Sign me in!

74  Download (0)

Full text


Master Thesis 2020-2022

MA Dutch as a second language and multilingualism University of Amsterdam

Name: Marieke van Brandwijk Student number: 11399678

Supervisor: Sible Andringa Second corrector: Roland Pfau

Sign me in!

How iconic and high idiosyncratic signs

support L2 vocabulary acquisition



When I was in my high school economics class, my teacher was trying to teach us about the concept of ‘leveling’ (‘nivelleren’ in Dutch). Because this was an unknown word to me, it turned into a vocabulary lesson, where the teacher tried to explain the meaning of ‘leveling’ in different ways. Not until he used his hands, held them flat with the palms facing down, one higher up than the other, and moved them through the air vertically to the same level, did I understand the meaning of this word: adjusting two components so they are at an equal level.

The two components both move toward each other, not just one to the other. This made me realize the added benefit signs can have when explaining concepts or novel words.

When teaching my students in my Dutch as a second language classes, I tend to

automatically use my hands in the same supporting manner as my economics teacher did with me, and I find it to be helpful in getting the meaning of a word across. The amount I use my hands to support my talking while teaching has increased as I was taking Sign Language of the Netherlands (Nederlandse Gebarentaal, NGT) courses. The signs I learned there, and the logical structures of those signs to represent words have proved useful in conveying word meaning in my classes.

Ever since those courses, signs have played a part in my life, both in educational settings as well as in personal settings when telling stories. Trying to find a topic for my master thesis was therefore not such a mammoth task as I first thought it to be. I sought it close to home, and figured I would like to know more about using signs in vocabulary education, and their added benefits.

After almost two years of both working on this thesis and leaving it be to pursue my ambition of becoming a Dutch as a second language teacher, this thesis has finally reached its conclusion. And, spoiler alert, it is an interesting one at that! I hope you will enjoying reading


this thesis much more than I enjoyed writing it, and most of all, if you are a fellow teacher in second language education, I hope you will start incorporating signs into your teaching methods.

A word of thanks

I would like to profusely thank the following people for helping me in many ways on this thesis. First of all I would like to thank my two supervisors. Nel, thank you for your patience with me not wanting to have any deadlines to try and avoid stress, but pushing me instead to try and make deadlineless timelines, which helped me a lot in keeping track of not only what still had to be done, but also of what I had achieved already. This helped my motivation a lot in the harder times. Sible, thank you for your Frisian down-to-earth nature which made me feel right at home at the master program (to quote you in my admission interview to the master: “of course you are admitted, how could you not, we went to the same high school!”), your everlasting encouragement to improve my thesis, and your indispensable explanations of generalized mixed linear models. Your idea to use cross situational word learning changed my whole thesis plan, and it has improved it 100% for the better. Tige dank!

In no particular order, here are the people I would like to thank for helping out with getting my thesis and my experiment together. Eveline, Nienke, and Lin An, for the many online study sessions which kept my head at my thesis instead of in the depths of YouTube, and who I could always come to for helpful advice, both linguistic and other. Rinse, my cousin, thanks to your skills with computers my videos had audio at the exact moment I wanted them to start, which would have taken me ages but you did it in an afternoon. Caroline Roset, for translating all my stimuli to Arabic, advising me on correct translations and getting me into contact with Nedaa. Nedaa, for taking the time to come and record the Arabic words, over and over again until the audio was good. Natalia Rivera Vera, for opening the lab up to me and helping me record the Arabic words Nedaa read out loud. Fleur and Cassandra, for recording the practice


trials in Arabic and English, after I had forgotten to make practice trials, and Fleur for lending me her white wall for making the corresponding videos. Marijke Scheffener, for not only igniting my enthusiasm for Dutch Sign Language and teaching me NGT for years, but also for letting me use her account for the online NGT dictionary to find the NGT signs for my

experiment. Dirk Vet, for making my vision for my online experiment come to life, changing it along with me changing my mind about it, and changing it again after testing it innumerous amounts of times. Your patience and skills have made this thesis to what it is now. Your help was indispensable. Everyone who helped me raise as many participants as possible, with special recognition for my mom and my aunt Corine, who both went to great lengths in asking their circle of acquaintances to participate. All of the participants to my experiment, whether they were able to finish the experiment or not, anonymous but highly appreciated. Andrew, for helping me find the right words and for your unlimited interest in anything I tell you. The schools Royal Kentalis and De Kameleon, for being so kind to let me interview

representatives to find out more about the use of signs in your education. Wendy Maxwell, for your enthusiastic answers to my questions about AIM. Thank you all for your time and effort.

Then I would like to thank all the people who were there for me mentally, again in no particular order. Hiddo, my brother from another mother, who I have known all my life, for always being in for a Harry Potter marathon, just like me. Jerney and Marlien, for always being on my side, providing me with both the necessary support and distraction. Nienke, Malou, Daniël, Sjoerd, Isabel, Fleur, and Martijn, also known as Eat Play Love, for all the unconditional love, boardgames and wonderful meals shared as well as the support and advice, whenever needed. With special recognition for Daniël and his Timon-hat for giving me enough joy to last me the rest of the year. Els, for the lockdown walks in which I could vent about my thesis and the support and advice you gave me. Snorrie, my cat, for your additions to my thesis by walking across my keyboard, even though your contributions did


not make it into the final thesis, as well as the necessary play breaks which did wonders for both our mental states. Both my and Bart’s family for all the support, distraction and Carina for the substantive contributions. My colleagues at the Academic Language Center and Dutch Studies for the amazing times teaching and learning about second language education, fueling my passion for it and making me feel so welcome and appreciated. Liesbet, thank you for all your guidance, your fun loving attitude to life, and your endless creativity which brought mine up too. To my apes Nienke and Eveline, thank you for being there whenever, for never failing to make me laugh and for being a shoulder to cry on when needed. Without you, life would be rather dull.

Lastly I would like to thank Bart, the love of my life, my biggest supporter and cheerleader (literally), for always being there for me, always lending me an ear, giving me kind words, wisdom and many ideas. Thank you for testing my experiment again and again, even twice at the same time in different tabs. Thank you for your support when I couldn’t see the forest for the trees, and for your everlasting and unconditional love. Words cannot adequately describe my love and appreciation for you.


Table of contents






2.3.1 Iconic, abstract, and high idiosyncratic signs ... 14


2.5HYPOTHESES ... 18

3. METHOD ... 20


3.2MATERIALS ... 20

3.2.1 Word selection ... 20

3.2.2 Audio ... 24

3.2.3 Video ... 24

3.2.4 Background questionnaire ... 24

3.2.5 Debriefing questionnaire ... 25


3.3.1 Trials ... 26

3.3.2 Vocabulary tests ... 27



4. RESULTS ... 30

4.1MAIN RESULTS ... 30

4.2AGE ... 32


4.4WORD CLASS ... 34




5.1.1 Discussion on sub results ... 39





5.5.1 To give the correct translation or not to give the correct translation, that is the question ... 41

5.5.2 Complete L2 vocabulary acquisition ... 42

5.5.3 Age ... 42

5.5.4 Word class ... 42

5.5.5 Pepper ... 43

5.5.6 Linguistic aptitude ... 44






Practice trials ... 49

Test trials ... 51








1. Introduction

Due to my own affinity with teaching Dutch as a second language (DL2), I am always looking to improve my classes and help students in their DL2 learning ambitions. Of the different aspects of learning a new language, vocabulary learning can be considered as one of the most extensive and challenging parts. The Common European Framework of Reference for Languages: Learning, teaching, assessment (CEFR) (Council of Europe, 2001) provides us with numbers to truly see the vastness of the vocabulary learning task for an L2 learner. To reach level A1 it requires the knowledge of 1000 lemmas, A2 requires 2000 lemmas, B1 (which is the level of DL2 state examination program I) requires 5000 lemmas and B2 (which is the level of DL2 state examination program II) requires 12.000 lemmas. Much research has been conducted that provides proof for a relationship between vocabulary size and language skills (a.o. Rogde, Melby-Lervåg and Lervåg, 2016; Malatesha Joshi, 2005). Evidently, learning second language (L2) vocabulary is an important and enormous task for the learner and it is therefore important we research new ways in which we can facilitate vocabulary learning.

One way in which L2 vocabulary learning can be improved is by seeing signs during learning. This technique is already used in certain schools, and has proven to be fruitful for the students (Maxwell, 2001; Rousse-Malpat, Verspoor & Visser, 2012; Clark and

Trofimovich, 2016).

In this master thesis I will report on our research, in which we have looked at how different types of signs can aid in L2 vocabulary retention. So far, there have been quite some studies that show that signs can aid L2 learners during the learning of novel L2 vocabulary in both adults and children (Tellier, 2008; Morett, 2014; Macedonia, Müller & Friederici, 2011;

Macedonia & Knösche, 2011 a.o.), but there is still much we do not know about this subject.


We decided to delve deeper into the different types of signs a sign language has, and whether there is a difference in the aid each type of sign can provide during the L2 vocabulary learning process. We looked specifically at the process during the learning phase, rather than only looking at the learning outcome.

In the next chapter we will go deeper into the topic of using signs in second language education, specifically vocabulary education, and we will go into the different types of signs a sign language has, and what we already know about their potential support in L2 vocabulary learning.


2. Theoretical background

2.1 Signs in the classroom

Incorporating signs in the classroom is not a novel idea; it is used already in some schools.

Royal Kentalis is an organization that aids people that are hard of hearing, deaf, deafblind, have a language development disorder, or have profound and multiple learning disabilities that impair communication. In their schools, signs are not only used to communicate with children with hearing impairments during their classes: in the classes for children that have a language development disorder, signs of a Dutch sign system (NmG: Nederlands met

Gebaren) are used as well to support communication. Though it is not mandatory for the staff to use signs, it is seen as a huge pro if they do, and it is a good example of how signs are already used in classrooms to support communication.

People with a non-standard IQ-score can benefit from using signs in vocabulary education as well. At KPO (V)SO-school De Kameleon in Roosendaal, the Netherlands, they teach students aged 4-18 with an IQ-score below 70 and/or with additional problems. In this school, teachers are encouraged to use signs from NmG when teaching vocabulary. The speech therapist in this school is a fan of using NmG, for she sees an improvement of vocabulary retention and recollection in the students, even in students that do not talk. However, when signs are used with someone with autism, they found that using a sign is often too much for them to process at the same time as the novel word. Teachers are free in choosing to use NmG, but when they do use it, it comes with an additional benefit for the teacher: it helps to keep their communication to the point and clear, which is important, especially for their target audience.

There are numerous other examples of teachers using signs to support the subjects they are trying to convey in the classroom. Smotrova and Lantolf (2013) pointed out that signs are an


important part of comprehensible input and that teachers use them naturally. Many studies have already shown the many advantages signs can have when used in the classroom

(Valenzo, Alibali & Klatzky, 2003; Barnett, 1983; Alibali & DiRusso, 1999; Maxwell, 2001 o,a,). Barnett (1983), for instance, showed that the use of signs in a classroom can increase recital speed of the students, reduce the time the teacher is talking instead of the students, can support the teacher’s instruction, and can decrease shame in students and increase their bond with each other. Many teachers already use signs, for instance pointing and/or nodding to point out that a student’s answer is correct, or trying to mime a new word. The signs allow both teachers and students to work more directly in the L2: they talk in the L2, not just about the L2. Signs also give the teacher more control over the classroom, atmosphere, and student activity. Valenzeno, Alibali and Klatzky (2003) researched children learning about symmetry and asymmetry. When half of the children participating in their experiment were taught about these concepts with the teacher using signs, and half were taught without signs, the children who had been taught about symmetry with signs scored better on the posttest than children who had been taught without signs. Alibali and DiRusso (1999) found that children that produced signs during learning mathematics had higher grades than children who did not sign to themselves. Clark and Trofimovich (2016) found the use of signs in L2 vocabulary

teaching by both students and teachers to be a positive stimulant for the students. Signs had a positive effect on participation of students and the amount of interaction during the class.

Students indicated liking the sign activities, and they said that the signs helped them remember the words if they were logically linked to the words or gave hints towards word length or pronunciation

Teaching methods based around incorporating signs in the classroom have already been made and are successful. In the 90s, Wendy Maxwell started to create the Accelerative Integrated Method (AIM). Originally made to facilitate learning French as a second language


in Canada, AIM uses a fixed sign paired with each uttered French word to help students in acquiring the vocabulary, as well as grammar and syntax. Multiple researches have been conducted on the effectiveness of AIM, and the results are positive: students who used AIM are more fluent and are better at engaging in sustained speech (Maxwell, 2001) and writing (Rousse-Malpat, Verspoor & Visser, 2012) than comparable students who used a traditional method to learn French. Rousse-Malpat, Verspoor and Visser (2012) also found a benefit in vocabulary acquisition: students taught with AIM retained more words than students taught with a regular teaching method. However, AIM is designed to offer less words in one go than regular methods, which gives students more chance to come across these words in different contexts, so this benefit cannot be acquitted to signs with certainty without further research.

2.2 Signs and L2 vocabulary learning

Research into combining visual stimuli with L2 vocabulary learning has become

increasingly popular, and shows that combining visual stimuli with L2 vocabulary learning tends to aid in L2 vocabulary retention (Chun & Plass, 1996; Tight, 2010 a.o.). For example, both Chun and Plass (1996) and Tight (2010) conducted experiments in which they combined words with a visual representative of the word (a video, picture or a physical object). In the studies they conducted they both found that learning a novel L2 word while being

concurrently presented with a visual stimulus related to the word led to a better retention of said word compared to not being presented with a visual stimulus. Furthermore, Tight (2010) found that teaching novel L2 vocabulary in a multi modal fashion was more effective than adjusting the teaching method to individual learning style preference.

Interestingly, improved L2 vocabulary retention is not only achieved when viewing a video, picture or a physical object, but also when letting participants view signs that match the words they are learning. (Kelly, McDevitt & Esch, 2009; Goldin-Meadow, Kim and Singer, 1999; Morett, 2014) For example, Kelly, McDevitt and Esch (2009) conducted an experiment


where participants were either shown a sign while hearing a novel L2 word, or would only hear the novel L2 word. Afterwards they conducted a memory test which showed that participants who had learned novel L2 words while seeing a matching sign retained these words significantly better than the participants who had not seen a sign in the learning phase.

Using a different method, Morett (2014), also proved that signs can aid in L2 vocabulary learning. She used a method where participants were either ‘explainers’ (they heard an L2 word and saw it in writing, either with or without a matching sign) or ‘learners’. Explainers had to teach learners the L2 words. This worked significantly better when the explainers saw a matching sign.

Experiments looking at the production of signs show that production also facilitates L2 vocabulary retention (Macedonia and Knösche, 2011; Mayer et al., 2015; Tellier, 2008;

Macedonia and Klimesh, 2014). Macedonia and Knösche (2011), for example, researched L2 vocabulary retention by making participants either repeat the novel L2 word and draw a matching image, or by making them repeat the novel L2 word and sign a matching sign. They found a significant difference in L2 vocabulary retention in favor of reproducing a matching sign. They also found that words learned while seeing a sign are used more often than words learned with only audio-visual stimuli when participants had to create novel sentences with the newly learned words. They concluded that the words learned with signs are more accessible in the brain. Mayer et al. (2015) found similar results when they conducted the same experiment, but added a control condition where participants only repeated the word. In a second version of their experiment, Mayer et al. (2015) changed the drawing a matching image condition to viewing a matching image. This experiment is similar to that of Tellier (2008), who had French children learn English words by either viewing a matching picture or producing a matching sign. Both experiments had the same outcome: producing a sign

benefited L2 vocabulary retention more than viewing a matching image. Similar results were


found by Macedonia and Klimesch (2014), when they conducted an experiment where participants either repeated the novel L2 word or repeated the word and produced a matching sign: producing a sign aided retention significantly compared to not producing a sign.

2.3 Usefulness of different kinds of signs

Not all signs are equally useful when learning novel L2 vocabulary. Janzen Ulbricht (2008) found that for children that had a lower speech rate in the L2 at the onset of the experiment, a sign linked to a morpheme was more useful than a sign linked to a sentence. For children who had a high rate of speech in the L2 at the onset of the experiment, the opposite was true.

Furthermore, signs do not help with every aspect of language learning. Kelly et al. (2014) found that signs do not aid comprehension at a phonetic level; signs used to highlight phonetic differences did not aid the participants in acquiring these phonetic differences. Finally, the nature of the relationship between the word and the corresponding sign affects its usefulness in the acquisition process as well.

2.3.1 Iconic, abstract, and high idiosyncratic signs

A distinction can be made between iconic and abstract signs. Iconic signs are signs where

‘aspects of its form are directly related to what is represented.’ (Bellugi & Klima, 1976). An example of an iconic sign is ‘to cut’ (see Appendix A for a photo of the sign and its

movement). Wherever you are in the world, if you are looking for scissors, or if you want to cut something, this would be this sign you would use to get that message across and it would be understood by someone not speaking your language, who knows what scissors are.

Because of this imminent understanding of the meaning of the signs, without needing any words to support it, this is an iconic sign.

Abstract signs on the other hand do not have a relationship between their meaning and their form; their meaning cannot be subtracted solely from their form. An example of an abstract sign is ‘doctor’ (see figure 1 below for a photo of the sign and its movement). This


movement is completely unrelated to its meaning. Even when you would be told what the sign means, there is no reason for why the form of the sign would come to that meaning. There is no mnemonic for it.

Figure 1

The sign for ‘doctor’ in NGT (“Online Gebarenwoordenboek”, n.d.). The movement of the tips of the fingers hitting the chin is repeated twice.

Iconic signs provide aid in novel L2 vocabulary retention (a.o. Macedonia, Müller and Friederici, 2011; Huang, Kim & Christianson, 2019). Macedonia, Müller and Friederici (2011), for instance, found that words presented with signs that have an iconic relationship with said word are better retained than words presented with abstract signs. This was also shown in an fMRI experiment they conducted afterwards: only when the motoric image that the sign creates is compatible with the internal representation of the semantics of the concept, will the sign cause better memory performance.

Besides iconic and abstract signs, there is a third type of sign called high idiosyncratic signs. An example of a high idiosyncratic sign is ‘memory’ (see Appendix A for a photo of


the sign and its movement). Upon seeing this sign, there are many possible meanings you could attribute to it. You could think it referred to the length of the hair of the speaker, or to the face in general. However, when you are told the meaning of the sign is ‘memory’, you could form a form-meaning relationship yourself, where the movement of the hand

symbolizes the anchoring of the memory in the mind. This form-meaning relationship is not evident before you know what the sign means, but the sign can become a mnemonic to its meaning after the meaning is revealed to you. This ‘second instance’ type of sign, where the meaning makes sense after learning of it, not before, is called a high idiosyncratic sign.

High idiosyncratic signs, like iconic signs, can provide aid in retention of novel L2 vocabulary as well. Huang, Kim and Christianson (2019) conducted an experiment to

establish whether presenting the L2 learner with a word combined with a sign will give the L2 learner an advantage in L2 lexical recognition compared to presenting the word without a sign. They had two types of signs: low idiosyncratic signs (signs which most people will correctly iconically link to their meaning, we would call these signs iconic) and high

idiosyncratic signs. In total they had three conditions for their experiment: participants were orally presented with either a novel L2 word with a low idiosyncratic sign, a novel L2 word with a high idiosyncratic sign, or just the novel L2 word without any sign. They used a within-participants and within-items design and had 30 native speakers of Midwestern

American English who had no prior experience with Mandarin Chinese. All participants were given one example in each condition, and were told they would have to take a multiple-choice vocabulary test at the end of the experiment. The experimenter, a native Mandarin Chinese speaker, presented 18 new words. When the experimenter would present a participant with a word, she would tell them to first watch and listen, and then repeat the Chinese word and the English meaning twice. They did not ask the participants to copy the signs and none of the participants did this on their own. The list of words was presented twice to each participant.


The results of the vocabulary test showed that the low idiosyncratic sign condition (the iconic signs) and the high idiosyncratic sign condition did not differ significantly from one another.

Both conditions did significantly differ from the no signs condition; the experimenters found an 8% to 10% advantage in L2 lexical recognition. However, this advantage faded when participants were presented with more words after 10 novel words. These findings tell us that a sign does not need to be iconic in order for it to aid in word recognition, as long as learners can still add meaning idiosyncratically, and it is not obviously iconic with another known word. Additionally, it seems that presenting novel words accompanied by signs in big groups does not aid to their retention.

2.4 The current study

The interesting outcome of the experiment conducted by Huang, Kim and Christianson (2019), namely that both iconic and high idiosyncratic signs shown concurrently with hearing a novel L2 word aid in L2 vocabulary retention equally, sparked our interest. The different signs aid towards a similar outcome in vocabulary retention, but do the learners have different paths to get to the same outcome? In order to get some insight into the learning process of participants learning novel L2 words aided by seeing iconic and high idiosyncratic signs, we have conducted an experiment using a cross situational word learning paradigm, in which we presented participants with Dutch as their mother tongue with 16 Arabic words and

concurrent signs, either iconic or high idiosyncratic.

Cross situational word learning was first performed by Yu & Smith (2007) who found that learners “calculate cross-trial statistics with sufficient fidelity and by doing so rapidly learn word-referent pairs even in highly ambiguous learning contexts”. This type of experiment allows us to gather data not only on the outcome of the word learning experiment, but on the learning process as well. Data is gathered in multiple learning blocks, which then can be analyzed using a generalized mixed linear models analysis.


Key to a cross situational word learning task is that participants do not receive feedback on their answers. The way they learn the words is by a process of elimination; if Arabic word A is presented with possible Dutch translations 1 and 2, and in the next block Arabic word A is presented with possible Dutch translations 2 and 3, then it can be deducted that the correct Dutch translation is 2. The signs are used in an assisting way; they can help participants decipher what the correct Dutch translation might be. After some testing we found out that participants relied too heavily on the signs, and were not connecting the Arabic words to the Dutch translations. Instead they were connecting the signs to the Dutch translations, mostly ignoring the Arabic spoken word. To combat this we introduced a vocabulary test in between each block in which participants were asked to link the Arabic spoken word to its Dutch translation without seeing a sign. We received oral feedback from participants that this helped them focus on the task of learning the spoken Arabic words.

2.5 Hypotheses

We want to subject our participants to a cross situational word learning task containing four learning blocks, in which they will learn 16 words in an L2, 8 of which are supported by seeing an iconic sign, and the other 8 are supported by a high idiosyncratic sign. The signs represent the meaning of the word. We hypothesize that, even though the learning outcomes are equal in Huang, Kim and Christianson’s experiment (2019), iconic signs help with the retention of the L2 words more than the high idiosyncratic signs during the learning process.

We hypothesize this because we think having to create a form-meaning relationship idiosyncratically between the sign and the word takes the focus and energy away from

learning the L2 word. This would support the resource depletion for output hypothesis (RDO) introduced by Barcroft (2006). The RDO suggests that by exhausting cognitive processing resources, those resources cannot be used to retain a word. Barcroft (2006) showed this in an experiment where participants had to write down the word they were learning. Writing turned


out to impair the learning process compared to learning the word without writing it down, and it was hypothesized that the process of writing created a cognitive overload: it occupied processors otherwise used to retain a word. In the iconic condition of our experiment, the form-meaning relationship between the sign and the Dutch translation is already formed, which leaves the participant with only two tasks: creating a form-meaning relationship between the Dutch translation and the Arabic word, and between the sign and the Arabic word. In the high idiosyncratic condition, the participant has another task on top of the previously mentioned tasks: creating an idiosyncratic link between the sign and the Dutch translation. The RDO suggests that this extra task would take impair the L2 vocabulary acquisition process, which is why our hypothesis is that the iconic signs will aid the acquisition of the novel L2 words more than the high idiosyncratic signs.

We also hypothesize that it will take longer to finish retaining a novel L2 word when said word is supported by a high idiosyncratic sign than when it is supported by an iconic sign for the same reason as stated above.


3. Method

3.1 Participants

All 53 participants were native Dutch speakers, with Dutch being their only mother tongue.

They did not have sight- or hearing problems, or if they did they corrected them to a standard that could be considered ‘regular’. They did not have any language disorders such as dyslexia.

They did not have any prior language proficiency in Arabic and neither in any sign language.

The participants were recruited through social media posts, as well as through word of mouth.

All ages above 18 were welcome, as were all social and educational backgrounds.

Participants were informed about the proceedings of the experiment (the information brochure can be found in Appendix E) and were asked permission to use their data

anonymously (permission brochure can be found in Appendix F). After reading and accepting the use of their data, the experiment started, but they were allowed to quit the experiment at any given time without stating a reason. The ethical commission of the University of

Amsterdam agreed to the anonymous gathering of the information that we did, which we have only used for this experiment.

3.2 Materials

3.2.1 Word selection

The sixteen words participants had to learn were first selected based on the sign in NGT (Nederlandse Gebarentaal, Dutch Sign Language). We decided to use signs from an already existing sign language because it gives us a standard, there is no variation in the hand shape or movement, and the signs already exist and are therefore easily retrievable. An added benefit, albeit not the purpose of the research, is that the participant learns some NGT. The signs were selected as potential candidates for iconic or high idiosyncratic signs, as judged by an NGT learner. To be sure signs were high idiosyncratic, native speakers of Dutch who did not have


any language proficiency in NGT were asked what they thought the sign meant. If they answered correctly immediately, the sign would be judged as being iconic. If they didn’t know what it meant, or weren’t sure, they would be told the meaning of the sign. If they experienced a moment of realization and understanding once the meaning of the sign had been revealed, it would be judged as being high idiosyncratic. If they still did not see the relationship between the sign and its meaning, the sign would be judged as being abstract, and would not be used as a stimulus in the current experiment.

The Dutch translations of the high idiosyncratic and the iconic signs were then searched in the CELEX lexical database (Baayen, Piepenbrock & Gulikers, 1995) to uncover their

frequency per million Dutch words. From the pool of words, eight nouns and eight verbs were selected, which were paired for frequency. The amount of words in the same frequency range was the same in the iconic and the high idiosyncratic words. The four words used as practice trials and the sixteen words used in the testing phase can be seen in table 1 and 2 respectively.

Table 1

Overview of words used in the practice trials

Dutch word

English translation

Arabic spoken translation

Word class

Orthog raphic count

Syllable count

Freque ncy per million


Wissen To erase n.a. (this practice trial was in English)

Verb 6 2 7 High


Knippe n

To cut n.a. (this practice

Verb 7 2 26 Iconic


trial was in English)

Naam Name مسا Noun 4 1 420 High


Regen Rain رطم Noun 5 2 55 Iconic

Table 2

Overview of words used in the testing phase

Dutch word English translation

Arabic spoken translat ion

Word class

Ortho graphi c count

Sylla ble count

Frequen cy per million


Bijlage Attachemen t (in an e- mail)

قحلملا Noun 7 3 5 High


Overdrijven To


ةغلابملا Verb 11 4 6 High

idiosyncratic Advertentie Advertisem


ةياعدلا Noun 11 4 17 High


Schoppen To kick لكرلا Verb 8 2 23 High


Herinnering Memory ةركاذلا Noun 11 4 95 High



Onderzoeke n

To research ثحبلا Verb 11 4 95 High


Plaats Place ناكملا Noun 6 1 661 High


Vertellen To tell رابخلإا Verb 9 3 514 High


Armband Bracelet راوسلا Noun 7 2 5 Iconic

Knuffelen To hug قناعتلا Verb 9 3 5 Iconic

Peper Pepper لفلفلا Noun 5 2 16 Iconic

Roeren To stir (with a spoon)

كيرحتلا Verb 6 2 24 Iconic

Vogel Bird ريطلا Noun 5 2 96 Iconic

Bellen To call (on the phone)

لاصتلاا Verb 6 2 101 Iconic

Huis House لزنملا Noun 4 1 630 Iconic

Schrijven To write ةباتكلا Verb 9 2 424 Iconic

In Appendix A, an overview can be found of the signs which accompanied these words.

The words were translated by dr C. J. Roset of the University of Amsterdam, a lecturer in the department of Arabic language and culture. Seeing as Arabic does not have an infinite verb form, she proposed to translate the verbs to verb substantives, i.e. ‘the writing’ instead of ‘to write’, which we did.


3.2.2 Audio

In each trial, participants heard the novel L2 word spoken in either English (in two of the practice trials) or in Arabic (in the rest of the practice trials and all of the test trials). The Arabic audio in the test trials was recorded by a native speaker, as was the English audio in the practice trials. The Arabic audio in the practice trials was recorded by a native Dutch speaker but learner of Arabic, due to practicality issues. Because the practice trials were only meant to familiarize participants with the task, we did not conceive this as a problem.

3.2.3 Video

In each trial, participants saw either an iconic or a high idiosyncratic sign.The signs were produced by the researcher, who is a non-native NGT signer, who is a learner of NGT and has mid-range proficiency, which we deemed enough to sign the signs for this task. Even though the videos in the practice trials were not shot on the same day as the videos in the test trials, we had the intention to keep a consistency to them by shooting them in the same place. This consistency was eventually not completely achieved, for I had not considered the experiment when deciding to cut my hair off to donate it to charity. We do not think the change in hair length from the practice trials to the test trials has affected the participants’ performance in any way.

3.2.4 Background questionnaire

Participants were asked nine questions to gather relevant information about the participant and to confirm that they were suitable candidates for the experiment. While recruiting

participants, the requirements to have ‘normal’ sight and hearing, no dyslexia or any other type of language disorder and to not have any language proficiency in Arabic or any sign language were already communicated. We decided to still ask these questions in the

questionnaire to check that all participants had read these requirements and had taken them seriously. The questions asked were (translated here from Dutch to English):


1. How old are you?

2. What is your sex (answer options available: male, female, I would rather not say, other)

3. What is your highest received education (answer options available: high school, MBO, HBO, WO)

4. What is your mother tongue?

5. Do you have any language proficiency in Arabic? (answer options available: yes, no) 6. Do you have any language proficiency in a sign language? (answer options available:

yes, no)

7. Do you have an eye disorder that you are not correcting with glasses or contact lenses?

(answer options available: yes, no)

8. Are you dyslexic or do you have another language disorder? (answer options available: yes, no)

9. Do you have hearing problems you don’t correct? (answer options available: yes, no) Regardless of the answers given, participants were always allowed to do the rest of the experiment. One participant’s data was excluded from the statistical analyses because they had filled in Frisian as their mother tongue. Two participants indicated already knowing one of the Arabic words, so these words from these participants only were excluded from the statistical analyses.

3.2.5 Debriefing questionnaire

After the final vocabulary test, participants were asked six follow-up questions. They served to give insight in how participants used the signs and whether they had correctly executed the experiment. The questions were (translated from Dutch to English)

1. Did you already know one or more of the Arabic words used in the experiment?

(answer options available: yes (with text box to fill in Arabic words they already knew), no)


2. Did you already know one or more of the signs used in the experiment? If yes, which?

(answer options available: yes (with text box to fill in signs they already knew), no) 3. Did you look up any words (the answer to this question is important for the interpretation of the data) (answer options available: yes, no)

4. Do you feel like you now know the Arabic words? Why (not)? (answer options available: yes, no. With textbox to elaborate)

5. Did you use the signs in the experiment in one way or another? If yes, how? (answer options available: yes (with textbox to elaborate), no)

6. Were the signs useful to you while learning the Arabic words? If yes, how? (answer options available: yes, no. With textbox to elaborate)

3.3 Task Procedure 3.3.1 Trials

We made sixteen stimuli consisting of a spoken Arabic word, recorded by a native speaker, paired with a video of the corresponding sign from NGT. The audio of the spoken word starts playing at the exact frame that the sign starts. The audio and the video of one word are only played once per trial. Participants are asked to choose between two Dutch answer options, one being the correct Dutch translation of the Arabic word, and the other being a foil answer. The foil answer is always one of the other fifteen possible translations. All Arabic spoken words are presented once in a block, and the experiment has a total of four blocks. This means that all Arabic words were presented in the test trials a total of four times. In each block, the foil used with each Arabic spoken word is chosen at random from the fifteen Dutch translations. It is made sure that there is never a translation-foil pairing which is the same as in another block. The foils are randomly selected for each participant, so the experiment has different translation-foil pairings every time it is opened, so no participant had the same translation-foil pairings as another. Correct answer options were distributed across the right and left equally,


so 50% of the correct answer options were the left button and 50% of the correct answer options were the right button in each block. None of the trials gave feedback as to whether the participant had selected the correct answer. In between trials was a ‘+’ sign in the middle of the screen for 500ms to focus attention of the participant to the middle of the screen. An example of a test trial can be found in Appendix C.

To help participants understand what is expected of them during the experiment, we devised four practice trials. Because the participants were not familiar with Arabic, and we wanted to let participants understand exactly what they were doing, we decided to first have two practice trials with an English spoken word, accompanied by a corresponding sign from NGT. The practice trials were identical to the test trials, apart from two of the four practice trials being in spoken English instead of spoken Arabic. Two of the practice trials, one Arabic and one English, had an iconic sign-meaning relationship, and the other two practice trials, again one Arabic and one English, had a high idiosyncratic sign-meaning relationship. An example of the setup of a practice trial can be viewed in Appendix B.

3.3.2 Vocabulary tests

The vocabulary test consisted of an audio of one of the sixteen Arabic words, and sixteen answer options displayed on the screen: all of the Dutch translations for all of the Arabic words. When a participant selected the translation for the spoken Arabic word, a new spoken Arabic word was played and the participant could once again select the Dutch translation from sixteen answer options. At no point was the amount of answer options reduced, nor was any feedback given on the answers. The spoken words were only heard once and never repeated. There were no NGT signs available in the vocabulary test. This ensured that

participants were pairing the Arabic audio to the Dutch translation, instead of the NGT sign to the Dutch translation. An example of the vocabulary test setup can be found in Appendix D.


3.4 Experiment Procedure

Before starting the experiment, participants filled out the background questionnaire as described in 3.2.4. After the questionnaire was finished, participants were told what was expected of them in the experiment. They then proceeded with the practice trials as described in 3.3.1. The first two practice trials were ‘to cut’ and ‘to erase’ in spoken English, the second trials were ‘name’ and ‘rain’ in spoken Arabic. After the practice trials, participants were told that the experiment was now starting and that they had to make sure they were not disturbed for the next 30 minutes at least. They were also reminded that they could stop participating in the experiment at any given time without reason.

The procedure of a block consisted of the sixteen test trials, as described in 3.3.1, followed by a vocabulary test, as described in 3.3.2. There were four blocks in total.

All in all the experiment took about 30 minutes to complete, which was communicated up front to the participants, so they could take that into account in choosing a correct time and environment to complete the study.

The experiment was online for 47 days. In total there were 54 participants who have

completed the experiment, of which 53 met the requirements. Many others unfortunately were not able to start or finish the experiment due to technical issues.

3.5 Data-analyses

To analyze the data gathered from the experiment, we used generalized linear mixed models in JASP. We used the correct responses to the four vocabulary tests in between each block to track each participant’s progress in acquiring the novel Arabic words. In a regular cross situational word learning paradigm, the correct responses in the trials would have been used, but as mentioned before, these responses were not representative of the acquisition of the Arabic words of our participants, but were more a representation of how well the


participants could pair the sign with the Dutch translation (which they did really well, most scored 100% correct in the first block). The vocabulary test combatted this problem because it had no signs for participants to lean on in their translation, and therefore we think that the correct responses to these vocabulary tests are a better representation of the acquisition of the Arabic words than the correct responses in the trials.

We looked into the correct responses on the vocabulary tests, how they evolved after each learning block and how this evolvement was different between iconic and high idiosyncratic signs. After that we looked at the effect age, educational level and word class respectively had on the correct responses per block per sign type. The results of these analyses are outlined in the next chapter.


4. Results

In this chapter we will present the results of our statistical analyses. We used generalized linear mixed models in JASP and R on our data, due to the design of the experiment. JASP was used to produce all of the figures seen below and the tables seen in Appendix G. R was used to check the random slopes, seeing as JASP had some trouble with generating these. It turned out that the random slopes did not add anything to our analyses, so we decided to only display the results found in JASP in this section. We will analyze the correct answers given in the vocabulary tests only.

4.1 Main results

The goal of our experiment was to gain insight into the learning processes of L2

vocabulary when presented with two types of signs: iconic and high idiosyncratic, and to see whether seeing one or the other type of sign while learning a novel L2 word makes a

difference to the learning outcome of the L2 word.

An overview of all of the results found in this experiment can be found in Appendix G.


Figure 2

Comparison of correct responses per block (bloknummer) per sign type (gebaartype): iconic (Ic) v. high idiosyncratic (HI)

Figure 2 shows the overall effect for all four vocabulary tests at the end of each learning block (bloknummer 1 to 4), as well as for the type of sign (gebaartype, Ic= iconic or HI = high idiosyncratic). As can be seen in figure 2, there is a significant effect between each of the blocks; the average correct scores of words supported by iconic and high idiosyncratic signs in the learning phase is significantly different in each block (p=0.246, see Appendix G, table 3 and 4). When separating the correct scores into the words supported by iconic and high

idiosyncratic signs, we see a main effect of the type of sign throughout all of the blocks (p=0.194 for block 1, p=0.414 for block 2, p=0.092 for block 3, see table 4 in Appendix G).

There is a difference in the chance of a correct response in the different sign type categories, which is significant: novel L2 words presented with an iconic sign were learned better than novel L2 words presented with a high idiosyncratic sign (p=0.044, see table 3 in Appendix


G). The difference between the types of signs is constant between blocks, so there is no significant difference in development across time.

Based on this model we have looked into further factors one by one, because it was not possible to display them all in one model.

4.2 Age

Due to the many different ages our participants had, JASP was unable to generate a plot from our data. The ages were divided into three age groups (one with the average age of 23.8, one with the average age of 41.3 and one with the average age of 58.8) to be able to look into difference across these age groups.

We found that age has an effect on the outcome of the experiment (p=0.003, see table 6, 7 and 8 in Appendix G for all of the data), but there is no significant difference in the age-effect between types of signs (p=0.518). Interestingly, when split out it is seen that age affects scores a lot in the first block (leeftijd*Bloknummer (1) p=0.016) but not in the second and third block (p=0.612 and p=0.670 respectively). This shows that younger participants had a significant advantage in the first block compared to older participants, but not in the second and third block.


4.3 Education level Figure 3

Comparison of level of education per block number per type of sign

For highest received education level, we had four options participants could choose from:

high school, MBO, HBO and WO. 4 participants had a high school education, 3 participants had MBO education, 10 participants had HBO education, and 36 had WO education. Because 4, 3 and 10 people respectively were not enough to draw conclusions from, we decided to fuse these into a ‘rest’ category, which we compared to the WO category.

Education level had no effect on the outcome of the experiment (p=0.294, see table 9 in Appendix G), and neither was there any significant difference between sign types when looking at education level (p=0.386, see table 9 in Appendix G). Even though figure 2 may suggest a difference between the two educational levels in block 4, this difference is not significant because the confidence intervals are overlapping (Rest 0.284<0.587 and WO 0.512<0.741, see table 10 and 11 in Appendix G)


4.4 Word class Figure 4

Comparison of nouns (nw) v. verbs (ww) per block per type of sign

When choosing the words for the experiment, we chose eight nouns and eight verbs (see table 2 in chapter 3 for an overview of the 16 words used). No significant effect due to word class was found (p=0.083, see figure 4 and table 12 in Appendix G). The effect is, however, bordering on significance. Zooming in on sign type, we can see a considerable difference for nouns compared to verbs within the iconic sign type category. This difference is not

significant (p=0.080, see table 13 and 14 in Appendix G), but bordering on significance. No such effect can be seen for the high idiosyncratic sign type; nouns and verbs were learned in the same manner.


Figure 5

Comparison of nouns (nw) v. verbs (ww) per type of sign

Whenever we left out block number as a fixed effect, we did get a significant effect for word class: overall, more correct responses were given for nouns in the iconic sign type category than in the high idiosyncratic sign type category (p=0.047, see figure 5 and table 15, 16 and 17 in Appendix G). No such effect was found for verbs.

4.5 Debriefing questionnaire

As explained in 3.2.5, after the participants had finished the testing phase of the

experiment, they were asked a few questions to gain insight into the way they had used the signs, whether they thought them useful and whether they had looked anything up.

Fortunately, none of the participants indicated looking up any words during the experiment.

The other three questions (whether the signs had been useful, and if so: in what way?, how they used the signs, and if they now know the Arabic words) were answered and interpreted in


a variety of ways. This means that participants who did not indicate something, didn’t necessarily not feel that way. They simply did not choose to write that in their remark. We therefore can only draw conclusions based on what was said, not on what participants didn’t say.

The answers to these two questions have been categorized as can be seen in table 18.

Table 18

Summary of answers given by participants in debriefing questionnaire

Summary of remark Number of

participants indicating this (percentage of total number of participants) The signs were useful to unravel the meaning of the Arabic

spoken words

34 (64%)

The signs were distracting the participant from the task 6 (11%) The signs were useful to remember the Arabic words 5 (9%) I did not use the signs during the vocabulary tests 4 (8%) I did use the signs during the vocabulary test 1 (2%) I imitated the signs at least one block 6 (11%) I visualized the signs during the vocabulary test 4 (8%)

I sometimes used the signs 1 (2%)

Signs helped me keep my focus 1 (2%)

I now feel like I know none of the Arabic words 17 (32%) I now feel like I know some of the Arabic words 20 (38%)


I now feel like I know half/the majority of the Arabic words

8 (15%)

I now feel like I know all of the Arabic words 6 (11%) I tried to connect the phonetic sounds of the Arabic words

to Dutch words I know

10 (19%)


5. Discussion and conclusion

5.1 Discussion on main results

The results of our experiment show that the differences in correct scores between blocks are higher than chance, meaning that participants overall had significantly more correct scores in each block compared to the previous block; they started learning the Dutch meanings of the Arabic words. The results show a significant difference between the two types of signs: iconic and high idiosyncratic (p=0.044) as well, and this difference is constant between blocks (p=0.246), meaning that there is no significant difference in the development across time.

These findings support our hypothesis that iconic signs support the learner more in the acquisition process of a novel L2 word than high idiosyncratic signs. The current experiment was not sufficient to gain 100% chance of a correct answer (in the final block, the high idiosyncratic condition had a 47% chance and the iconic condition had a 66.5% chance of a correct answer). In other words, final retention was not reached. We think this is due to the low number of blocks we had and/or to the high (compared to the recommended 10 by Huang, Kim and Christianson (2019)) number of novel L2 words we presented to participants. We therefore have no definitive answer to our second hypothesis: that final retention was reached sooner in the iconic condition than in the high idiosyncratic condition. We can, however, extrapolate from our current findings that it seems likely that this hypothesis is confirmed, seeing as there is no difference in development seen over time between the two sign conditions. We did not incorporate a control condition without a sign, because previous research had already shown that using signs to learn novel L2 vocabulary is more effective than not using signs or using abstract signs (Huang, Kim & Christianson, 2019; Macedonia, Müller & Friederici, 2011; Tellier, 2008 a.o.). We are solely looking at the effect that seeing iconic and high idiosyncratic signs have on the learning process of novel L2 vocabulary.


5.1.1 Discussion on sub results

As for the age and educational level of the participants, and the word class of the stimuli, we found that only age had an effect on the outcome of our experiment: the percentage of correct scores declines as the average age increases. This is the case for each block and for both sign conditions. Vocabulary learning is mainly dependent on memory performance (Magnusson &

Greitz, 1997), and that performance, particularly when it is necessary to form new

connections to recall newly acquired information (Burke & Mackay, 1997), is impaired in the ageing process (McIntyre & Craik, 1987; Schacter et al. 1994). It is therefore not a surprise that we see a decline in percentage of correct scores when the participants are older.

5.2 Our results compared to those of Huang, Kim and Christianson (2019)

Huang, Kim and Christianson (2019) showed that both iconic and high idiosyncratic signs aid in L2 vocabulary acquisition more than an abstract sign would. The results of using iconic and high idiosyncratic signs in L2 vocabulary acquisition did not significantly differ from one another, so their conclusion was that both iconic and high idiosyncratic signs aid equally in L2 vocabulary acquisition. Looking at the learning process of L2 vocabulary supported by iconic and high idiosyncratic signs in our experiment taught us that supporting iconic signs increase the chance of a correct answer in the vocabulary test more than the high idiosyncratic signs do. This increase is independent of educational level of the learner, but ageing does come into play, most likely due to memory constraints. Whether there is a difference between acquiring nouns and verbs remains to be seen and is a good topic for future research.

Our findings are in contrast with the results of the experiment of Huang, Kim and Christianson (2019). Their participants heard the target words and saw their accompanying signs a total of four times, the same amount of times as the participants were presented with the target words in our experiment. One of the main differences between Huang, Kim and Christianson’s experiment (2019) and ours is that they did not use cross situational word


learning, but instead presented the learners orally with the target words in the target language twice, and then presented them orally with the correct translation. The experimenter repeated this practice twice, before subjecting them to a vocabulary test. A possible explanation for the difference in results could be related to not being given the correct translation but that having to choose between two possible translations gave participants a learning advantage. Future research should shed light on this apparent contradiction of findings.

5.3 The difference in aid high idiosyncratic and iconic supporting signs can provide As stated in our hypotheses, the difference between the high idiosyncratic and the iconic condition could be explained on the basis of the resource depletion for output hypothesis (RDO) (Barcroft, 2006). This explanation is also supported by some of the remarks

participants wrote down in the second questionnaire in the experiment: “I think [the signs]

had a lot of influence subconsciously. Signs like bird, house, pepper were logical and I knew them fast. Not attachment and advertisement.” This remark stands out because the participant, unknowing of the different sign types, highlights exactly this distinction: bird, house and pepper were iconic signs (logical to the participant). Attachment and advertisement were high idiosyncratic signs (not logical to the participant). Another participant remarked: “I don’t have any knowledge of a sign language, but I did recognize some of them immediately, like

calling”. This shows us that the form-meaning relationship was already present in these participants, they did not have to be formed on the spot, as with the high idiosyncratic signs.

In line with the RDO, this would suggest that participants had more trouble with the high idiosyncratic signs because they had an extra task in this sign condition, compared to the iconic sign conditions.

5.4 What can we take away from this?

We would like to highlight that the experiment consisted of a learning phase, in which the participants saw the congruent signs (either iconic or high idiosyncratic) with the L2 word and


the Dutch translation, and a vocabulary test, in which the participants did not see the signs.

The current effects found of both sign conditions were found in the correct answers given in the vocabulary test. This means that, in spite of participants not seeing any signs during the vocabulary test, seeing them during the learning phase was enough to diverge the support each sign type could provide.

The results of our experiment regarding the level of education, namely that the level of education does not have an effect on the aiding properties of iconic and high idiosyncratic signs in L2 vocabulary learning, make the translation from an experimental environment to the L2 classroom easy: any adult L2 learner from any educational background can benefit equally from seeing corresponding iconic or high idiosyncratic signs in the L2 vocabulary education. These benefits were still present in the vocabulary tests, despite not seeing the signs there. We therefore think that using signs in the classroom will still benefit learners during testing phases and might increase testing scores, even when the test does not contain the signs. All in all we highly recommend L2 teachers and methods to start incorporating high idiosyncratic and especially iconic signs into their L2 vocabulary education.

5.5 Future research

We have some suggestions for future research that stem from the results we found in this experiment.

5.5.1 To give the correct translation or not to give the correct translation, that is the question

As stated above, we think it would be fascinating to look into the difference in acquisition when participants are given the correct translation (as a feedback mechanism), versus when they have to choose the correct translation between two options without getting feedback. If the results of this experiment show that there indeed is a difference in acquisition, and giving participants two possible translations instead of giving the correct translation, it would


confirm our suspicion as to why our results differ from those of Huang, Kim and Christianson (2019). The outcomes of this experiment would also have implications for teaching strategies in L2 vocabulary education.

5.5.2 Complete L2 vocabulary acquisition

It would be interesting if our experiment would be replicated with less target words but with more blocks to the point that all participants achieve 100% correct scores in the last vocabulary test. In this way, it would be possible not only to look at the full acquisition process from start to finish, but also to see whether the two sign conditions significantly differ from one another at the final block or not. We hypothesize that there would still be a

significant difference, since the difference between the iconic and high idiosyncratic condition is not changing across the four blocks in the current experiment.

5.5.3 Age

Seeing as we found a significant difference between age-groups for both the iconic and the high idiosyncratic condition, it would be interesting to find out more about how much signs aid L2 vocabulary acquisition in different age groups compared to not having a supporting sign in the learning phase. Specifically, the difference between adults of different age groups has had very little research, especially compared to the amount of research that has been conducted on the topic of children versus adults. It would be interesting to know more about how different adult age-groups are affected by the support of signs in L2 vocabulary learning.

5.5.4 Word class

As pointed out in paragraph 4.4, we found a p-value bordering on significance for nouns in the iconic condition, which was actually significant when block number was taken out of the analyses as a random effect (see table 15-17 and figure 5). This most definitely deserves to be researched more in the future, as it is possible that this will gain us new insights in the

difference in acquisition between verbs and nouns, as well as how signs come into play there.


5.5.5 Pepper

One of the target words in the iconic sign condition was pepper. This word stood out in that it was mentioned by many of the participants in the debriefing questionnaire, as well as in informal discussions after the experiment was finished. The overall consensus was that people found the Arabic word for pepper significantly easier to remember than the other target

words. A multitude of reasons were given by the participants: the Arabic word sounded more cheerful than the other words, the Arabic word had the same rhythm as the accompanying iconic sign, it had a unique sound compared to the other target words. Multiple times, participants indicated that they knew none of the words after the experiment had ended,

except for pepper. Or that they had only remembered a couple of the words, including pepper.

Due to the remark about the apparent cheerfulness of the Arabic word for pepper compared to the other novel Arabic words, we think it is interesting to look at (diverging) prosody as a factor in L2 vocabulary retention.

The remarks about rhythm and the repetitive nature of the word for ‘pepper’ spark our interest on the effect of these elements on L2 vocabulary retention. The Arabic word for pepper is لفلفلا [ɛlfɪlfɪl] and the NGT sign for pepper can be seen in Appendix A. Seeing as the [ɛl] part of the word was present in many words, it is plausible participants saw this as either an article or at the very least did not focus so much on that part of the word as they did on what followed it, seeing as the first syllable was always constant. [fɪlfɪl] has a repetitive nature of one syllable [fɪl], which corresponds to the repetitive nature of the NGT sign (as can be seen in Appendix A, the motion is repeated twice). Even looking at the Dutch translation (peper), there is a repetitiveness to the word (pe is used twice). This makes for a similar rhythm (everything is repeated twice) as well as a repetitiveness in the nature of the target word, the Dutch translation, and the supporting NGT sign. It would be interesting to look into the effect of corresponding rhythms in both target words and supporting signs. One could, for




Related subjects :