Replication study of: Bongaerts et al. (1997).

(1)

1

Replication study of: Bongaerts et al. (1997).

Age and ultimate attainment in the pronunciation of a foreign language.

Name: Amy Jackson S1856812 Supervisor: Prof. Dr. M.H.Verspoor Date: 18-06-2018

Wordcount: 11856

(2)

2 Table of contents Abstract 3 Introduction 4 Motivation 4 – 5 Background 6 – 11 Statement of purpose 11 – 13 Method 14 Subjects 14 – 15 Materials 15 – 18 Procedures 18

Design and Analyses 19 – 20

Results 20 – 21

Discussion/Conclusion 21 - 25

References 26 – 27

(3)

3 Abstract

This study replicates Bongaerts et al. (1997), T., Summeren, van C., Planken, B., & Schils, E. (1997) on age and ultimate attainment in the pronunciation of a foreign language. The goal of Bongaerts et al. (1997) was to determine whether it is possible to attain a nativelike level in the

pronunciation of a foreign language with a late age of onset. Bongaerts et al. (1997) found that only a few of his participants got close to nativelike level. As English is still relatively close to Dutch and it is difficult to prevent early exposure of English to Dutch people, I chose a different target language. This replication study focuses on Dutch learners of Finnish, which is very distant from the Dutch language and early exposure is rare.

There have been many studies on the subject of late second language learning and on a possible Critical Period Hypothesis. In the background section, several of these studies (e.g. Birdsong & Molis (2001), Piske et al. (2001), Moyer (1999), and DeKeyser (2000) are discussed and their position on the Critical Period hypothesis. The findings of most of these studies, like Bongaerts et al. (1997) and this replication study, are not in line with the Critical Period Hypothesis.

For the current study there were 3 groups of participants: - Dutch learners of Finnish (12)

- Highly proficient Dutch learners of Finnish (6) - Native speaker control group (4)

As the current study is on pronunciation and not on vocabulary and/or grammar, I decided to focus solely on Bongaerts et al. (1997) second study, in which he elaborates on his previous article in 1997

In his second study, Bongaerts et al. (1997) had the participants record the sentences. The results were interesting, most of the Dutch learners as well as the highly proficient learners were not considered Native like. However, it was surprising to find that one of the low proficient learners actually received a higher nativeness rating than any of the highly proficient learners and was not so far removed from one of the native speakers. In conclusion these results do not refute the Critical Period Hypothesis, but do suggest that it is possible to achieve near nativelike levels by some learners.

(4)

4 Introduction

Bongaerts et al. (1997), T., Summeren, van C., Planken, B., & Schils, E. (1997) did a study on age and ultimate attainment in the pronunciation of a foreign language. The goal of the study was to determine whether it is possible to attain a nativelike level in the pronunciation of a foreign language with a late age of onset. The results suggest that it is not impossible to achieve an authentic, nativelike pronunciation of a foreign language even with a late age of onset. Bongaerts et al. (1997) included two studies in order to get these results. The first one showed that, amongst other things, the fact that the control group as well as the judges were from different parts of England, their judgement was influenced. The second study was a slightly simplified version of the first one, but got more accurate results. Initially this study had as a goal to replicate both studies, as they were more detailed, but in the end, this proved to be more complicated than I had first anticipated. This will be further explained in the section “method”. Therefore, this thesis will be a replication study of the second study.

Motivation

I have been interested in ultimate attainment and age for quite some time now and I believe that it is possible to attain a nativelike level in a foreign language even with a late age of onset. The results of Bongaerts et al. (1997) suggest that this is indeed possible for at least late Dutch learners of English.

However, even though the Dutch learners of English, tested by Bogaerts (1997) have not been overly exposed to English language before the age of 12, in this day and age (and also already in 1997) it is almost impossible to avoid exposure to the English language. Even in 1989 there were already concerns as to the enormous impact the English language had on the Dutch and came to a conclusion that there are multiple reasons to use English instead of Dutch (Bakker 1989). Research from 1998 also points out that in 1996 one third of television commercials on Dutch television were in English (Gijsbers et al 1998).

Also, the English language is still relatively close to Dutch, as they are both part of the (West) Germanic branch of the Indo-European language family and thus closely related to each other and to Frisian, German and the Scandinavian languages (Smith & Swan 2001). These two factors might be of influence on achieving ultimate attainment in the pronunciation of a foreign language.

(5)

5 as used in Bongaerts et al. (1997), but instead of using Dutch learners of English, I will use Dutch learners of Finnish for my testing groups. Finnish is very uncommonly heard in the Netherlands, so it is a lot more likely to find learners who have not been exposed to the Finnish language before the age of approximately 12 years old. Furthermore, as the Finnish language is a member of the Finno-Ugric language family, it is quite different from the Indo-European family, to which both Dutch and English belong (Karlsson 1983). I expect using a language more distant from Dutch will shed new light on this study.

Furthermore, many different languages have been used as the target L2 being learned in late second language acquisition. Most studies have used English as their target L2 language. Other languages include German, Hebrew, Spanish and Thai. With L1 languages as Spanish, English, Arabic, Japanese, Persian, Thai, Italian, Dutch, Mandarin, Taiwanese, Russian, Swedish and Korean (Piske et al. 2001). Interesting to me is that the combination Dutch (L1) - Finnish (L2) has not been examined yet. And as Suter (1976) and Purcell & Suter (1980) argue, non-native pronunciation of an L2 is to some extent dependent on their L1.

An additional reason for choosing the Finnish language over other languages distant from Dutch is that I have my bachelor's degree in Finno-Ugric languages and cultures, specialised in the Finnish language and culture. This has aided me in assuring the tests will be similar to the English ones and helped me find sufficient Dutch learners of Finnish.

(6)

6 Background

Bongaerts et al. (1997) shows that it is not impossible to achieve an authentic, nativelike

pronunciation of a second language after a specified biological period of time. The paper reports on two studies, which had the aim to determine whether or not late second language learners who had achieved a nativelike performance in the pronunciation of a second language could be identified. The study included speech samples from a native speaker control group and by two groups of learners, one of which consisted of highly successful learners only, and one of average level (Bongaerts et al. (1997).

Bongaerts et al. (1997) shows that it is possible to attain a nativelike level for Dutch learners of English, who have not been exposed to the English language more than incidentally before the age of approximately 12. The age of 12 was chosen based on Scovel 1988's argument that there is a critical period for the acquisition of the pronunciation of a second language only because pronunciation is “the only aspect of language performance that has a neuromuscular basis,” requires “neuromotor involvement,” and has a “physical reality” (p. 101). He predicted that learners who start to learn a second language later than age 12 will never be able “to pass themselves off as native speaker” (p. 185). Scovel (1988, p. 181), however, does allow for the possibility that there may be some “super exceptional” foreign language learners, about 1 in 1.000 in any population of late learners, who are not bound by critical period constraints.

Critical Period Hypothesis (CPH) and Universal Grammar (UG)

There are many studies that report on second language acquisition and the critical period. The Critical Period Hypothesis states that there is a limited developmental period during which it is possible to acquire a language to normal nativelike levels. Once this window of opportunity is passed, however, the ability to learn language declines (Birdsong 1999). This means that it would have to be impossible to achieve ultimate attainment in second language acquisition with a late age of onset.

The Critical period hypothesis is often associated with Universal Grammar. The knowledge of language, linguistic competence, is assumed to be represented in the form of a generative

grammar, an abstract system of principles and rules which produce the grammatical sentences of a language. Principles and rules account for such formal properties of language as syntax,

phonology, morphology and certain aspects of semantics. This so-called black box is very rich. It enables us to understand and produce utterances which we have never heard before. We know

(7)

7 that certain utterances are grammatical, others ungrammatical, that some sentences are

ambiguous and that some meanings can be represented by more than one sentences. Chomsky raised the question: What constitutes knowledge of language?, which is directed towards the knowledge possessed by adults. The more important question for L2A is how children come by this knowledge (White 1989). This abstract system of principles and rules which produce grammatical sentences of a language can been seen as a black box. According to the Critical Period Hypothesis, at a certain age (somewhere around puberty) you lose access to this black box, to your so-called Universal Grammar, making ultimate attainment in second language acquisition with a late age of onset almost impossible.

There are many studies that support the Critical Period Hypothesis and universal grammar; however, more and more studies appear that would contradict the Critical Period Hypothesis as they show results of late second language learners which are not distinguishable from native speakers. Piske et al. (2001) reviewed earlier studies on factors that might affect the degree of foreign accent in an L2. They found that the most important factor is indeed the Age of L2 learning. But as the earlier researches lacked adequate experimental control the importance of other variables, such as:

- Length of residence in an L2-speaking country - Gender

- Formal instruction - Motivation

- Language learning aptitude

- Amount of native language (L1) use

Thus Piske et al. did their own study in which they examine the influence of Italian-English

bilinguals’ age of L2 learning, length of residence in an L2-speaking environment, gender, amount of continued use of native language and self-estimated L1 ability on degree of L2 foreign accent. They found that age of L2 learning as well as continued L1 use affect the degree of foreign accent.

Large differences have been found between studies in design and methodology, so there is no surprise that L2 foreign accent studies have at times produced conflicting results, depending on, amongst other things, the factors taken into account, motivation of the subjects, proficiency of the subjects, L1, rating method and whether or not there is a native speaking control group (Piske et al. 2001).

(8)

8 Birdsong & Molis (2001) researched several aspects of the Critical Period Hypothesis in order to challenge it. Several types of behavioural evidence would constitute support for a maturational view of the limits on L2 attainment in order to proof that limits on attainment are maturational in nature:

- Linguistic performance should correlate negatively with the age at which L2 learning begins. (L2 learning should start prior to the end of maturation → before mid/late teens)

- Among late learners, correlations of learning onset with performance would suggest a role for mechanisms other than maturation.

- There should be few, if any late learners who perform in the range of native controls. This null result would suggest biological constraints.

- This should apply to L2 acquisition generally → Critical period-type effects and near-zero incidence of nativelike attainment should be observed no matter what the pairing of L1 and L2.

When looking at these three types of evidence that would favour a critical period account of L2 acquisition limits Birdsong & Molis (2001) study found the following:

Evidence of post maturational age effects and native language effects, this contradicts the statement that age effects prior to, but not after the end of maturation could be found.

Furthermore, modest evidence of nativelike performance was observed, while the Critical Period Hypothesis states that there should be few, if any, late learners who perform in the range of native controls. As evidence has been found, this contradicts the statement on near-zero incidence of nativelike levels of attainment. Finally, according to the Critical Period Hypothesis, critical period-type effects and near-zero incidence of nativelike attainment should be observed no matter what the pairing of L1 and L2. However, Birdsong & Molis (2001) found data suggesting the results may vary depending on different L1-L2 pairing.

The results in the replication study by Birdsong and Molis (2001) reject the Critical Period

Hypothesis, as in order for the Critical Period Hypothesis to be true a unique sensitive period must be shown, distinct from any general age-related increase or decline in cognitive abilities. Second, the biologically regulated receptivity of the learner must be demonstrated to be specific to linguistic, and not just any, stimuli (Long 1990).

(9)

9 DeKeyser (2000), for example, states that there can be no exceptions to the age effects that the Critical Period Hypothesis seeks to explain. DeKeyser (2000) showed that very few of his adult subjects scored within the range of his child subjects on a grammaticality judgement test. He states that the few that did had a high level of verbal analytical ability. This is a replication study of Johnson and Newport (1989) and aims to explain the apparent exceptions in their study. He cites that if the scope of the Critical Period Hypothesis is limited to implicit learning mechanisms, then it appears that there may be no exceptions.

De Keyser chose Hungarian as the L1 language as it is a non-Indo-European language and even the basic structures of English morphosyntax present many elements that have to be acquired by these speakers. The participants had to take 2 tests:

- Grammaticality judgement test - Aptitude test

The result of these tests was that only a few of the adults managed to achieve a result that

overlapped with the subjects with the age below 16. Of these adults all but one had a score of 6 or higher on the aptitude test, which indicates a high level of verbal aptitude. DeKeyser dismisses this exception by claiming that as the subject was doing postdoctoral studies in the natural sciences, he must be of above-average analytical ability and that his aptitude test score is not indicative of his analytical abilities.

Even though DeKeyser dismisses the existence of an exception to his hypothesis, there is a clear exception to it. First of all, he dismisses the aptitude test that was supposed to be a reliable test in order to prove his hypothesis that adults can only achieve ultimate attainment if they have a high level of verbal analytical ability. So, if the test score is not indicative for this test subject, it might also not be indicative for the other subjects either. Furthermore, it is an assumption, and not a fact that this subject ought to have above-average analytical ability, as he is doing postdoctoral studies in the natural sciences. Especially as students of exact sciences often have more difficulty with language acquisition.

Also, with his study he confirms Piske et al.’s theory on language learning aptitude being an important factor in late second language learning. DeKeyser supports the Critical Period

(10)

10 states that it should not be possible to achieve ultimate attainment in second language acquisition if age of L2 learning commences after puberty.

Moyer (1999) also challenges the Critical Period Hypothesis. The results of her study did not show any overlap between the students and the native speaker control group, apart from one

exception, who even got rated with a higher native level than one of the native speakers. This exceptional speaker appeared to be the only subject that aimed for professional motivation as the goal driving his study. He was largely self-taught and cited a strong desire to acculturate and to sound German. He only started learning German at the age of 22 and had no other language background, other than English, before this age.

So, the Moyer (1999) study results do not clearly support the Critical Period Hypothesis either as motivation and instructional variables would be able to override the impact of age, as all apart from one still performed in a non-native range. However, the results do show that several types of variables exist in close and significant relationship to language learning success. Examples of these are motivation, language aptitude, formal instruction and amount of native language (L1) use. The results also show that age of exposure does not significantly explain the outcome in isolation from other influences. Moyer (1999) found that maturation did not offer a unique explanation of the variance in outcome after controlling for professional motivation and type of feedback.

The findings in Moyer (1999) support my own belief that age is not necessarily a constraint for ultimate attainment in second language acquisition. Even though most of her subjects could be clearly distinguished from the native speaker control group, there was still one that was not distinguishable from the native speakers. What makes this even more remarkable is that, even though the tasks selected in this specific study were chosen for their level of difficulty for English learners of German, this exceptional speaker, who had no experience with languages apart from English before the age of 22, still managed to be indistinguishable from native speakers. This indicates that motivation is a large factor. This is something I have not focussed on in my own study, but definitely worth doing more research on.

All of these studies, even DeKeyser (2000), show that late L2 learners can indeed achieve near nativelike levels, depending on many variables. They are usually called exceptions, especially by supporters of the Critical Period Hypothesis. However, if exceptions are found in all the studies done on late second language learning, can we still call them exceptions.

(11)

11 Yet there is still too much more research to be done. As Moyer (1999), I believe that the L1-L2 pair can be of importance in achieving ultimate attainment in second language learning. Also, many studies, such as Birdsong and Molis (2001), did not take age of first exposure into account. Bongaerts et al. (1997) did take this into account, though no exposure to English whilst being Dutch can appear to be rather difficult. Therefore, this study uses the language pair Dutch (L1) - Finnish (L2).

Statement of purpose

In this paper I will report on a replication study of Bongaerts et al. (1997), using Dutch learners of Finnish rather than of English. As mentioned above, even though Bogaerts' test group had been minimally exposed to the English language before the age of 12, in this time and age English is everywhere. Due to this, complete prevention of exposure to English is almost impossible. Finnish, however, is very uncommonly heard in the Netherlands and therefore it should be less

complicated to find learners of Finnish who had not been exposed to the language at all before the age of 12. As the Finnish language is a member of the Finno-Ugric language family, it is quite different from the Indo-European family, to which both Dutch and English belong (Karlsson 1983). These two factors will shed new light on whether it is possible to achieve a nativelike

pronunciation of a second language with a late age of onset.

In the study of Bogaerts 1997 the test subjects had to perform 4 tasks, which were recorded: - A spontaneous speech sample by talking in English about their most recent holiday abroad - Read out a brief English text (length: 84 words)

- 10 English sentences - A list of 25 English words

The sentences, which varied in length from 9 to 12 syllables, were selected such that they

contained phones that range from very similar to very different from Dutch phones (Bongaerts et al. 1997).

In order to do a proper replication study, a Finnish text of approximately the same length and sentences, which contain phones with a similar range in variety to Dutch phones, are selected. While selecting words for the wordlist, it is taken into consideration that Finnish has a very consistent pronunciation, as there is almost one-to-one correspondence between letters and

(12)

12 sounds (Korpela 2010). Finnish has 8 vowels:

- a [a] as in English ‘father’, but shorter and in Dutch ‘als’ - e [e] as in English ‘let’ and in Dutch ‘ekster’

- i [i] as in English ‘feet’ and in Dutch ‘fiets’

- o [o] as in English ‘ought’ and in Dutch ‘opruimen’ - u [u] as in English ‘wood’ and in Dutch ‘oer’

- y [y] has no English equivalent, but as in Dutch ‘Utrecht’ - ä [æ] as in English ‘bat’, but no Dutch equivalent

- ö [ø] has no English equivalent, but in Dutch ‘stomme’

And 13 consonants: - d [d/t] - h [h/ɦ/ç/x] - j [j] - k [k] - l [l/ɬ] - m [m] - n [n/ň] - p [p] - r [r] - s [s] - t [t] - v [ʋ]

Other consonants, such as b [b/p], c [k/s], g [g/k/ŋ], f [f/ʋ], q [k], w [ʋ], x [ks] and z [ts] are only found in loan words. Many sounds, especially consonants, are similar to Dutch, apart from r [r], which is a rolling ‘r’ articulated in the alveolar area of the mouth. While Dutch speakers tend to use a guttural ‘r’. Another challenging sound for Dutch learners of Finnish is ‘h’. In Dutch this sound only occurs in the middle of a word in the form of ‘ch’ [x] e.g. ‘lachen’ (to laugh). In Finnish it often occurs in the middle of the word, but alone e.g. ‘lehmä’ [lexmæ]. So, it looks different, while the pronunciation is similar.

The most difficult for Dutch learners of Finnish is not necessarily the pronunciation of these sounds, but the confusion between sounds. For example, ‘a’ in Dutch can also be [a:] while only

(13)

13 one ‘a’ is written. Even more difficult for Dutch learners of Finnish are vowels like ‘y’ and ‘u’, as Finnish ‘y’ would be written as ‘u’ in Dutch and the Finnish ‘u’ would be a diphthong in Dutch: ‘oe’. This could cause some confusion when reading texts, words and sentences out loud, especially for learners of Finnish. However, as these are known sounds in Dutch language as well, albeit in a different spelling, it is expected that the non-native groups will be able to pronounce these sounds properly as long as they remember that certain letters combine with different sounds in Finnish.

Pronunciation is the easiest and at the same time one of the most difficult parts of the Finnish language. As stress of a word is always, without exception, on the first syllable of a word. And every single letter is always pronounced. Vowels in Dutch have different pronunciation depending on whether they occur in an open or closed syllable. E.g. ‘halen’ (to get), where ‘a’ is [:a] and hallen (halls), where ‘a’ is [ɑ]. This is very different from Finnish pronunciation where in e.g. ‘ala’ (area) and ‘alla’ (under) ‘a’ in both cases is [ɑ]. As mentioned before, every letter is pronounced in Finnish, the importance of this becomes evident when we look at the following words. If certain letters are or are not pronounced, the meaning of a word could change:

Tuli  fire Tuuli  wind Tulli  customs

Another difficult thing about the Finnish language for Dutch speakers is that Finnish, compared to Dutch, is very monotonous. In Dutch the pitch at the end of the sentence rises in order to imply a question, whilst Finnish uses suffixes to indicate a question sentence (Jackson 2009).

The expectation is that the languages being so far apart will show advantages as well as disadvantages to the learner. As it is shown that the greater the difference between the two languages, the more negative effects of interference are expected to be. Though, if the languages are near, phones that sound alike, but are slightly different, might be more difficult to learn (Bhela 1999).

Jackson (2009) explains that the stress in a Finnish word is always on the first syllable of the word. The expectation is that this study will show that this should not show any difficulties with two or three syllable words, but as in Dutch stress can also be on other syllables it is unusual, and it is expected to prove more difficult for the Dutch learners of Finnish and the near-natives to correctly place the stress in words with more than three syllables.

(14)

14 Method

As this is a replication study we will first look at the subjects used by Bongaerts (1997). In his studies three groups of test subjects were required:

- Highly successful Dutch learners of British English (11, mean age 30) - Dutch learners of British English, any level (20, mean age 42)

- British English control group (10, mean age 27)

The Dutch learners of English had not been overly exposed to the English language before the age of approximately 12 and had not begun learning the language until after puberty.

In order to replicate the study, I first thoroughly went through the original one. The original study consisted of 4 tests:

- Spontaneous speech sample - Short text (84 words)

- 10 sentences - 25 words

The sentences varied in length from 9 to 12 syllables and were selected such that they contained phones that range from very similar to very different from Dutch phones.

Subjects

3 groups of participants similar to the original study were found:

- 6 Highly successful native Dutch learners of Finnish (mean age 42) - 12 native Dutch learners of Finnish, any level (mean age 29) - 4 Finnish control group (mean age 24)

The first group of participants were found via a network of Dutch speakers of Finnish on Facebook. Many of them have been living in Finland for years and are highly successful in their language acquisition. The second group of participants consist of fellow students from my studies on Finno-Ugric languages and culture. The third group of participants were formed by native speaker Finnish acquaintances.

None of the Dutch learners of Finnish had any exposure to the Finnish language before the age of approximately 12 years old and they all started learning Finnish after puberty. Some at university, some through self-study and living in Finland and some via language courses in the Netherlands or

(15)

15 in Finland. See table 1.

Table 1: Method of language learning

Participant Type Instruction

1 Learner University 2 NN Self-taught 3 Learner University 5 NN University 6 Learner University 7 NN Self-taught 8 NN Self-taught

9 Learner Language course

10 NN University

12 Learner University

17 NN Language course

22 Learner Self-taught

Two near native speakers were excluded from the study, one due to personal circumstances and one due to incomprehensive recordings.

Materials

The original tests are stated above, I originally intended completing the following tests - Spontaneous speech sample

- Short text (72 words) (see appendix 1) - 10 sentences (see appendix 2)

- 25 words (see appendix 3)

(16)

16 that they contained phones that range from very similar to very different from Dutch phones. These sentences were checked by my former professor of Finno-Ugric languages and cultures, Cornelius Hasselblatt. Below the sentences are listed, each sentence will be analysed on what Finnish phones are similar and different from Dutch phones

1. Kyllä, Kytäjä on Hyvinkäällä sijaitseva kylä. (Yes, Kytäjä is a village next to Hyvinkää.) Differences in this sentence are the double consonants in ‘kyllä’ and in ‘Hyvinkäällä’. The letter “ä”, which is used multiple times in this sentence is a phone that is unfamiliar in the Dutch language. Every other phone is also used in Dutch, though it might have different spelling. Also, the words ‘kyllä’ and ‘kylä’ are purposefully chosen for the same sentence, as the importance of distinguishing single and double consonants/letters is tested by this. ‘kyllä’ means ‘yes’ and ‘kylä’ means village.

2. Hän ei halunnut luopua periaatteestaan. (He did not want to abandon his principles.) This sentence was selected for its diphthongs in ‘luopua’ and the multiple syllables and double vowels and consonants in ‘periaatteestaan’. In order to pronounce ‘periaatteestaan’ properly one needs to remember that the stress is on the first syllable and how double vowels and consonants are pronounced. This sentence will most likely be the most difficult for the non-natives.

3. Harri on ylpeä hänen kauniista pikkusiskostaan. (Harri is proud of his beautiful little sister.) The most difficult part of this sentence is the name ‘Harri’, due to the difference in

pronunciation of ‘r’. The name ‘Harry’ is also a common Dutch name, the similarities of the word in combination with the different pronunciation of the ‘r’, makes this a challenging sentence.

4. Kissa hyppäsi pöydältä tuolille. (The cat jumped from the table onto the chair.)

Seemingly a fairly simple sentence, but the diphthong ‘öy’ looks difficult, even though it is similar in pronunciation to the Dutch diphthong ‘eu’. What makes this sentence

challenging is the word

‘hyppäsi’. The double consonant in this word causes it to have slightly more aspiration time than Dutch would have. As in Dutch only one ‘p’ would naturally be pronounced.

5. Joonas uskoo, että Suomesta tulee maailmanmestari. (Joonas believes that Finland will become the world champion.)

This sentence was specifically chosen for the double vowel ‘o’, as the Dutch pronunciation of ‘o’ is similar, but slightly different from the Finnish pronunciation. This is expected to be difficult for the participants of this study.

(17)

17 6. Olisin mieluummin kotona, kuin täällä. (I would rather be at home, than here.)

The diphthong ‘ui’ in ’Kuin’ in this sentence is a sound that is similar to the Dutch ‘oei’ but slightly different, just like ‘o’ in sentence 5.

7. “Anna kirja!”, Essi huusi. (“Give me the book!”, Essi screamed.)

As this an imperative sentence, it is expected to be difficult for the non-natives. Even though it is an imperative sentence, the stress is on the first syllable and the sentence will stay monotonous compared to Dutch. It is expected that the non-natives will have a slight pitch raise while putting stress on the first syllable of ‘kirja’. Also, the Finnish ‘r’ in the middle of the word will be tough, especially as the ‘i’ is articulated in the back of the mouth while the ‘r’ is formed in the alveolar area of the mouth.

8. Olen yrittänyt soittaa Sallalle koko aamu. (I have tried calling Salla the entire morning.) The repetition of the double consonant ‘l’ in Sallalle is expected to be difficult for Dutch speakers, as all four consonants need to be pronounced.

9. Laura valitsee että hän ei valita (Laura chooses not to complain.)

The difficulty of this sentence lies in the word ‘valita’ as it is crucial that the word is

pronounced correctly. ‘Välitä’ sounds similar and if ‘a’ is not pronounced correctly, it could easily sound like ‘ä’. In which case the sentence would mean “Laura chooses not to care” instead of the intended meaning: “Laura chooses not to complain”. Or other combinations if ‘valitsee’ is pronounced as ‘välitsee’, which e.g. would create the sentence “Laura cares that she does not complain”.

10. Kuka rakastaa tätä kaunista punaista kukkaa? (Who loves this beautiful red flower?) This sentence will most likely be the easiest one to pronounce. It contains minor difficulties in pronunciation such as ‘ä’ and ‘u’. The most difficult will probably be ‘kukkaa’, as it contains a double consonant and double vowel combination.

I had the subjects use Audacity to record the test sentences, to ensure that the recordings would be easy for me to edit into an anonymous judging system, for which I used Audacity to edit and Google Forms for judging and analysing.

I intended to edit the spontaneous speech samples to filter out most of the errors, where possible, as this is a study on pronunciation and not on grammar. This was also done in the first study of Bongaerts et al. (1997). If the spontaneous speech sample turned out to be too influenced by grammatical or vocabulary errors I would have excluded it from the research. In Bongaerts et al.

(18)

18 (1997) ' second study, only the sentences were tested, but as I intended to do a complete

replication study, I decided to try to make it work. As it turned out, this was too ambitious, and I decided to continue with replicating Bogaerts (1997)'s second study.

Procedures

In order to establish similar variety in the sentences and words used for my replication study I consulted literature on Finnish pronunciation to verify the knowledge I already had due to my Bachelor's study. I verified my findings with former teachers (Cornelius Hasselblatt and Marja-Leena Hellings) and consulted them for additional phones I might have missed.

For a similar text I found a short Finnish text from a Finnish newspaper website, containing 72 words.

Each of the subjects had been asked to record the tests; unfortunately, there were some

complications with the spontaneous speech sample. It appeared to be too difficult for a lot of the non-natives. Also, due to the duration of the judging, after discussing this with my supervisor, I decided to only use the sentences. This is still similar to Bongaerts et al. (1997), as in his second study he only used sentences as well. I made a selection of the recorded sentences, but made sure all near-native recordings were taken into account in the judging forms, which I set up for

anonymous judging via Google forms. All judges (5 native Finnish speakers with no linguistic background) received the forms. The recordings were mingled, so that it would not be obvious which recording would be by a native speaker. This insured that Native, Near-Native and learners of Finnish were completely mixed up and prediction would not play a factor in judging and thus would not influence the outcome. Bongaerts et al. (1997) did the exact same thing, as described in the section about the second study.

Every sentence was recorded three times in a row and from the last two I selected the sentence with the best pronunciation. This eliminates most speech errors due to suddenly reading a sentence out loud. Each word was recorded only once. The participants were not allowed to practice any of the tests.

The judges graded each voice sample on a scale from 1 to 5:

- 1 meaning the speaker has a very strong accent and is definitely non-native - 5 meaning the speaker has no foreign accent at all and is definitely native.

(19)

19 Design and analyses

All the recordings were set up for anonymous judging via google forms. The form was divided into six parts and five test recordings in order to get the judges used to the recordings.

Every part consisted out of fifteen to twenty recordings: - Part 1: 15 - Part 2: 15 - Part 3: 15 - Part 4: 20 - Part 5: 19 - Part 6: 19

In between every part the judges were instructed to take a small break in order to make sure that they did not lose their concentration due to judging the recordings becoming a drag. After the recordings were judged, I worked out an excessive Excel file (Appendix 4) in which I registered which subject had been given which grade for each sentence. The subjects are named by number in order to keep their name anonymous. I calculated the total amount of each score (1 to 5) that had been given to each subject. And then calculated the percentage a subject was scored as native. In order to define that I used the sum of 1 and 2 as non-native total and the sum of 4 and 5 as native total. I divided the native total by the total amount of ones, twos, fours and fives given to a single subject over all their judged recordings which defined the percentage of Nativelikeness for that subject. For example: Subject number 10, which is in the near native speaker’s group, had received the following scores on judging (sum of the score from all five judges over all 10 sentences: - 1: 1 time - 2: 15 times - 3: 8 times - 4: 12 times - 5: 13 times

I left out the threes as they do not give an indication for native or non-native. That leaves 1 + 16 = 17 times non-native score and 12 + 13 = 25 times native score. From a total of 50 – 8 = 42 scores. 42 divided by 25 scores this subject a nativelike level of 60%. (appendix 5).

(20)

20 I analysed the scores of each participant in this way and then calculated the average

Nativelikeness level per group by adding up their Nativelikeness level and dividing this by the number of participants in a group. So, for the learners group made: 6 + 21 + 18 + 71 + 5 + 0 + 0 + 0 = 121 / 8 = 15,125%. By doing this for each group, I got an average Nativelikeness level per group, and not only per person.

Results

This study yielded the following results, in tables 2 you will find the average Nativelikeness score of the learners participant group. The participants are numbered in order to preserve their anonymity. In table 3 you will find the Nativelikeness score of the Near-Natives and in table 4 the Nativelikeness score of the native speaking control group is listed.

Table 2: Score Nativelikeness (NL) learners

Learners Score NL 3 6% 6 21% 11 18% 12 71% 15 5% 19 0% 22 0% 24 0% Average score NL: 15%

Table 3: Score Nativelikeness (NL) near-natives

Near-natives Score NL 2 61% 5 56% 7 29% 8 49% 10 60% 17 22% Average score NL: 46%

(21)

21 Table 4: Score Nativelikeness (NL) natives

Natives Score NL 4 83% 13 100% 18 100% 23 100% Average score NL: 96%

The results show that not even the native speakers scored a 100% from the judging. It must, however, be noted that the native speaker which did not score a 100% level of Nativelikeness is from a different part of Finland than the other native speakers.

The Nativelikeness level of the learners’ and near-natives’ groups varies a lot. The learners are picked out really easily apart from one. Number 12 even has a higher Nativelikeness level rating than any of the near-native speakers.

The near-natives’ results vary enormously. Most of the near-natives have a Nativelikeness level of around or above 50%, meaning that most of the time, they will not be picked out as foreigners. Two of the near-native speaker group were clearly picked out as non-natives, this raises the question whether they were placed in the wrong group and should have been placed in the learners group instead.

Appendix 5, however, does show that participant number 2 got similar ratings for sentences 4 and 10 as the native speaking control group. Whilst multiple speakers (e.g. numbers 5 and 8) managed to get similar or even better ratings than number 4, who is a native speaker, for sentence number 7.

Discussion/Conclusion

This study was a replication study of Bongaerts et al. (1997), which focussed on ultimate

attainment in pronunciation of Dutch learners of English with a late age of onset. Replication of this study was interesting as Dutch, and English are both part of the Indo-European language family and are therefore closely related. Another factor was that one of the demands for the participants of this study was that they had not been overly exposed to English before the age of 12. As explained in this paper it seems improbable that all subjects had been excluded from English during their childhood. Finnish language, however, is part of the Finno-Ugric language

(22)

22 family and therefore unrelated to Dutch. It is uncommonly heard in the Netherlands and therefore it proved fairly easy to find participants that had not been overly exposed to Finnish before the age of 12.

The study consisted of three groups of participants: Learners of Finnish (average level), highly successful learners of Finnish (near-natives) and a native speaking control group. During the study participants were asked to record several sentences, which were then scored by native speaker on a scale from 1 to 5. 1 being definitely non-native, 5 being definitely native. On average the learners of Finnish scored between 6% and 71%, the highly successful learners between 22% and 61% and the native speaking control group between 83% and 100%.

Looking at the results per sentence, see Appendix 6, it is shown that some sentences were more difficult than others (see Table 5). On average, sentences 10 and 1 were pronounced best and 8 and 2 worst. Sentence 2 is difficult for non-natives, especially due to the last word in the sentence: “periaatteestaan”, as it as many double phonemes after each other, which all need to be

pronounced. Something that was predicted in my hypothesis. In sentence 8 there are several difficult aspects: “yrittänyt” has ‘y’ which sounds like the Dutch letter ‘u’ and can be confusing. Another difficult aspect of the word “yrittänyt” is the ‘r’. In Finnish it is articulated in the alveolar section of the mouth, whilst in Dutch it is formed in the uvular section. In the beginning of the word, for example “rakastaa” in sentence 10, the alveolar ‘r’ is more easily formed, but this study found that in the middle of a word it becomes more difficult. Another difficult part of sentence 8 is “Sallalle”, again the multiple double consonants, but also the consistency of Finnish pronunciation, which appears to be very difficult for Dutch speakers.

Sentence 10, which is a question, appeared to be less difficult than predicted in the hypothesis of this study. Non-natives appear to have difficulties with putting the stress of a longer word on the first syllable, but keeping the pronunciation of a sentence, especially in a question, monotonous showed to be less complicated than expected. However, a few participants, e.g. participant number 24, did pronounce sentence number 10 as this study predicted and raised the pitch at the end of the sentence, the way Dutch speakers do whenever a question is asked.

Another of this study's hypotheses is that, even though certain letters have different sounds in Finnish than in Dutch, this should not prove to be very difficult for the non-natives. All these sounds are familiar sounds in Dutch, though they have different spelling. Sentence 10 was, as

(23)

23 predicted, clearly the easiest sentence to pronounced. Sentence 10 and 1, the easiest sentences of this study, include ‘u’ and ‘y’, pronounced in Dutch as ‘oe’ and ‘u’ and sentence number 4 includes ‘ö’ which resembles the Dutch unstressed ‘e’. Sentence number 5 is about halfway on the difficulty scale, which matches the prediction of this study that the pronunciation of ‘o’ in Finnish being similar, but slightly different from the Dutch pronunciation would appear difficult for some learners. Also, the imperative sentence, number 7, proved to be challenging for the non-natives. Recordings showed that the predictions of this study came true and the pitch of most of the non-natives rose during the first syllable of the word ‘kirja’, alongside with the difficulty of pronouncing the Finnish ‘r’ this made this a difficult sentence for them to pronounce.

Table 5: Ranking of difficulty of sentences. 1 being the sentence pronounced best on average, 10 being the sentence pronounced worst on average.

1 10. Kuka rakastaa tätä kaunista punaista kukkaa? 2 1. Kyllä, Kytäjä on Hyvinkäällä sijaitseva kylä. 3 3. Harri on ylpeä hänen kauniista pikkusiskostaan. 4 4. Kissa hyppäsi pöydältä tuolille.

5 5. Joonas uskoo, että Suomesta tulee maailmanmestari. 6 6. Olisin mieluummin kotona, kuin täällä.

7 7. “Anna kirja!”, Essi huusi. 8 9. Laura valitsee että hän ei valita

9 8. Olen yrittänyt soittaa Sallalle koko aamu. 10 2. Hän ei halunnut luopua periaatteestaan

Above results imply that it is possible to achieve a native level of pronunciation of a second language with a late age of onset. It is, however, uncommon. Though even a native speaker might be considered a foreigner, if they have a different accent, foreigners, even those who have been living in Finland for years, are mostly picked out easily.

Of course, this was a rather small-scale replication of Bongaerts et al. (1997) original study, but it still gives a clear perspective on Second language acquisition with a late age of onset. The near-native speakers, were all considered to be highly proficient speakers of their second language and had all been living with and amongst Finnish speakers, using the language as their main language to communicate in. It was surprising to see that only half of them were rated a Nativelikeness percentage of above 50% and that one third even scored under 30%. Whilst one of the learners,

(24)

24 who I had expected to be picked out immediately, was scored 71% on Nativelikeness. Admittedly, the rest of the learners did get picked out as foreigners without (hardly) any doubt by the judges.

The original study shows that some, but not all, recordings of the highly successful learners

received the same ratings as the native speaker control group. This replication study shows similar results, however in a much smaller form. However, in the original study there were a few highly successful learners how managed to convince the judges that they were native speakers.

Unfortunately, this did not come true for this replication study. I therefore conclude that it is, difficult, but possible to achieve a nativelike level in the pronunciation of a second language with a late age of onset, but it is more likely achieved in a language closely related to your mother

tongue.

The goal of this replication study was to see whether the fact that Dutch is closely related to English and that Dutch learners of English, unavoidably, must have had at least some exposure to the English language before age 12 is of influence on the initial study on whether it is possible to achieve ultimate attainment in second language acquisition with a late age of onset. The study did not show results very different from Bongaerts (1997), but it does show different struggles for participants in e.g. pronunciation of every letter in Finnish versus not pronouncing certain letters in English words. This study also mentioned the method the participants had used on order to learn Finnish, it would be interesting to research whether certain learning methods are more efficient than others and to look at the variables Piske et al. (2001) mentioned. I expect that language aptitude and motivation will make an important difference in second language

acquisition. If one is driven to actually sound native, they will work harder on focussing on how a language should be pronounced than one that is merely interested in learning the language with the goal to be able to communicate in that language, not caring if people will here that they are non-native.

The Critical Period Hypothesis states that it should not be possible to achieve ultimate attainment in second language acquisition if age of L2 learning commences after puberty. However, this study, like Bongaerts (1997), Piske et al. (2001), Molis & Birdsong (2001), and Moyer (1999), shows that it is indeed possible to achieve nativelike attainment. Even DeKeyser (2000), who is a supporter of the Critical Period Hypothesis, finds nativelike attainment from his study. Participants in these studies that are found to have achieved nativelike attainment are often put away as “exceptions”. However, it is remarkable that “exceptions” are found in every study on Second Language

(25)

25 Acquisition with a Late Age of Onset. To conclude, if exceptions are found in all the studies done on late second language learning, can we still call them exceptions?

I therefore argue that the Critical Period Hypothesis is not as black and white as it is often interpreted. The amount of exceptions found in numerous studies proves that it is possible to achieve Ultimate Attainment with a late Age of Onset. These exceptions are often called super exceptional learners, whilst I argue that their success might be caused by other factors than just age. It has been proven in multiple studies that it becomes more difficult to learn a new language once one gets older, but not impossible. I argue that there are multiple variables worth research that could play a part in the success of second language acquisition. Variables such as, motivation, feedback, language aptitude, L1 use, L1 – L2 pairing, length of residence in L2 speaking country and formal instruction. I expect that more research on these variables will shed new light on the Critical Period Hypothesis.

(26)

26 References

Abondolo, D. (1998). The Uralic languages. London: Routledge.

Bakker, J.J. (1987). De woordenaar, tien redenen. Onze taal, 56, 73.

Birdsong, D. (1999). Second Language Acquisition and the Critical Period Hypothesis. London: Lawrence Erlbaum Associates.

Birdsong, D., Molis, M. (2001). On the evidence for maturational constraints in second-language acquisition. Journal of Memory and Language, 44, 235-249.

Bhela, B. (1999). Native language interference in learning a second language: Exploratory case studies of native language interference with target language usage. International Educational Journal, 1(1) 22-31.

Bongaerts et al. (1997), T., Summeren, van C., Planken, B., & Schils, E. (1997). Age and ultimate attainment in the pronunciation of a foreign language. Studies in Second Language Acquisition, 19(04) 447-465.

DeKeyser, R.M. (2000). The robustness of critical period effects in second language acquisition. SSLA, 22, 499 - 533.

Flege, J.E., MacKay, I.R.A., Piske, T. (2001). Factors affecting degree of foreign accent in an L2: a review. Journal of Phonetics, 29, 191-215.

Gijsbers, I., Gerritsen, M., Korzilius, H. & van Meurs, F. (1998). Engels in Nederlandse TV- reclame. Onze taal, 67, 175-177.

Haas, M. R. (1969). The prehistory of languages. Paris: Mouton.

Hasselblatt, C. (2000). De boom van de Finoegristiek. Maastricht: Shaker publishing B.V.

Jackson, A.I. (2009). Finse cultuur en taalstructuur in vergelijking tot de Nederlandse cultuur en taalstructuur. Not published.

(27)

27 Karlsson, F. (1983). Finnish an essential grammar. Helsinki: WSOY.

Korpela, J. (2010, April 28). Finnish pronunciation. Retrieved from

https://www.cs.tut.fi/~jkorpela/finnish.pronunciation.html

Moyer, A. (1999). Ultimate attainment in phonology. The critical factors of age, motivation, and instruction. SSLA, 21, 81 - 108.

Muñoz, C. & Singleton, D (2011). A critical review of age-related research on L2 ultimate attainment. Language Teaching, Volume 44, Issue 1, 1-35.

Ruhlen, M. (1994). The origin of language. New York, NY: John Wiley & Sons, Inc.

Scovel, T. (1988). A time to speak. A psycholinguistic inquiry into the critical period for human speech. Rowley, MA: Newbury House.

Smith, B & Swan, M. (2001). Learner English: A teacher’s guide to interference and other problems. Cambridge: University Press (Second edition).

Swadesh, M. (1972). The origin and diversification of language. London: Routledge & Kegan Paul

White, L. (1989). Universal grammar and second language acquisition. Amsterdam/Philadelphia: John Benjamin’s Publishing Company

(28)

28 Appendix 1

Harvinainen paksujalka liiteli Kuusamon Kantokylään

Pohjois-Suomessa paksujalka on vieraillut viimeksi Kemissä vuonna 2004 ja Siikajoella vuonna 1999.

Suomessa harvoin nähty kahlaajalintu paksujalka tutkii kesän tuloa Kuusamossa. Lintu havaittiin Kuusamon taajaman itäpuolella Kantokylässä kuluvan viikon alussa.

Suomessa paksujalka nähtiin viimeksi Porissa kaksi vuotta sitten. Pohjois-Suomessa paksujalka on vieraillut viimeksi Kemissä vuonna 2004 ja Siikajoella vuonna 1999.

Lintu pesii harvakseltaan pääsääntöisesti Keski- ja Etelä-Euroopassa. Se asustelee myös Afrikassa ja Lounais-Aasiassa.

Laji on vähentynyt viime aikoina.

Appendix 2

11. Kyllä, Kytäjä on Hyvinkäällä sijaitseva kylä. 12. Hän ei halunnut luopua periaatteestaan

13. Harri on ylpeä hänen kauniista pikkusiskostaan. 14. Kissa hyppäsi pöydältä tuolille.

15. Joonas uskoo, että Suomesta tulee maailmanmestari. 16. Olisin mieluummin kotona, kuin täällä.

17. “Anna kirja!”, Essi huusi.

18. Olen yrittänyt soittaa Sallalle koko aamu. 19. Laura valitsee että hän ei valita

(29)

29 Appendix 3 1. käsi (hand) 2. kesä (summer) 3. löyly (steam) 4. vahva (strong) 5. laki (law) 6. käyttää (to use) 7. pussi (bag)

8. röyhkeys (insolence) 9. matto (mat)

10. häiritä (to disturb) 11. koko (size)

12. rakastaa (to love) 13. lakki (cap) 14. bussi (buss) 15. pöytä (table) 16. yhteys (shared) 17. ovikello (doorbell) 18. verhot (curtains) 19. mato (worm) 20. kokko (bonfire)

21. ruokaöljyä (cooking oil) 22. vesi (water)

23. aurinko (sun) 24. saari (island) 25. meri (sea)

(30)

30 Appendix 4

Part 1 Participant Sentence 1 2 3 4 5

1 10 10 0% 20% 20% 20% 40% 2 24 1 80% 20% 0% 0% 0% 3 12 4 0% 20% 0% 80% 0% 4 4 2 0% 0% 40% 20% 40% 5 19 6 80% 0% 20% 0% 0% 6 8 3 0% 0% 40% 40% 20% 7 22 7 0% 60% 40% 0% 0% 8 3 9 20% 60% 0% 20% 0% 9 23 5 0% 0% 0% 0% 100% 10 13 8 0% 0% 40% 20% 40% 11 2 5 0% 20% 20% 20% 40% 12 6 2 60% 40% 0% 0% 0% 13 7 3 0% 40% 40% 0% 20% 14 10 9 20% 20% 20% 20% 20% 15 17 1 0% 0% 20% 20% 60%

1 5 1 0% 40% 20% 0% 40% 2 17 4 20% 40% 20% 20% 0% 3 11 10 0% 40% 40% 20% 0% 4 7 6 0% 20% 20% 20% 40% 5 8 7 0% 40% 0% 20% 40% 6 18 3 0% 0% 20% 20% 60% 7 2 8 0% 40% 0% 60% 0% 8 17 9 40% 0% 40% 20% 0% 9 15 7 60% 20% 20% 0% 0% 10 10 2 0% 40% 20% 0% 40% 11 3 3 80% 0% 20% 0% 0% 12 5 8 20% 40% 40% 0% 0% 13 7 5 0% 80% 20% 0% 0% 14 19 3 80% 20% 0% 0% 0% 15 2 4 0% 20% 0% 60% 20%

1 4 7 0% 40% 0% 20% 40% 2 8 1 0% 20% 0% 60% 20% 3 17 2 60% 40% 0% 0% 0% 4 11 1 40% 40% 0% 20% 0% 5 5 10 0% 20% 40% 40% 0% 6 10 5 0% 20% 20% 60% 0% 7 6 8 60% 20% 0% 20% 0% 8 7 10 0% 40% 20% 40% 0% 9 8 9 40% 40% 0% 20% 0% 10 2 2 0% 40% 40% 20% 0% 11 18 6 0% 0% 0% 20% 80%

(31)

31 12 17 5 40% 40% 20% 0% 0% 13 5 7 0% 0% 40% 20% 40% 14 24 10 60% 20% 20% 0% 0% 15 8 4 0% 20% 40% 20% 20%

1 10 1 0% 40% 0% 20% 40% 2 17 3 60% 40% 0% 0% 0% 3 22 5 60% 40% 0% 0% 0% 4 23 9 0% 0% 0% 0% 100% 5 2 10 0% 0% 0% 20% 80% 6 8 8 40% 40% 20% 0% 0% 7 15 2 80% 20% 0% 0% 0% 8 7 4 0% 40% 40% 0% 20% 9 5 6 0% 60% 0% 40% 0% 10 2 9 0% 20% 40% 40% 0% 11 12 10 0% 20% 0% 40% 40% 12 13 9 0% 0% 0% 0% 100% 13 17 7 100% 0% 0% 0% 0% 14 10 6 0% 40% 0% 20% 40% 15 3 5 40% 40% 20% 0% 0% 16 5 4 0% 40% 40% 20% 0% 17 8 5 20% 40% 0% 40% 0% 18 6 4 40% 20% 20% 20% 0% 19 5 5 0% 20% 0% 40% 40% 20 8 2 20% 40% 0% 40% 0%

1 7 8 40% 40% 20% 0% 0% 2 2 1 0% 40% 20% 40% 0% 3 4 4 0% 20% 0% 0% 80% 4 11 8 60% 20% 20% 0% 0% 5 8 10 0% 20% 20% 0% 60% 6 2 7 20% 80% 0% 0% 0% 7 19 8 100% 0% 0% 0% 0% 8 10 7 0% 60% 0% 20% 20% 9 5 2 20% 40% 20% 0% 20% 10 24 5 20% 60% 20% 0% 0% 11 17 6 40% 40% 20% 0% 0% 12 15 8 80% 20% 0% 0% 0% 13 7 1 0% 40% 20% 20% 20% 14 12 1 0% 40% 20% 20% 20% 15 10 3 0% 20% 40% 20% 20% 16 17 8 80% 20% 0% 0% 0% 17 13 5 0% 0% 20% 20% 60% 18 7 9 60% 20% 0% 20% 0% 19 18 10 0% 0% 0% 20% 80%

(32)

32

1 22 4 100% 0% 0% 0% 0% 2 7 2 20% 60% 0% 0% 20% 3 2 6 0% 20% 40% 40% 0% 4 23 1 0% 0% 0% 20% 80% 5 10 8 20% 20% 20% 40% 0% 6 3 2 40% 40% 20% 0% 0% 7 5 3 0% 20% 0% 20% 60% 8 17 10 20% 0% 60% 20% 0% 9 19 1 100% 0% 0% 0% 0% 10 8 6 0% 60% 20% 20% 0% 11 2 3 0% 20% 20% 40% 20% 12 11 6 80% 0% 0% 20% 0% 13 5 9 0% 20% 20% 40% 20% 14 17 1 20% 0% 20% 40% 20% 15 4 10 0% 0% 20% 0% 80% 16 10 4 0% 20% 20% 20% 40% 17 6 9 40% 20% 0% 20% 20% 18 7 7 40% 40% 0% 20% 0% 19 15 6 60% 20% 0% 20% 0% Appendix 5

Numbers 3, 6, 11, 12, 15, 19, 22 and 24 belong to the learners group Numbers 2, 5, 7, 8, 10 and 17 belong to the near-native group Numbers 4, 13, 18 and 23 belong to the native group

Participant Sentence 1 2 3 4 5 Sum 1+2 Sum 4+5 Total NL

2 5 0 1 1 1 2 2 8 0 2 0 3 0 2 4 0 1 0 3 1 2 2 0 2 2 1 0 2 10 0 0 0 1 4 2 9 0 1 2 2 0 2 1 0 2 1 2 0 2 7 1 4 0 0 0 2 6 0 1 2 2 0 2 3 0 1 1 2 1 1 15 9 17 8 16 25 41 61%

3 9 1 3 0 1 0

3 3 4 0 1 0 0

3 5 2 2 1 0 0

3 2 2 2 1 0 0

(33)

33 Participant Sentence 1 2 3 4 5 Sum 1+2 Sum 4+5 Total NL

4 2 0 0 2 1 2

4 7 0 2 0 1 2

4 4 0 1 0 0 4

4 10 0 0 1 0 4

0 3 3 2 12 3 15 18 83%

5 1 0 2 1 0 2 5 8 1 2 2 0 0 5 10 0 1 2 2 0 5 7 0 0 2 1 2 5 6 0 3 0 2 0 5 4 0 2 2 1 0 5 5 0 1 0 2 2 5 2 1 2 1 0 1 5 3 0 1 0 1 3 5 9 0 1 1 2 1 2 15 11 11 11 17 22 39 56%

6 2 3 2 0 0 0

6 8 3 1 0 1 0

6 4 2 1 1 1 0

6 9 2 1 0 1 1

10 5 1 3 1 15 4 19 21%

7 3 0 2 2 0 1 7 6 0 1 1 1 2 7 5 0 4 1 0 0 7 10 0 2 1 2 0 7 4 0 2 2 0 1 7 8 2 2 1 0 0 7 1 0 2 1 1 1 7 9 3 1 0 1 0 7 2 1 3 0 0 1 7 7 2 2 0 1 0 8 21 9 6 6 29 12 41 29%

8 3 0 0 2 2 1 8 7 0 2 0 1 2 8 1 0 1 0 3 1 8 9 2 2 0 1 0 8 4 0 1 2 1 1 8 8 2 2 1 0 0 8 5 1 2 0 2 0

(34)

34

8 2 1 2 0 2 0

8 10 0 1 1 0 3

8 6 0 3 1 1 0

6 16 7 13 8 22 21 43 49%

10 10 0 1 1 1 2 10 9 1 1 1 1 1 10 2 0 2 1 0 2 10 5 0 1 1 3 0 10 1 0 2 0 1 2 10 6 0 2 0 1 2 10 7 0 3 0 1 1 10 3 0 1 2 1 1 10 8 1 1 1 2 0 10 4 0 1 1 1 2 2 15 8 12 13 17 25 42 60%

11 10 0 2 2 1 0

11 1 2 2 0 1 0

11 8 3 1 1 0 0

11 6 4 0 0 1 0

9 5 3 3 0 14 3 17 18%

12 4 0 1 0 4 0

12 10 0 1 0 2 2

12 1 0 2 1 1 1

0 4 1 7 3 4 10 14 71%

13 8 0 0 2 1 2

13 9 0 0 0 0 1

13 5 0 0 1 1 3

0 0 3 2 6 0 8 8 100%

15 7 3 1 1 0 0

15 2 4 1 0 0 0

15 8 4 1 0 0 0

15 6 3 1 0 1 0

14 4 1 1 0 18 1 19 5%

17 1 0 0 1 1 3

17 4 1 2 1 1 0

(35)

35 17 2 3 2 0 0 0 17 5 2 2 1 0 0 17 3 3 2 0 0 0 17 7 5 0 0 0 0 17 6 2 2 1 0 0 17 8 4 1 0 0 0 17 10 1 0 3 1 0 17 1 1 0 1 2 1 24 11 10 6 4 35 10 45 22%

18 3 0 0 1 1 3

18 6 0 0 0 1 4

18 10 0 0 0 1 4

0 0 1 3 11 0 14 14 100%

19 6 4 0 1 0 0

19 3 4 1 0 0 0

19 8 1 0 0 0 0

19 1 1 0 0 0 0

10 1 1 0 0 11 0 11 0%

22 7 0 3 2 0 0

22 5 3 2 0 0 0

22 4 1 0 0 0 0

4 5 2 0 0 9 0 9 0%

23 5 0 0 0 0 1

23 9 0 0 0 0 1

23 1 0 0 0 1 4

0 0 0 1 6 0 7 7 100%

24 1 4 1 0 0 0

24 10 3 1 1 0 0

24 5 1 3 1 0 0

(36)

36 Appendix 6

Table 6: Results per sentence Participant 3 (learner)