easily learned than noncognates? Language Teaching Research 19(1). 9-27.
90
Skehan, Peter. 1989. Individual differences in second-language learning. London: Edward Arnold. Spinhof, Herman and Allerd L. Peeters. 1999. Opinies over nasynchroniseren en
ondertitelen [Opinions about Dubbing and Subtitling]. Hilversum: NOS.
Vanderplank, Robert. 1988. The value of teletext sub-titles in language learning. ELT Journal 42(4). 272-281.
Van Lommel, Sven, Annouschka Laenen & Géry d’Ydewalle. 2006. Foreign-grammar acquisition while watching subtitled television programmes. British Journal of Educational Psychology 76. 243-258.
Watson-Gegeo, Karen & Ann Sarah Nielsen. 2003. Language Socialization in SLA. In Dougty, Catherine J. & Michael H. Long (eds.) The handbook of second language acquisition, 155-177. Oxford: Blackwell Publishing Ltd.
Xu, Jianwei & Kris Van de Poel. 2011. English as a Lingua Franca in Flanders: A study of university students’ attitudes. English Text Construction 4(2). 257-278.
Zarei, Abbas Ali. 2009. The effect of bimodal, standard, and reversed subtitling on L2 vocabulary recognition and recall. Pazhuhesh-e Zabanha-ye Khareji 49. 65-85.
Zipf, George K. 1935. The psychobiology of language: An introduction to dynamic philology. Cambridge: MIT Press.
91
7 Appendices 7.1 Appendix 1: Questionnaire
Vul in:
! Naam:
! Klas:
! Leeftijd:
! Jongen/Meisje:
! Moedertaal:
! Andere talen die je spreekt:
! Heb je al taallessen of taalkampen gevolgd voor het Engels?
Kruis aan wat past:
1. Kijk je vaak naar televisie en films?
Ja
Nee
2. Hoe vaak kijk jij naar Engelstalige programma’s of films?
Nooit
Een paar keer per jaar
Een paar keer per maand
Een paar keer per week
Elke dag
92
3. Als je naar Engelstalige programma’s en films kijkt, is dat dan met Nederlandstalige
ondertiteling?
Altijd
Af en toe
Nooit
4. Waar kijk je het meeste naar?
Nederlandstalige programma’s en films
Engelstalige programma’s en films met Nederlandstalige ondertiteling
Engelstalige programma’s en films zonder ondertiteling
Engelstalige programma’s en films met Engelstalige ondertiteling
5. Waar kijk je het liefst naar?
Nederlandstalige programma’s en films
Engelstalige programma’s en films met Nederlandstalige ondertiteling
Engelstalige programma’s en films zonder ondertiteling
Engelstalige programma’s en films met Engelstalige ondertiteling
93
6. Naar welke muziek luister je het meest?
Nederlandstalige muziek
Franstalige muziek
Engelstalige muziek
7. Hoe vaak kom je met het Engels in contact op het internet?
Nooit
Een paar keer per jaar
Een paar keer per maand
Een paar keer per week
Elke dag
8. Op welke websites kom je dan in contact met het Engels?
(Meerdere antwoorden mogelijk)
YouTube
Een gaming website
Blogs
Andere:…………………………………..
94
9. Speel je vaak Engelstalige games op de Wii, Playstation, Xbox of op het internet?
Nooit
Een paar keer per jaar
Een paar keer per maand
Een paar keer per week
Elke dag
10. Kom je nog op andere manieren in contact met het Engels?
Engelstalige vrienden of contacten
Op reis
Via familie
Andere……………………………
Nee
11. Vind je het belangrijk om Engels te leren?
Ja,
omdat:
Nee,
omdat:
95
12. Wat vind je van je eigen Engelse kennis, in vergelijking met je leeftijdsgenoten?
Heel slecht
Slecht
Matig
Goed
Heel goed
13. Vind je Engels een moeilijke taal?
Ja,
omdat:
Nee,
omdat:
96
7.2 Appendix 2: Transcript of the 15-minute Film Fragment Once, not long ago, in deepest Africa, there was a great rain forest, a wild, magical land,
barely touched by humankind. It's a special day in the dark heart of this forest. There's a new
life, a precious baby boy named Oscar. What's so special about this little guy? Well,
something truly amazing will transform his life. And that's what our story is all about.
Beneath this rain forest canopy is a world that you or I may never set eyes on. It's a
world with few people, but still, it's full of drama, sadness, and joy. About three months have
passed since little Oscar brought light into his mother's life. Isha couldn't be happier with her
new baby boy. Oscar's world revolves around Mom, but there are 35 other chimpanzees in
his group that make a kind of extended family, and they're never far away. Perhaps one day
Oscar will become an old-timer like Grandpa, who will turn 50 this year. But there are other
kids, too. They'll become Oscar's playmates.
It's a rich society, but in the heart of the jungle. Now, Oscar won't be in his prime until
he's over 20, but one day, he'll look like this big guy. This is Freddy, and he's large and in
charge. His white beard shows his years of wisdom and experience. Right now, Freddy is
untouchable. And like most alpha males, he couldn't care less about the young upstarts who
would just love to take over his job someday. Yeah, yeah. Seen it all before. You need brains
to make a living in this forest. The jungle toolbox can help, but only if you're bright enough
to use it. Sometimes it just takes a little home-schooling available from your friends and
family. See? Like this. Right there. Even Rufus catches on, eventually. There you go. Bugs
on a stick.
It's finally time for Oscar to leave his mother's side. For Isha 's little boy, this is
independence day, and independence comes one step at a time. Well, that's enough
independence for one day. Oscar's forest home is a prime piece of chimpanzee real estate
that appears rich and lush, but it's a hard place to make a living. The jungle itself is a living,
breathing thing that doesn't want to be eaten. Finding food here is a full-time job that
involves the whole family, including Oscar. Now that he's grown up a little, he can start to
explore on his own. But he's not above catching a lift when he can. Really, why walk when
you can hitch a ride? Keep it moving, Mom. Freddy leads the search for breakfast. It helps to
have someone who knows where they're going, especially when the restaurant is never in the
same place. Today the diner isn't even in the same neighborhood. To find the fruit they'll
need, they will have to travel much further and take a risk. Freddy knows the only fruit
available now is on the far side of their kingdom, out of their safe zone.
97
This border has been a battleground for generations. Everyone is on maximum alert and
treads as quietly as possible. Freddy is leading them close to their greatest enemies. This
ridge marks the start of territory owned by a group of rival chimpanzees. The forbidden fruit
is finally in reach. Shh, Oscar. Shh. Freddy sees his team are outnumbered two to one. The
enemy has a formidable leader, Scar. He and his team have had Freddy in their sights for
years. Oscar clings to Isha. They dare not make a move. The rivals have caught their scent.
Oscar and the others are in real danger. Scar's drum is a threat. For Isha, it's run or die.
Freddy leads the escape. By running off Freddy's group, their enemies have claimed a small
victory. But that's just the beginning. Scar and his troops are intent on winning the war, not
just a single battle. The prize, a grove of nut trees in the center of Freddy's land. The grove is
an oasis of food in this difficult forest and one that Scar's army has to conquer if they're
going to survive and thrive. Isha knows that Scar is the greatest threat to little Oscar, and the
run-in was too close for comfort. She'll do everything possible to keep her son safe. Oscar
will rely on his mother's milk for a good while yet, and to keep that milk flowing Isha relies
on the rich nuts in Freddy's kingdom. The magnificent nut groves, so sought-after by Scar,
thrive here because of the rich soil, and another magic ingredient, life-giving water.
98
7.3 Appendix 3: Words in the Vocabulary Test
Noun Verb Adjective
Cognate Noncognate Cognate Noncognate Cognate Noncognate
Frequent 1. Day
2. Life
3. World
4. Time
5. Mother
9. Forest
10. Independence
11.Boy
12. Threat
17. Make
18. Live
19. Rain
20. Seek
25. Like
26. Know
27. Keep
28. Close
33. Rich
34. Special
35. Full
36. Far
41. Little
42. Safe
43. Real
44. Possible
Non-
Frequent
6.
Territory
7. Rival
8. Leader
13. Enemy
14. Danger
15. Battle
16. Victory
21. Grow
22. Give
23. Find
24. Eat
29. Travel
30. Survive
31. Escape
32. Catch
37. Wild
38. Good
39. Deep
40. Light
45. Difficult
46. Small
47. Bright
48. Happy
49. Amazing
50. Precious
99
7.4 Appendix 4: Test Results per Participant
7.4.1 Group 1: E-D
First name Percentage Points received Points available
X 98,00 49 50 X 96,00 48 50 X 84,00 42 50 X 44,00 22 50 X 60,00 30 50 X 52,00 26 50 X 72,00 36 50 X 94,00 47 50 X 88,00 44 50 X 66,00 33 50 X 52,00 26 50 X 90,00 45 50 X 80,00 40 50 X 60,00 30 50 X 56,00 28 50 X 78,00 39 50 X 56,00 28 50 X 90,00 45 50 X 68,00 34 50
Average 72,84
100
7.4.2 Group 2: E-E
First name Percentage Points received
Points available
X 92,00 46 50 X 86,00 43 50 X 78,00 39 50 X 48,00 24 50 X 72,00 36 50 X 42,00 21 50 X 44,00 22 50 X 56,00 28 50 X 46,00 23 50 X 92,00 46 50 X 86,00 43 50 X 82,00 41 50 X 70,00 35 50 X 64,00 32 50 X 36,00 18 50 X 88,00 44 50 X 90,00 45 50 X 88,00 44 50 X 68,00 34 50
Average 69,89
101
7.4.3 Group 3: E
First name Percentage Points received
Points available
X 48,00 24 50 X 40,00 20 50 X 78,00 39 50 X 82,00 41 50 X 76,00 38 50 X 58,00 29 50 X 60,00 30 50 X 70,00 35 50 X 74,00 37 50 X 78,00 39 50 X 70,00 35 50 X 74,00 37 50 X 88,00 44 50 X 58,00 29 50 X 84,00 42 50 X 70,00 35 50 X 66,00 33 50 X 76,00 38 50 X 32,00 16 50
Average 67,47
102
7.4.4 Group 4: D
First name Percentage Points received
Points available
X 40,00 20 50 X 90,00 45 50 X 66,00 33 50 X 48,00 24 50 X 36,00 18 50 X 68,00 34 50 X 76,00 38 50 X 54,00 27 50 X 54,00 27 50 X 68,00 34 50 X 64,00 32 50 X 54,00 27 50 X 26,00 13 50 X 74,00 37 50 X 60,00 30 50 X 86,00 43 50 X 46,00 23 50 X 62,00 31 50 X 88,00 44 50
Average 61,05
103
7.5 Test Results per Lexical Item
Nr. Word Average Score Group 1 E-D Group 2 E-E Group 3 E Group 4 D 1 Day 100,00 100,00 100,00 100,00 100,00 2 Life 96,05 94,74 100,00 94,74 94,74 3 World 100,00 100,00 100,00 100,00 100,00 4 Time 100,00 100,00 100,00 100,00 100,00 5 Mother 97,37 100,00 94,74 100,00 94,74 6 Territory 68,42 78,95 68,42 68,42 57,89 7 Rival 38,16 52,63 42,11 36,84 21,05 8 Leader 75,00 84,21 63,16 78,95 73,68 9 Forest 52,63 57,89 63,16 42,11 47,37 10 Independence 9,21 15,79 10,53 0,00 10,53 11 Boy 98,69 94,74 100,00 100,00 100,00 12 Threat 3,95 15,79 0,00 0,00 0,00 13 Enemy 43,42 52,63 52,63 36,84 31,58 14 Danger 90,79 100,00 89,47 89,47 84,21 15 Battle 86,84 94,74 89,47 94,74 68,42 16 Victory 47,37 68,42 52,63 26,32 42,11 17 Make 96,05 100,00 94,74 100,00 89,47 18 Live 85,53 100,00 94,74 78,95 68,42 19 Rain 67,11 63,16 63,16 84,21 57,89 20 Seek 19,74 21,05 26,32 21,05 10,53 21 Grow 63,16 57,89 73,68 68,42 52,63 22 Give 100,00 100,00 100,00 100,00 100,00 23 Find 81,58 78,95 78,95 89,47 78,95 24 Eat 97,37 100,00 100,00 94,74 94,74 25 Like 80,26 94,74 84,21 78,95 63,16 26 Know 71,05 94,74 73,68 63,16 52,63 27 Keep 43,42 47,37 36,84 57,89 31,58 28 Close 64,47 84,21 78,95 36,84 57,89 29 Travel 26,32 36,84 31,58 21,05 15,79 30 Survive 61,84 68,42 68,42 57,89 52,63 31 Escape 35,53 42,11 42,11 26,32 31,58 32 Catch 30,26 42,11 36,84 15,79 26,32 33 Rich 44,74 42,11 52,63 47,37 36,84 34 Special 100,00 100,00 100,00 100,00 100,00 35 Full 73,68 57,89 63,16 78,95 94,74 36 Far 57,90 63,16 63,16 57,89 47,37 37 Wild 75,00 89,47 73,68 73,68 63,16 38 Good 100,00 100,00 100,00 100,00 100,00 39 Deep 94,74 100,00 94,74 94,74 89,47 40 Light 89,47 94,74 94,74 89,47 78,95 41 Little 88,16 89,47 94,74 84,21 84,21 42 Safe 68,42 63,16 78,95 84,21 47,37
104
43 Real 78,95 84,21 73,68 84,21 73,68 44 Possible 69,74 84,21 68,42 68,42 57,89 45 Difficult 63,16 63,16 57,89 68,42 63,16 46 Small 63,16 52,63 68,42 78,95 52,63 47 Bright 13,16 21,05 15,79 5,26 10,53 48 Happy 97,37 94,74 94,74 100,00 100,00 49 Amazing 71,05 84,21 84,21 78,95 36,84 50 Precious 10,53 15,79 5,26 15,79 5,26