Working Memory Training in
Relationship with Vocabulary
Development
Investigating the relationship between working memory training
and vocabulary development
Anna Logkaki S2000202
1
Contents
Contents ... 1
Acknowledgements ... 5
Abstract... 6
Introduction ... 7
Chapter 1: Vocabulary Development ... 10
1.1 Introduction ... 10
1.2 How vocabulary learning can be viewed through Behaviorism, Cognitivism and Constructivism ... 11
1.2.1 Vocabulary learning through Behaviorism ... 11
1.2.2 Vocabulary learning through Cognitivism ... 12
1.2.3 Vocabulary learning through Constructivism ... 12
1.3 Dynamic Systems Theory (DST) ... 14
1.4 Mental Lexicon ... 17
1.4.1 Levelt’s model of speech production ... 18
1.4.2 The Revised Hierarchical Model by Kroll and Stewart (1994) ... 19
1.4.3 The Distributed Feature Model of lexicon by De Groot (1992) ... 20
1.5 Dynamic view of the mental lexicon ... 22
1.6 Conclusion ... 23
Chapter 2: Working Memory ... 25
2.1 Introduction ... 25
2.2 Background ... 25
2.2.1 Information-processing approach-Atkinson and Shiffrin’s Multistore Model ... 25
2.2.2 Distinction between short-term memory (STM) and working memory (WM) ... 27
2.3 The Multicomponent Model of Working Memory ... 28
2.3.1 The phonological loop ... 28
2.3.2 The visuospatial sketchpad ... 29
2.3.3 The central executive ... 29
2.3.4 The episodic buffer ... 29
2.4 Working Memory capacity and language proficiency ... 30
2.5 Conclusion ... 31
2 3.1 Introduction ... 33 3.2 Historical Background ... 33 3.3 Previous studies ... 34 3.3.1 Keyword technique ... 34 3.3.2 Repetition... 35 3.3.3 Oral production ... 35
3.3.4 Word card strategy ... 36
3.4 Conclusion ... 36 3.5 Statement of purpose ... 37
Chapter 4: Method ... 38
4.1 Introduction ... 38 4.2 Participants ... 38 4.3 Materials ... 40 4.3.1 Notebook-printed papers ... 40 4.3.2 Paper cards ... 404.3.3 Categories of the words ... 41
4.3.4 Recordings ... 42
4.3.5 Vocabulary Tests and design ... 42
4.3.6 Working memory training sessions-website ... 44
4.4 Procedure ... 45
4.4.1 Teaching of the Greek alphabet and weekly assessments ... 45
4.4.2 Vocabulary learning procedure ... 46
4.4.3 Working Memory trainings ... 46
4.4.4 Time schedule ... 49
4.5 Description of the tasks of the Pre-test and the Post-test ... 51
4.5.1 Task 1 ... 52
4.5.2 Pre-test Task 2 ... 52
4.5.3 Post-test Task 2 ... 53
4.5.4 Task 3 ... 54
4.5.5 Rationale for the order followed when conducted each task ... 55
4.6 Description of the games in the working memory training sessions ... 56
3
4.7.2 Categorization of the data-variables... 62
4.7.3 Analyses for each one of the three categories ... 63
Chapter 5: Results and Discussion ... 66
5.1 Results ... 66
5.1.1 Working Memory Training Scores ... 66
5.1.2 Vocabulary Learning Scores ... 71
5.1.3 Pre-test and Post-Test Scores ... 73
5.2 Discussion ... 76
5.2.1 Working Memory Training Scores ... 76
5.2.2 Vocabulary Learning Scores ... 77
5.2.3 Pre-test and Post-Test Scores ... 77
Chapter 6: Conclusion ... 79
References ... 81
Appendix A: Pre-task and Post-task ... 87
A1) Pre and post-task 1 (Word recall task) ... 87
A2) Pre-task 2 (Vocabulary task- Greek language) ... 87
A3) Pre and post-task 3 (Word list free recall task) ... 88
A4) Post-task 2 (Vocabulary task-Kinyarwanda language) ... 90
B1) The Greek Alphabet ... 91
B2) Greek Words in Categories ... 92
B3) Greek Words in Lists to be learned (one list per week) ... 93
B4) Instructions-short description of the 4 categories of the games ... 94
Appendix C: Weekly Tests and Final Test ... 96
Test 1 ... 96
Test 2 ... 97
Test 3 ... 98
5
Acknowledgements
First of all, I am really grateful to all my friends who voluntarily participated in the experiment and they were really willing to help me. This paper is gratefully dedicated to all those that without their help this paper wouldn’t reach to its end. I really thank them for the time they spent and their patience.
I would like to thank Dr. Wander Lowie for his supervision, patience and support. Even when I was stressed and frustrated after our meetings he managed to make me feel better and inspired me to continue enthusiastically with this paper.
6
Abstract
7
Introduction
Read stated that words are “the basic building blocks of language” (2000:1). According to this statement it is understandable why there has been a considerable focus on vocabulary learning not only from the part of the learners but also from the part of the teachers. In the field of linguistics there have been many studies on vocabulary learning and on methods that it can be promoted more effectively. Numerous studies have been carried out not only on how learners acquire and develop new vocabulary but also on vocabulary strategies.
According to the principles of Dynamic Systems Theory (DST), we can argue that vocabulary learning can be viewed as a dynamic system. It is dynamic in its own right as it includes factors that should be taken into account as they can influence the whole learning process. Factors that can influence vocabulary learning can be learner’s characteristics such as a learner’s age, motivation, aptitude and many others. One component of aptitude is considered to be the Working Memory (WM). Baddeley stated that “Working memory involves the temporary storage and manipulation of information that is assumed to be necessary for a wide range of complex cognitive activities” (Baddeley, 2003: 189). As De Bot et al. stated “WM must be seen as an active system in which information is stored and manipulated and which is required for complex tasks like language comprehension” (De Bot et al., 2005:70).
8
influential models is the multiple-component model of working memory designed by Baddeley and Hitch (1974). According to this model, working memory has been identified by the three following components: the central executive component, the phonological loop and the visuo-spatial sketchpad.
We have used an updated version of the multiple-component model of working memory as it was revised by Baddeley (2000) to base our research for a better understanding of the relationship between working memory and vocabulary learning. Even though many studies have shown that working memory plays an important role in vocabulary development, no study has shown whether working memory can be trained. That was the main trigger for the present study to focus on the training of the working memory and the relationship it has with vocabulary learning. Specifically our research questions are the following:
1) Can Working Memory be trained?
2) If so, can Working Memory training affect vocabulary development?
In more detail, the present work consists of six chapters and deals with the following issues.
Chapter 1 begins with a reference on to how vocabulary development could be viewed throughout the framework of some basic theories of language acquisition. The chapter includes some of the basic principles of the Dynamic Systems Theory (DST), a theory that was used as a basic framework through the study that has been carried out. There is a short exploration on the different models of the mental lexicon and some issues concerning vocabulary teaching and learning are discussed.
9
there is a focus on the updated version of the Multicomponent Model of working memory as it was designed by Baddeley (2000).
Chapter 3 aims to combine vocabulary development and working memory capacity. This chapter provides an overview of some previous studies that have been carried out to reflect some important issues on the testing of working memory in connection to vocabulary development/L2 learning skills. It works as a link to the previous chapters whist the research questions and the hypotheses are also introduced.
10
Chapter 1: Vocabulary Development
1.1 Introduction
11
1.2 How vocabulary learning can be viewed through Behaviorism, Cognitivism and Constructivism
We chose to deal with three of the most general and influential theories to depict how vocabulary learning could be viewed through behaviorism, cognitivism and constructivism. Below we present in each separate section some of the main principles of each theory in order to present later our own view of vocabulary development.
1.2.1 Vocabulary learning through Behaviorism
Behaviorism was one of the theories in the field of psychology that was later developed as a theory in many fields and also as a leaning theory in linguistics. Some of the most famous originators of the behaviorism that significantly contributed a lot in this field were Ivan Pavlov, Burrhus Frederic Skinner, Edward LeeThorndike, and John Broadus Watson. One of the main principles of this theory is the principle of
“stimulus-response” which means that when a stimulus is presented to subjects and subjects respond to it, they produce “behavior”. According to the behaviorist theory, learning is a product of teaching and learners are passive recipients. Its main argument derives from the theory that learning a language is just a pure habit formation and that is why repetition is supposed to be one of the “golden” rules according to this theory. Thus, this theory has been criticized a lot as a mono-dimensional one since it failed to account for internal influences that deal with the learning process.
12
should have a more active role than being just a passive recipient of knowledge as the behaviorist model presents him to be. From a behaviorist point of view, vocabulary learning seems to be rather static without any sign of development.
1.2.2 Vocabulary learning through Cognitivism
John Watson, the founder of the behaviorism, argued that people learn different ways to behave and there is no such thing as memory. In contrast to this argument, cognitivism looks beyond the mono-dimensional theory of behaviorism. According to cognitive approach, memory plays an important role to the learning process as learning is facilitated by connecting previous knowledge for the new knowledge to take place. Cognitivists argue that the memory system is an active system, responsible for processing and encoding information that is linked to the learning process. Even though this theory states that learner can be viewed as an active processor of learning, it mostly focuses on the mental activities of the learner rather than on the learning process as such.
1.2.3 Vocabulary learning through Constructivism
13
things before integrating the new information. Constructivists view the learner as a unique individual with his own needs and argue that learning is an active and constructive process.
Additionally, Vygotsky contributed significantly in this theory by adding a social aspect in it. Constructivism, considers that learning is developing through interaction. According to Vygotsky’s theory, learning occurs in a zone which called “zone of proximal development” (ZPD). This term is defined as “the difference between the child’s developmental level as determined by independent problem solving and the higher level of potential development as determined through problem solving under adult guidance or in collaboration with more capable peers” (Vygotsky, 1978:85).
14
Figure 1.1: Representation of Behaviorism, Cognitivism and Constructivism as a
comparison of Atomic Theory Development to Learning Theory Development. (http://www.usask.ca/education/coursework/802papers/mergel/brenda.htm) The three theories are compared with the Atomic Theory Development. Specifically, Behaviorism which argued that the mind is a “black box” is compared with Dalton’s theory who argued that matter is made up of indivisible atoms. Later, Thomson found that in the matter they are electrons-that is compared with the cognitivism and its argument that learners go through mental processes. Nowadays, as the technology has developed we know that not only matter consists of positive and negative electrons but they interact with each other to construct the final form of matter (constructivism).
1.3 Dynamic Systems Theory (DST)
If we go back to all the theories that have dealt with the aspects of language we can state that none theory managed fully to depict all its aspects giving a complete picture of the language itself. Undoubtedly, there are many aspects to be examined such as the way the language is acquired, the way it changes and proceeds and the list goes on. Dynamic Systems Theory (DST) originally developed in the field of biology and then developed in many fields of science. It can also be applied to language development in the field of SLA. Most if not all of the researchers who have dealt with DST believe that language can be viewed as a dynamic system. However, a dynamic system constitutes different subsystems which are interconnected and affect each other making the development nonlinear. As Capra (1996) stated, a dynamic system is a system “whose essential properties arise from the relationships between its parts” (1996:27). Every system depends on internal and external resources, which interact. De Bot et al., (2005) stated that “the two main properties of dynamic systems are that all variables interact and that this continuous interaction keeps changing the system as a whole over time” (2005:14).
15
others. Because of the interaction, the system always changes but it can sometimes be rather stable; states of lesser variability are called “attractor states”. Moreover, there are many factors that we have to take into account when dealing with the developmental process such as the matter of causality. Every factor is influenced by another and even the issue of a learner’s motivation has to do with his attitude and has to be viewed through the scope of a dynamic system. Consequently, according to DST all changes in variability should be viewed as indicators of development.
Lowie and Caspi (2010) state that a dynamic system may be a chaotic but a self-organizing system. The authors, in order to capture this dynamic nature of a system drew an analogy of a flock of birds as a dynamic system. They wanted to show that birds individually nested in a greater system can influence it as they interact with each other over time by causing continuous changes inside the system. These interactions cause the different kind of shapes the flock of birds can take and the different kind of movements the birds can perform. Abraham (2003:299) summarized some of the basic principles of DST to link them to pedagogy.
1. Dynamical systems are self-organizing. 2. Dynamical systems are not linear.
3. Behavior is the product of multiple, contributing systems, each of which itself has a history.
4. Self-organization can be coupled with changes in the brain.
5. Dynamical systems can be understood only within the context of the larger whole (Capra, 1996; Kelso, 1995; Thelen & Smith, 1994).
16
and this can be accounted as a basis for L2 constructions. Usage-based approach focuses on the frequency of information and pays attention to meaning through lexical and grammatical patterns while Ellis (2002) also states that input frequency is one of the main factors to drive the L2 acquisition process.
From a DST perspective variability is seen as an intrinsic property of the developmental process. Variability is clearly defined through each individual learner’s performance. Variability depends on the different kinds of learning styles and can show us the interaction of the subsystems when acquisition takes place. Comparing to traditional SLD analyses which focus on the general picture of development, DST focuses on the developmental process itself. That means that the variability of the data, even when it is caused by internal or external factors, is considered as potentially valuable information about the stages of the process.
17
1.4 Mental Lexicon
The learners’ capacity of storage, use and memorizing is considered an important attribute of the L2 vocabulary learning. Takač, has stated that “L2 vocabulary development is also influenced by the organization of the mental lexicon” (2008:11). There has been much research on the various issues on the mental lexicon that deal with the organization and the development of it even though many conclusions are based on assumptions. Even nowadays the question whether there is a distinction between L1 and L2 mental lexicon as two separate systems is considered one of the complexities in dealing with the mental lexicon.
18
the mental lexicon below we are going to refer to some of the more influential models.
1.4.1 Levelt’s model of speech production
19
Figure 1.2: Levelt’s model of speech production (1993)
1.4.2 The Revised Hierarchical Model by Kroll and Stewart (1994)
After the first introduction of the Levelt model, Kroll and Stewart (1994) created the Revised Hierarchical Model (RHM) to explain the relation between the concepts in the L1 and in the L2 lexicon and to argue that the lexicons for different languages are essentially separated. Pavlenko (2009) stated that “The unique strength of the RHM is in capturing the developmental change in linking between the L2 and L1 word forms and lexical concepts” (2009: 143). However, as it can be seen from the Figure 1.3 below, the model deals closely only with the words’ concepts while as we are going to see later a word constitutes of many different features.
20
1.4.3 The Distributed Feature Model of lexicon by De Groot (1992)
21
Figure 1.4: The Distributed Feature Model (adapted from De Groot, 1992)
22
1.5 Dynamic view of the mental lexicon
Under the umbrella of the multilingual mental lexicon the words ‘association’ and ‘activation’ are considered key words. Even in L1 is really difficult to retrieve words if you don’t use them. De Bot et al.(2005) have stated that the words that are heard, seen or used most often are the words that are most easily accessed again and will have the most associations with other information such as how it is used” (De Bot et al.,2005: 49). In line with the DST, mental lexicon is viewed as a dynamic system. It incorporates words that are connected to speech production, concepts and meanings, transfer from L1 to L2. There are many factors that interact with each other and as mental lexicon in this view, is not static but changes over time it is understandable why some times it is difficult to retrieve words. Acquiring new words and ‘losing’ others is a part of this complex system that shows that mental lexicon is regarded as a network and is a dynamic system in its own right. However, each retrieval strengthens the connection between the form of the word and its meaning (Baddeley, 1990). Moreover, below we present what De Bot et al. have stated for terms of ‘elaboration’ and ‘rehearsal’ in connection to the terms of ‘association’ and ‘activation’ in order to comment on the dynamic nature of the mental lexicon:
“Elaboration, or the linking of new information to existing knowledge, makes perfect sense when the lexicon is regarded as a network in which all lexical items may be linked. Rehearsal will increase the level of activation of lexical items. Since the level will decrease over time, it is important not to allow the level of activation to drop beyond a threshold level. If a lexical item is reactivated before the activation level has dropped, the increase is more effective.”
23
1.6 Conclusion
Many years ago Stevick (1976) has shown that memory plays an important role in vocabulary learning and that words are stored in a network of associations. Nagy and Scott (2000) state that what is important in order to understand a learner’s vocabulary growth we need first to recognize five aspects of the complexity of word knowledge. These five aspects include a word’s incrementality, multidimensionality, polysemy, interrelatedness and heterogeneity. The analysis of these terms is out of the scope of this paper but a reference to these five aspects is considered important in order to give evidence to our support that vocabulary development is a dynamic system as each word has its own function and qualification and thus words can be seen as subsystems incorporated in a big dynamic system as vocabulary development.
All the above have been said in order to give our own interpretation for what vocabulary development accounts based on the DST principles. Vocabulary development is a dynamic system in its own right. The development concerns a process that cannot be viewed as static. It is understandable that when something is about to develop, and specifically a learner’s vocabulary development is seen as a process that changes over time. It is understandable that mental lexicon includes a vast number of words and human memory is an important component which gives access to the words. In terms of DST mental lexicon and human memory can be characterized as flexible. Memory contributes to the process of word recognition and Takač points out that both word recognition and production are considered processes that the mental lexicon offers as multiple accesses to information (Takač, 2008).
24
25
Chapter 2: Working Memory
2.1 Introduction
According to Dynamic Systems Theory (DST) language is seen as a system which changes all the time by including various subsystems which influence each other. By viewing a learner’s language development as a dynamic system we have to state that learners’ characteristics function as subsystems. Each one of these characteristics is viewed as a factor that affects and interacts with each other. Even though, in the literature of linguistics there is a very large number of factors that affect the process of second language learning, some of those refer to the learners’ age, attitude, motivation, aptitude, gender, learning style and strategies. One of the components of aptitude is working memory (WM). This section attempts to provide a better understanding on why memory plays an important role in vocabulary learning with a focus on working memory in particularly. We believe that working memory is a dynamic system and in order to base our argument we are going to look closer how it works according to Baddeley’s revised multicomponent model of Working Memory (2000).
2.2 Background
26
27
Figure 2.1: Atkinson-Shiffrin Multistore Model (1968). Information processing through various memory sub-systems. (From The psychology of learning and motivation”by R.C. Atkinson and R.M. Shiffrin, 1968. In J.S. Halonen, and J.W. Santrock, 1999, Psychology contexts and applications (3rd ed.). USA: The McGraw Hill Companies, Inc.).
28
2.3 The Multicomponent Model of Working Memory
Much research that has been carried out in the field of cognitive psychology has been based on the Multicomponent Model of Working Memory as it was designed by Baddeley and Hitch (1974). Baddeley and Hitch designed this model to explain how information processing is carried out by providing a description of both capacities and properties of the system of working memory. At first place the model constituted of three subsystems. The phonological loop, the visuospatial sketchpad and the central executive.
However, the tripartite working memory model raised various problems such as in explaining how the information of two subsidiary systems could be connected to each other and in which way could working memory be linked to long-term memory. That is why Baddeley (2000) added a fourth component in the Multi-component model which accounts for the episodic buffer and is supposed to support the capacity for conscious awareness (Baddeley & Andrade, 2000). More specifically each sub-system is shortly described below in order to get a better understanding of what are the properties of each sub-system and what is the relationship among them. Figure 2.2 illustrates the current version of the multicomponent model.
2.3.1 The phonological loop
29
have similar sounding names such as T,C,V,D,B,G. Moreover, the same happens with the recalling of series of similar words like cat, mat, map. Such series were correctly recalled on less than 20% of occasions while a score above 80% was found for the recalling of series of dissimilar words (Baddeley, 1966). Moreover, many studies have shown that phonological memory is important for vocabulary learning (Gathercole & Baddeley, 1998; Baddeley, Gathercole & Papagno, 1989) which gives support that phonological loop is an important system.
2.3.2 The visuospatial sketchpad
The visuospatial sketchpad is responsible for the storage and the maintenance of visual and spatial information. Studies have shown that there are different two systems, one which is responsible for the visual information and one for the spatial information (Hartley et al., 2001; Klauer & Zhao, 2004).
2.3.3 The central executive
The central executive’s role is to control the interaction between the phonological loop and the visuospatial sketchpad. Baddeley has characterized it, it is a type of “supervisor” or “scheduler” that is “capable of selecting strategies and integrating information from several different resources” (Baddeley, 1986).
2.3.4 The episodic buffer
30
Figure 2.2: The current version of the multi-component working memory model.
The episodic buffer is assumed to be capable of storing information in a multi-dimensional code. It thus provides a temporary interface between the slave systems (the phonological loop and the visuospatial sketchpad) and LTM. It is assumed to be controlled by the central executive, which is responsible for binding information from a number of sources into coherent episodes. Such episodes are assumed to be retrievable consciously. The buffer serves as a modeling space that is separate from LTM, but which forms an important stage in long-term episodic learning. Shaded areas represent “crystallized” cognitive systems capable of accumulating long-term knowledge, and unshaded areas represent “fluid” capacities (such as attention and temporary storage) themselves unchanged by learning. Baddeley's (2000) revised working memory model. LTM = long-term memory. From "The Episodic Buffer: A New Component of Working Memory?" by A.D. Baddeley, 2000, Trends in Cognitive Sciences, 4, p. 421. Copyright 2000 by Elsevier Science Ltd. Reprinted with permission.
2.4 Working Memory capacity and language proficiency
31
capacity. Higher levels of proficiency allow for larger holistic units to be stored in memory” (2010:23).
Based on the issue of the relationship between working memory capacity and language proficiency, Gass and Lee (2011) carried out a research to investigate whether L1 WM scores are predictive of L2 WM scores and whether this relationship depends on L2 proficiency. 110 American students who took courses in the Spanish language participated in the study. 48 of them were students from a first-year college course and 73 from a third-year grammar course. The learners participated in a computer-based session which involved a reading span task in their L1, a reading span task in their L2, a reverse digit span task and a task measuring inhibition. The results showed that the relationship between the L1 WM scores and the scores of the L2 WM for more proficient learners was strong in comparison to the scores of the less proficient learners which leads the researchers to conclude that the relationship between working memory scores and proficiency increases with greater proficiency.
2.5 Conclusion
32
learner has to encode the information through a series of cognitive processes. The multicomponent model can account for these processes. Our assumption is that if working memory accounts for complex activities such as vocabulary development that means that the training of working memory can facilitate vocabulary development
From what has been said above, according to Baddeley (2009) any kind of memory system whether it is physical, electronic or human requires three stages. The first concerns the capacity to either encode or enter information to the system, the second concerns the capacity to store this information while the third concerns the capacity to find and retrieve it. Consequently, it is understandable that even though each separate stage serves a different function they all interact. In line with DST, working memory is a dynamic system as it includes all these capacities as they have already been described in the Multi-component model. In addition to this, vocabulary learning constitutes a dynamic system in the sense that includes the process of linking the new information to the one that already exists, the process of encoding new information and store it in the learner’s mental lexicon. Furthermore, as a dynamic system which includes various subsystems, it is not static but it changes over time.
33
Chapter 3: Exploring the relationship between Vocabulary
Development and Working Memory
3.1 Introduction
In this study we focus on the training of the working memory and on the effect it has on vocabulary development. This chapter works as a connection to the previous chapters by showing how previous studies linked WM to learning process. We have already stated that vocabulary development and working memory can be seen as dynamic systems. In this chapter we will include our statement of purpose to introduce the experiment conducted and the results found.
3.2 Historical Background
34
fast we learn new information and on how fast we tend to forget it accordingly. This study showed us that even many years ago there was a tendency for the memory to be linked to learning.
3.3 Previous studies
3.3.1 Keyword technique
There were many researches and methods that tried to link memory to vocabulary learning. It was believed that some techniques linked to memory would be really helpful for learners to learn vocabulary more effectively. Back in the 1975, Atkison and Raugh used the keyword technique in their study. They gave their participants 120 Russian words with their English translation by providing them the keywords. According to the results of Atkison and Raugh’s study and to a more recent one of Taguchi (2006) keyword technique is proved to be really helpful in recalling new words.
35
3.3.2 Repetition
Repetition is considered to be very important in the process of vocabulary learning and as we saw it was one of the “golden” rules of behaviorism which nowadays is still supported. It is advisable that the learners should repeat the words they learn in a continuous basis in order to ensure that the words will be retained in their long-term memory and be part of the new lexicon they build. It is important to refer to one of the first studies which showed the importance of the repetition in language learning process. Anderson and Jordan (1928) in their study measured recall after learning in 3 stages. The percentages of material retained after one week was 66%, after three weeks was 48%, 39% and after eight weeks was 37%. This shows that for repetition is really important to occur the soonest possible after the new items have been learnt. 3.3.3 Oral production
Suzanne Graham (1997) in her book Effective Language Learning: Positive Strategies for Advanced Level Language Learning investigated some of the learning processes of students that started learning a language at an advanced level. In her book gives some practical recommendations for learning and teaching strategies. One of the strategies included is “Saying words out loud while learning”. The oral production of the new word plays an important role in the process of a learner’s vocabulary learning. Henning (1973) found that especially at the first stages of language learning, learners used to store words in their memory according to the sound of the words. According to this, we assume that the oral representation and repetition of the new words would be beneficial to learners’ memory.
36
stated that “acquiring vocabulary items involves linking a novel phonological and/or lexical sequence to its meaning” (1998:547) Consequently, that means that it is important for the learner to be able to associate the phonological or lexical form of the word with its meaning in order to understand its meaning. In this line we support the strategy of providing the learners recordings with the pronunciation of each word to be learned in order for the learning to be more effective.
3.3.4 Word card strategy
In SLA, long time ago research has shown that large quantities of vocabulary can be learned quickly and efficiently by learning words in lists of paired words that include words with either a synonym in the target language or their translation in the L1 or even pictures to illustrate their translation. (Crothers and Suppes 1967, Kellogg and Howe 1971, Nation 1983). The keyword method is still being used to strengthen the recall of the words to be learned.
3.4 Conclusion
37
consciousness as defined from the formers include both a cognitive and an affective stance toward words. It can be thus seen that there is a clear relationship between vocabulary development and memory.
3.5 Statement of purpose
Many studies so far have shown that working memory plays an important role in language acquisition but no study has shown whether WM can be trained. The present study deals closely with working memory training. An experiment was carried out to investigate whether WM can be trained and the effect it has on vocabulary development. 10 subjects took part in this study that lasted 5 weeks. All the participants had to learn 60 Greek words during this period. In order to measure the working memory training, 5 participants that constituted the Experimental Group had to take 20 training session during this period and also to follow the same procedure of learning the 60 Greek words as the participants of the Control Group. The Participants of the Experimental group had to play various games in order to train their working memory.
38
Chapter 4: Method
4.1 Introduction
In order to investigate whether Working Memory can be trained and what is the effect of the training to vocabulary learning, an experiment was conducted that lasted 5 weeks. Ten subjects took part in the experiment and their aim was to learn 60 Greek words during this period. The choice of the Greek language was made for three reasons. First of all, none of the participants had ever had any contact with the Greek language so that meant that there was no level of proficiency to take into account. Secondly, even though 60 words might not seem such a big number of words to be learned, the fact that the participants had to learn a new alphabet made the learning process more challenging for a small-scale study as this one. Thirdly, as I was both the teacher and the examiner, it was more convenient for me to choose my native language in case there was need to provide participants with more information about the structure of the Greek language. Two groups of 5 subjects each were formed in which the participants were assigned randomly: the Experimental (E) and the Control (C) Group. The distinction of the groups was made to separate the participants that constituted the Experimental Group who besides the learning of the new words, also completed some weekly working memory training sessions during the 5 weeks of the experiment.
4.2 Participants
39
Background section, some researchers conclude that the relationship between working memory scores and proficiency increases with growing proficiency. According to that all the participants’ level of proficiency in the English had to be taken into account so that all of the participants had succeeded either in the TOEFL or the IELTS exams. Their age ranged from 23 to 28 with an average of 25.5. The participants were selected according to their scores obtained from a pre-test. At a first place, the pre-test that is going to be analyzed later, was conducted to 20 potential participants. Because of a big range of different scores, a decision was made to exclude some people who scored either too high or too low to include participants who scored similarly. The experiment was conducted from April to May 2011. Below the Table 4.1 provides some information about each group’s participants concerning their age and their first language (L1).
Table 4.1: Participants’ profile
40
4.3 Materials
4.3.1 Notebook-printed papers
Initially, one printed paper of the Greek alphabet was given to them for the first two lessons that concerned the introduction to the new alphabet. The paper was structured in a way that included the letters of the Greek alphabet, the presentation of each letter’s pronunciation in Latin and in Greek letters and the Greek diphthongs. Additionally, each participant was provided with a notebook to be used for the Greek lessons. An example of a Greek lesson can be seen below when the vocabulary learning procedure is going to be described. They were instructed to use a different page every time a new set of words was introduced to them in order to help them connect each list with the position of the words in the notebook.
The words to be learned were 60 in total so 3 lists of 20 words each were created. The participants were assigned to learn one list per week. Each list was divided into two sub lists of words: Out of the 20 words, one sub list included a first set of ten words and the second sub list the other 10 words that remained. Every time that a new sub list was about to be introduced, the participants were provided with a printed paper of each sub list.
4.3.2 Paper cards
41
were used to connect the color of the cards to the words of each list. Specifically, yellow cards were used for words of the first list, green cards for the second list and blue cards for the third list. As each participant in the process of learning the new letters of the Greek alphabet adopted his own handwriting, the cards that were given to them were blank on purpose in order for the participants to practice and have their own handwriting on the cards.
The participants were suggested to use the cards weekly in order to facilitate them in their learning. The use of the cards was also presented to them as a mean of reassurance that they actually knew the words before they would be tested. Every time that they had to learn new words, the participants were suggested to mix the cards and by looking at the translation on each card to write down the Greek word. They were also advised to continue with this process till they were sure that they knew how to write all the words correctly.
4.3.3 Categories of the words
42
4.3.4 Recordings
We saw that it is really important for a learner in order to acquire vocabulary to link the new word with its phonological sequence (Atkins & Baddeley, 1998). That is why every time that a new set of words was introduced, the participants were provided with recordings of the words that I prepared as a native speaker of Greek. At first stage, when the participants had to learn the new alphabet a recording was given to them with the Greek alphabet and another one with the letters recorded in a random order which was used as a practice test of the Alphabet. In addition, before the participants were about to be tested in each list (20 words in total), a recording was given to them with the 20 words in random order. They were instructed after hearing one word to pause the recording and write down the word that they heard and at last to check their answers and to be sure that they knew how to write correctly each word. The same procedure was also followed for the learning of the new alphabet which was identical between the two groups.
4.3.5 Vocabulary Tests and design
43
The design of the three weekly tests was identical. Each test consisted of four exercises. The first exercise was a completion task. The participants were given a sentence in English and they had to write the Greek word that was missing. In this exercise the first letter of the Greek word was provided. The second exercise was a translation exercise. Each one of the sentences given included the English translation of the word to be tested underlined with the instruction to be translated in the Greek language. The third exercise was “a fill in the blanks” exercise. The English sentence explicitly implied the Greek word to be chosen. In the last exercise, some Greek adjectives were given and the participants had to write their antonyms.
The Final Test consisted of seven exercises. The number of words to be tested was convenient enough to allow us include various types of exercises and to give a different format from the one used for the weekly tests. More specifically, the first exercise was a matching exercise in which the participants had to match the Greek word given to its English definition. In the second exercise some pictures were given and the participants had to write the Greek word for what each drawing represented. In the third exercise the participants had to write the antonyms of six Greek adjectives. The fourth exercise had the same design as the completion exercises of the weekly tests with the only exemption that the sentences given concerned sayings, quotes or proverbs. The fifth exercise was a mixture of fill in the blanks and translation items. The sixth was a translation exercise that required only the translation of colors and the last one was a crossword.
44
include in the Final Test similar sentences that used in the weekly tests. In the Appendix C all the tests are included.
4.3.6 Working memory training sessions-website
45
4.4 Procedure
All the participants attended a five-week session which included the pre-test, an introduction to the Greek alphabet, the learning of 60 Greek words, the Working Memory trainings, the weekly tests, the final test and the post-test. The Table 4.2 gives a detailed schedule of the order and the time when these tasks had to be accomplished. Both as a first step and a pre-test, the 10 participants were tested in three different tasks (below is given the detailed description of each task). After the completion of the pre-test a shuffle was made to distinguish the participants in Group E (Experimental) and in Group C (Control). The participants in the Group E attended 4 training working memory sessions per week so that each participant had to complete the same working memory tasks (in total 20 sessions). All the participants were tested in the same Vocabulary tests and all of them had the post-test after they had been tested on the Final Test that included all the 60 Greek words to be learned. More details for the working memory training sessions of the Experimental group are given in section 4.4.3.
4.4.1 Teaching of the Greek alphabet and weekly assessments
46
ways of testing the 10 words. For the first assessment, the examiner pronounced the Greek word so the participants had to write down the Greek word followed by its translation and for the second assessment, the examiner pronounced the English translation so the participants had to write down the Greek word. In the third meeting the participants were assessed in the list that they had already learned.
4.4.2 Vocabulary learning procedure
Most of the times the participants were tested either in groups or individually because of the difficulty to meet the same schedule for all the participants. The introduction-teaching of the new words was done after their assessment. The examiner before writing on the blackboard the new word, she pronounced the new word and asked from the participants to try to write down the word as they could hear it. This was done in order to check their perception on the new alphabet. The goal was to get the participants into the process of thinking how the new word could be written according to the rules they had already learned from the Greek alphabet lessons. After that, the new word was introduced in the blackboard and the participants were asked to decide in which vowel the accent should be put. The examiner pronounced the word and the participants were instructed to pronounce the word in groups and then individually. The same process was followed for all the words during the teaching of the new vocabulary while a similar one for the lessons of the Greek alphabet.
4.4.3 Working Memory trainings
47
account to be compared later with the results of their vocabulary tests. The games included memory matrices, rhyme workout, face memory workout, spatial recall, memory match, face-name recall and many others. On average each participant played in total 25 different games, each one at least twice.
The internet site offered games from the five following categories: 1) Memory, 2) Speed, 3) Attention, 4) Flexibility, and 5) Problem-solving. The games that the participants played were taken from the first four categories as the Problem-solving category included games that had to do with comparisons and equations, dealing mostly with math problems. A meeting with each participant was scheduled in advance in order to be tested individually. In order for someone to get access to all the games of the lumosity site he has to be subscribed. As I was the examiner, I was enrolled creating my personal password so none of the participants could have access on the site if I wasn’t present. That was done in order for the participants to be controlled in the amount of games they played so that all of them had the same training. I personally, wrote down all the scores obtained for each participant in every training. The scores were put in an Excel file in order to be transferred later to the SPSS program.
48
Table 4.1: The use of the Memory games
Memory
Memory Matrix, Moneycomb
• Sense of 3-D environment • Remembering location of objects • Recalling visual patterns
Face Memory Workout, Familiar Faces, Name Tag
• Learning people's names for the first time • Remembering the name of someone you've met
Memory Lane, Memory Match, Monster Garden, Rhyme
Workout
• Learning and reasoning • IQ and intelligence
• Working with information in your head
Table 4.2: The use of the Speed games
Speed
Speed Match, Spatial Speed Match • Thinking faster
• Faster reaction time
• Speeding up cognitive processes
Rotation Matrix
• Sense of direction • Visualization • Reading maps
Table 4.3: The use of the Attention games
Attention
Birdwatching, Eagle Eye, Playing Koi, Observation Tower, Space
Junk • Peripheral vision • Sports • Driving safely Lost in Migration • Avoiding distraction
49
Table 4.4: The use of the Flexibility games
Flexibility
Brain Shift, Brain Shift Overdrive, Disconnection, Disillusion
• Multitasking
• Shifting your focus of attention • Cognitive control
Color Match
• Avoiding errors and mistakes • Resisting temptation
• Impulse control
Route to Sprout
• Thinking before acting
• Strategically planning the best approach • Flexibly solving new problems
4.4.4 Time schedule
50
Table 4.5: Time schedule during the study for both Experimental and Control Group
C Group (Control Group)
E Group (Experimental Group)
Week 1
Pre-test (3
tasks) 1) Word recall task(Pre-task 1) 1 ) Word recall task(Pre-task 1) 2) Vocabulary task(Pre-task 2) (Greek Language) 2) Vocabulary task(Pre-task 2) (Greek Language)
3) Word list free recall task(Pre-task 3)
3) Word list free recall task(Pre-task 3)
Design
Assignment of the participants into 2 Groups
Assignment of the participants into 2 Groups
E-Experimental Group and C-Control Group
E-Experimental Group and C-Control Group
Introduction
Learning the Greek alphabet (Lesson 1)
Learning the Greek alphabet (Lesson 1)
Learning the Greek alphabet (Lesson 2)
Learning the Greek alphabet (Lesson 2)
4 WM training sessions
Week 2
1) Learning of 10 new words (1st set)
1) Learning of 10 new words (1st set)
List 1
2) Testing on the 10 words and learning of 10 new words (2nd set)
2) Testing on the 10 words and learning of 10 new words (2nd set) 3) Testing of the 10 new words 3) Testing of the 10 new words
4) Testing of the 20 words and learning 10 new words
4) Testing of the 20 words and learning 10 new words
5) 4 WM training sessions
Week 3
1) Testing on the 10 words and learning of 10 new words (1st set)
1) Testing on the 10 words and learning of 10 new words (1st set) List 2 2) Testing of the 10 new words 2) Testing of the 10 new words
3) Testing of the 20 words and learning of 10 new words (2nd set)
3) Testing of the 20 words and learning of 10 new words (2nd set)
4) 4 WM training sessions
Week 4
1) Testing on the 10 words and learning of 10 new words (1st set)
1) Testing on the 10 words and learning of 10 new words (1st set) List 3 2) Testing of the 10 new words 2) Testing of the 10 new words
3) Testing of the 20 words and learning of 10 new words (2nd set)
3) Testing of the 20 words and learning of 10 new words (2nd set)
4) 4 WM training sessions
Week 5
Final Test Testing of the 60 Greek words Testing of the 60 Greek words Post-test (3
Tasks) 1) Word recall task(Post-task1) 1) Word recall task (Post-task1)
2) Vocabulary task(Post-task 2) (Kinyarwanda Language)
2) Vocabulary task(Post-task 2) (Kinyarwanda Language) 3) Word list free recall
task(Post-task 3)
3) Word list free recall task(Post-task 3)
51
Figure 4.1: A summative time schedule in a pyramid
4.5 Description of the tasks of the Pre-test and the Post-test
The pre-test and the post-test consisted of the same three tasks. The design and the procedure followed in each task are given below. The first task was a word recall task, the second was a vocabulary learning task and the third one was a wordlist free recall task. As both first and third task were the same in the pre-test and in the post test, we are going to analyze them by giving the names Task 1 and Task 3 accordingly. A more detailed description is going to be given for Task 2 that concerned the vocabulary learning task as it differed in the pre-test and the post-test. The order that was followed when the tasks were conducted was the same as the order of the description of the tasks that follows below. The rationale behind the specific order of the tasks is given in section 4.5.5.
W5:Final Test and Post test,
WM
W4: List 3, assessments, Test 3,
WM
W3: List 2, assessments, Test 2, WM
W2: List 1, assessments, Test 1, WM
52
4.5.1 Task 1
The first task was a word recall task. The participants were asked to carefully listen to a recording with a list of random words and repeat them back to the examiner in the same order. The words selected were short one or two-syllable words and they presented in CVC order. The lists were presented in ascending order, with two trials for each list length. The decision made for the two trials for each length was based on the piloting of the test that showed that three trials made the participants tired and not focused enough to reach a longer list. All participants were given the same lists, which increased in length from two to eight words. Each correct answer was assigned a score equal to the number of words in the list. However, for an answer to be considered correct, no mistakes should be made. In order for the task to be reliable, the words were recorded from a native speaker of English and each participant was tested individually.
4.5.2 Pre-test Task 2
53
adjectives, whilst the pairs were presented in a mixed order of “easy” and “difficult” words to be learnt.
The 20 Greek words were presented by the examiner to each participant individually followed by the oral pronunciation of the Greek word. After the oral presentation of the words, a list was given to each participant in which the Greek words were included in the same order they were presented with their translation in English. The estimated time needed for each word to be learnt was 5 seconds so a time of 1 minute and 40 seconds was given to each participant in order to memorize the 20 words. After the first minute passed, the examiner informed the participants that they had 40 seconds left in order for them to speed up with the learning and use wisely the time left.
As soon as the time needed passed, a paper was given to each participant with a list of all the Greek words that had been presented to him in a random order. The words appeared randomly and the participants were aware of it. This was done because the purpose of the test was not just to memorize the words in the order that was presented to them but try to “learn” them during the time given in order to measure the vocabulary learning as such. Each participant was asked to write down the translation of as many words as he could recall out of the 20. Each correct answer earned a score of 1 point for a total possible score of 20.
4.5.3 Post-test Task 2
54
present the Greek word with Greek letters. The other idea would be to choose directly 20 unknown Greek words presented by Greek letters but as I wanted the condition to be the same as in the pre-test I wanted to choose a new language to be tested.
Initially, the language that was chosen was Bulgarian but as one of the participants was from Serbia and he had some words in common in his native language, he got the total possible score that was 20 out 20. It was then considered that the test was not reliable so in terms of reliability of the scores to be obtained another language had to be chosen. Finally, Kinyarwanda language was chosen to be sure that there is no connection to any of the participants’ native languages. Again some easy words were chosen. Some examples of cognate words in the two languages: Kinyarwanda “repubulika”, English “republic”, Kinyarwanda “manda”, English “mandate”, Kinyarwanda “guverinoma”, English “government”, Kinyarwanda “minisitiri”, English “minister”. However, after piloting the test, the language was found to be more difficult than the Greek language in the pre-test so a total time of two minutes was given to the participants. After the first minute passed the examiner informed the participants that they one minute left.
4.5.4 Task 3
55
The participants’ were instructed to remember these words and then to write down as many words as they could recall no matter the order in which the words have been presented.
The program offers 15 lists, each one consisted of 15 words (see Appendix). The lists appear randomly and each participant was tested twice in order to gain a total score from the two lists in which he was tested. The third task was the same for the pre and the post-task. However, again in terms of reliability for the results the only thing that I needed to pay attention to for the post-test was that the participants had to be tested in different lists than the ones of the pre-test.
4.5.5 Rationale for the order followed when conducted each task
56
4.6 Description of the games in the working memory training sessions
As it has already been stated before, the participants had in total 20 Working Memory sessions. During these sessions they played in average 25 different games, each one for at least two times. This section aims to give a more clear idea of how the games looked like and as it is impossible to describe all the 25 games in particular we chose to have a closer look to four different games, each one of one of the categories to be tested. Below each game a picture is given with a screenshot taken from each game. 4.6.1 Memory Matrix
Memory Matrix is a memory game that exercises spatial recall and working memory. A pattern appears on the screen starting with a small number of tiles. The participants have to repeat the pattern by clicking on the correct tiles. The total number of trials that is given is 15. There are no levels in the game but we could say that each level represents the number of the tiles to be found. The first trial begins with a pattern of 9 tiles where 3 tiles appear simultaneously to be bold for 2 seconds and after the tiles disappear a message appears at the bottom of the pattern saying “Now click the tiles that appeared” so the players have to click on the right tiles. As soon as the player proceeds to a higher level, the pattern grows bigger. For instance, in level 4 a pattern of 12 tiles appear, in level 5 an equal one of 16 tiles and in level 6 one with 20 tiles. Depending on the difficulty of the level, more time is automatically given for the players to memorize the pattern.
57
correct it because as soon as they click on the wrong tile their attempt is cancelled so they have to play again the previous level in order to proceed into a more difficult one with more tiles to appear on the screen. For instance, if a player manages to reach the level 8 but cannot pass it, the next trial begins with the previous level that is level 7. As soon as the player manages to pass it for the second time the next trial will begin with 8 tiles. 10 points are given for each correct tile to be found plus some extra points as soon as the players manage to pass each level. Additionally, on the top of the game a bar appears including information about the number of tiles to be found, the number of the remained trials and the score obtained till this time.
58
4.6.2 Speed Match
Speed Match is a speed game that exercises information processing and working memory. The players have to remember each new symbol that it appears and to compare it with the symbol that appeared immediately before it. Do the symbols match or not? The players can respond either by clicking on the boxes that appear at the bottom of the game as it can be seen from the picture below or by using the keyboard by clicking on the right arrow if the symbols match or on the left arrow if they don’t match. As the game measures the accuracy of the correct answers in combination with the speed of the players, the participants were suggested to use the arrow keys of the keyboard as it was easier to use and more effective as to be as quick as possible in order to get the higher possible score. The players have to be quick as they have a limited period time of 45 seconds to get as many correct responses as possible. Additionally, the more correct responses they could get in a row, the more points each correct answer was worth it.
59
4.6.3 Lost in Migration
Lost in Migration is an attention game that exercises focus and response inhibition. A flock of 5 birds appear on the screen in a random place. The players have to focus and spot the direction of the center bird without letting the other birds to distract them. The players can respond either by clicking on the arrows that appear at the bottom of the game as it can be seen from the picture below or by using the arrows keys on the keyboard to define the direction of the center bird. The scoring of this game is same as in the Speed Match game. The players have a limited period time of 45 seconds to get as many correct responses as possible. 25 points are given for each correct answer. As soon as the players have many correct answers in a row combined with a quick speed of response he may take 50, 75 or even 100 points.
60
4.6.4 Color Match
Color Match is a flexibility game that exercises response inhibition and information processing. Two words appear on the screen in two boxes which are next to each other. The box on the right includes the meaning of the word that is a color while the box on the left includes a word. The players have to decide if the meaning of the word on the left matches the ink color of the word on the right without getting distracted from the meaning of the word on the right. The players can respond either by clicking on the boxes that appear at the bottom of the game as it can be seen from the picture below or by using the arrow keys on the keyboard in order to respond faster as they have a limited period time of 45 seconds to get as many correct responses as possible. An example from the instructions of the game is given in Picture 4.4 below to make it more explicit. For instance, this pair matches as the word on the right is written with red ink.
61
62
4.7 Analyses
4.7.1 Introduction
This section presents the analyses of our data while the next section provides a report of our results and then a discussion is followed that links the analysis of the results to the research questions. Before starting with the analyses and the choices that have been made on what analyses we chose to do, the research questions to be answered are presented below:
1) Can Working Memory be trained?
2) If so, can Working Memory training affect vocabulary development? 4.7.2 Categorization of the data-variables
All the data collected from the scores of the pre and post-tasks, the vocabulary tests and the scores from the working memory trainings were put into an Excel file and then transferred into SPSS program that was used for the analysis and the report of the results. Moreover, the data obtained from the experiment that was conducted can be distinguished into three different categories. Below a presentation of these categories is presented in order to give a better understanding of what were the variables we dealt with:
1) Working Memory Training Scores
• Time (The number of times the participants played each game-independent
variable)
• Memory Matrix score (Each individual’s score-dependent variable)
• Speed Match score (Each individual’s score-dependent variable)
63
2) Vocabulary Learning Scores • Test 1 score
• Test 2 score
• Test 3 score
• Final Test score
The above are the dependent variables which concern the individual score for both the Experimental and the Control Group, having as independent variable the participants of each group.
3) Pre-test and Post-Test Scores • Pre-task 1 score (Word recall task)
• Pre-task 2 score (Vocabulary task)
• Pre-task 3 score (Word list free recall task)
• Post-task 1 score (Word recall task)
• Post-task 2 score (Vocabulary task)
• Post-task 3 score (Word list free recall task)
The above are the dependent variables which concern the individual score for both the Experimental and the Control Group, having as independent variable the participants of each group.
4.7.3 Analyses for each one of the three categories 1) Working Memory Training Scores
64
mentioned in the method section, the participants of the Experimental Group had 4 working memory trainings per week, each one included an average of 3 to 5 games. In total they had 20 working memory trainings during the period that the experiment was conducted.
However, as each game measures different skills and also some games include levels it is understandable that the possible scores obtained from each game were not standardized. In order to draw a conclusion on whether each participant showed significant development in each game, some graphs were developed for each individual on four specific games, as a representative sample whilst various correlation analyses were done in order to investigate the strength of the relationship between the times the participants play each game and their individual scores in each one of them. We were interested in how the group performed as a whole in each game but we also analyzed each individual’s performance.
