• No results found

The pronunciation of foreign vowels in English loan words in Dutch

N/A
N/A
Protected

Academic year: 2021

Share "The pronunciation of foreign vowels in English loan words in Dutch"

Copied!
42
0
0

Bezig met laden.... (Bekijk nu de volledige tekst)

Hele tekst

(1)

1

The pronunciation of foreign vowels in English loan

words in Dutch

Name: Kelvin Finkers Student number: S3219968

First supervisor: dr. C.S. Gooskens Second supervisor: dr. R.M. Knooihuizen Date of second draft: 22-06-2020

MA thesis, Department of European Linguistics, Faculty of Arts, Rijksuniversiteit Groningen

(2)

2 Table of contents

- Abstract 3

- Introduction 3

- Literature 4

o Loan words and their incorporation into Dutch 4 ▪ Loan words in Dutch through the ages 5

▪ English loan words in Dutch 7

o Vowel Systems 9

▪ Northern Dutch vowels 10

▪ Received Pronunciation and General American vowels 11 o Pronunciation of loan words and L2 pronunciation 14

o Statement of purpose 15

- Methodology 16

o Participants 16

o Materials 17

o Procedures 20

o Design and analyses 21

- Results 21 - Discussion 29 - Conclusion 37 - Bibliography 39 - Appendices 41 o Appendix 1 41 o Appendix 2 42

(3)

3 Abstract

This study investigates the pronunciation of foreign vowels in English loan words that are introduced in the Dutch language. After an analysis of the Dutch and English vowel systems, the formant values of the four vowels that are present in Received Pronunciation English but absent in Dutch (/æ/, /ʊ/, /ʌ/ and /ɒ/) were measured. To accomplish this, a recording task was created that featured 32 English loan words, of which 24 contained at least one of the four foreign vowels. Visualisations of the

measurements suggested that the time since the introduction of the loan word had a pronuncial effect on 3 of the vowels (all except /æ/), most of which was on the first formant. However, none of these effects were significant. Furthermore, the age of the speaker showed a significant result on the pronunciation of /ʊ/ and a smaller effect on the pronunciation of /æ/ and /ʌ/. The influences of speaker age were all on the second formant value. The effect for two of these vowels, /ʊ/ and /æ/, were significant. Nonetheless, contrary to our expectations, none of the vowels followed the exact predicted pattern of the young speakers with recent loan words being the closest to the English vowel and older speakers with older loan words being closest to the Dutch vowel, with the other categories lying somewhere in between.

Introduction

The use of loan words, words from one language that are introduced into another without translation, is universal to almost every language. One of the world’s

languages that has played a large part in both the incorporation of other’s loan words and the spread of its own words into other languages, is Dutch. The clearest example of a language that has borrowed many Dutch loan words is Indonesian. During

colonial times, Indonesian has incorporated many words from the Dutch language, including words such as tante (‘aunt’), handuk (from handdoek, ‘towel’), kantor (from kantoor, ‘office’) and korupsi (from corruptie, ‘corruption’). Maier (2005) even argues that as much as 20% of words in Indonesian might be of Dutch origin. On the other side, Dutch has also acquired a part of its vocabulary both from language which border its language area (for example German aha-erlebnis (‘eureka moment’) and umlaut, French bureau (‘desk’) and paraplu (‘umbrella’)) and languages that are distant from it (such as Japanese karaoke and Javanese pienter (from Javanese pinter, ‘smart’).

One of the most closely related languages to Dutch is the English language, and, being relatively close to the area where Dutch is predominantly spoken, English has been one of the languages with which Dutch has had the most contact over the years. First it was used in trade with the English, and now it serves as a lingua franca all over the world for most people who do not speak each other’s languages. In the Netherlands, it is seen as a language which is useful to be able to speak, to such an extent that English is a compulsory subject throughout high school. Moreover, the Netherlands does not actively reject the use of English in parts of society such as entertainment. For example, as one of only a few countries in Western Europe, the Dutch do not dub their movies and tv shows, but rather add subtitles in their own language. These factors all add to the fact that the Dutch have the highest average understanding of English among all non-native English-speaking countries

(4)

4 As a result of the active contact between the two languages, Dutch has borrowed vocabulary from English into its own language over the span of many centuries, a process which was quite recently only increased because of the common use of English in technology and on the internet (Van der Sijs, 2012). One might wonder what happens to these loan words once they have been introduced into the Dutch language. Are they used in the same manner as their original English forms, or are there morphological features that should be added to make them compatible to Dutch? And on the subject of pronunciation, does their pronunciation remain similar to the pronunciation that is used in English, or is their pronunciation altered to a more “Dutchified” form? This last question is especially interesting, as there are some sounds, both vowels and consonants, that are not present in the Dutch sound system. How do these sounds behave when they are borrowed into the Dutch language via loan words? Does the age of the speaker play a significant role? Are the pronunciations of more recently borrowed loan words more akin to English than their counterparts which have been in the Dutch language for multiple decades? Do we see different patterns for these variables when examining different sounds, and do any of these variables have more influence on their pronunciation than the others?

In this study, we will try to answer all of these questions, as we will investigate the pronunciation of these non-native sounds in English loan words in Dutch, or more specifically, the vowels in English that are absent from the Dutch vowel system. Literature

Loan words and their incorporation into Dutch

Durkin (2014) explains that all words have both a form and a meaning, and either of them can be borrowed by another language. He further states that some linguists restrict the sense of “loan words” to only include the borrowings which maintain both the form and (at least some aspect of) the meaning, but that a broader sense of the word also includes other categories such as loan translations, also known as

calques, where the meaning of the word is borrowed and the form of the word is translated into the borrowing language. A well-known example of this is the word skyscraper, which was borrowed by most of the European languages via loan translation, albeit not always literally. Calques of skyscraper include Dutch

wolkenkrabber and German Wolkenkratzer (litt. ‘cloud-scratcher’), Danish skyskraber (litt. ‘cloud-scraper’) and French gratte-ciel (litt. ‘scratches-sky’).

The Dutch language borrows a large part of its vocabulary from other languages, mostly from neighbouring languages such as German, French and English. Unlike some of the European language unions, such as the French Académie Française, the Dutch Language Union tries to encourage the use of Dutch in as many parts of society as possible, but do not actively repulse the introduction of foreign words into its language. It achieves this not by creating Dutch equivalents for loan word, but by adopting foreign words and assimilating them by applying Dutch grammar rules to them. Examples of this include the infinitive verb suffix -en in words such as

facebooken (‘to be on Facebook’), the past tense suffixes -te(n) and -de(n) in words such as racete (‘raced’) and streamden (‘streamed’ (pl.)), the diminutive suffix -je in words such as cupcakeje (‘little cupcake’) and the present perfect verb circumfix ge -t/d such as in Ik heb gefaxt (‘I have faxed’) (Onze Taal, 2011)

(5)

5 An ongoing debate among speakers is whether we need the amount of (mostly

English) loan words we have and use nowadays in Dutch. This is because some speakers worry about an anglicisation of the Dutch language when native Dutch words are also available. However, Bakker (1987) lists 10 reasons why these loan words have purpose in the Dutch language. Other than the fact that some words simply don’t have a Dutch alternative, he also states that borrowed words fulfil a need for variation in the language (for example: ciao and bye are both used as alternative farewells), a necessity to use euphemistic forms and a wish to use shorter forms, because Dutch words can get rather long due to compounding, while English has a lot of 1 or 2-syllable words.

How these loan words end up in the Dutch language is complicated, as there are many different outcomes, but Van der Sijs (2005) distinguishes four stages of adaptation that foreign words undergo in order to become integrated in a language:

1: Use of a foreign word by a specific social group, as a foreign element in the language

2: Broader use of the word in multiple contexts

3: Word is on the verge of integration thanks to widespread use, possible change of orthography, pronunciation and/or formation

4: Word is completely assimilated into the language, only recognised as loan word by linguists

Only during the second and third stages can the foreign word technically be considered a loan word, since it is just a foreign element in the first stage and is completely assimilated into the language in the fourth stage.

The boundaries of these stages are quite vague. There is not a certain point when a word is widely used enough to change from phase 1 to phase 2. Moreover, the transition from phase 2 to phase 3 is characterised by a change in spelling or pronunciation, but this does not occur for all loan words. However, especially the transition from phase 3 to phase 4 is fairly unclear, as, according to Poplack, Sankoff & Miller (1988), “there is no natural way of identifying a discrete point within this interval before which the word is not an integrated loan, and after which it suddenly becomes one.” (50)

Loan words in Dutch through the ages

Thus, loan words go through different stages in order to be incorporated into Dutch. But where did the borrowing of vocabulary from other languages into Dutch start? The first indication of loan words in Dutch can be traced back to the 8th-9th century AD, when the Dutch loaned its first words from Latin. These were generally religious words, such as duivel (‘devil’) from diabolus, monnik (‘monk’) from Vulgar Latin

monicus and bisschop (‘bishop’) from episcopus. Nevertheless, traces of earlier loans have been found in Dutch, but these words were descendants of Proto-Germanic forms that were already loaned from Greek, Latin and Celtic. Examples of these descendants include kerk (‘church’), ultimately from Greek kuriakón, wijk

(‘district/neighbourhood’), ultimately from Latin vicus, and ambacht (‘handcraft’), ultimately from Celtic ambaxtos (Van der Sijs, 2001).

(6)

6 In fact, in the earlier stages of Dutch, the large majority of loan words in the language came from (Vulgar) Latin words that were acquired by Germanic tribes, who turned them into their own Proto-Germanic forms. This becomes apparent when comparing the Dutch forms of some of these early loan words to their equivalents in other

Germanic languages. Words such as kalk (‘chalk’), peper (‘pepper’) and poort (‘port’), all have similar forms not only in the other West Germanic languages such as English and German, but also in the Scandinavian languages such as Swedish, Danish and Norwegian Bokmål, as a result of the fact that these were loaned in Proto-Germanic from Latin calx, piper and porta respectively. We also see some loan words that were only obtained by West-Germanic languages, such as zalm (‘salmon’), which has similar forms in English (salmon) and German (Salm), but not in the Scandinavian languages, as the divide between the Northern and Western Germanic languages started when the Roman Empire was at its peak, and the spread of the Latin word salmo only reached West-Germanic territory (Van der Sijs, 2005).

The first instance of words loaned from modern languages occured in the beginning of the 13th century, when Dutch acquired its first loan words from French, such as sint (‘saint’), proeven (‘to taste/to try’) and toren (‘tower’), which were loans from the Old French words seint, prover and tur, which would later become saint, prouver and tour in contemporary French. The first German loan words in Dutch can also be traced back to this period, when words such as louter (‘a bunch of’), Hanze (‘Hanseatic’) and gans (‘whole/entire’) were loaned into Middle Dutch from Middle High German lauter, Hanse and ganz (Van der Sijs, 2001).

Medieval Latin and French would remain the predominant sources of loan words in Dutch for the last centuries of the Middle Ages, with the occasional loan word from modern languages such as High German, Spanish, Italian or the Scandinavian languages. The large span of Latin loan words in Dutch is due to the fact that Latin was widespread in Europe in three waves: the first wave being the Roman

colonisation under the Roman Empire, the second wave being the result of the spread of Catholicism, mostly in the 7th century by missionaries, and the third wave being around the Renaissance at the end of the Middle Ages, when Latin once again became the language of science and art. (van Keymeulen, 2008) This means that the period of Latin linguistic influence spans more than 1500 years, and each wave brought a multitude of loan words to Dutch, first via Proto-Germanic and later directly into Old or Middle Dutch. However, that does not mean that Modern Dutch has not stopped borrowing words from Latin; even until the first half of the 20th century, Latin loan words such as cervix and mortuarium (‘mortuary/moruge’) were mentioned for the first time in Dutch (Van der Sijs, 2001).

(7)

7 Nevertheless, while Latin has contributed words to Dutch for more than a millennium, it has not been the most influential language to Dutch in terms of loan words. That title belongs to the French language, which, as we have stated before, has been contributing to Dutch since the 13th century. However, in this time period of 700 years, it is estimated that the French language has supplied the Dutch language with three times as many loan words as Latin (Van der Sijs, 2005). This extreme quantity of loan words compared to other neighbouring languages comes as a result of a centuries-long French influence in the Benelux region. Starting in the 9th century, it reached its peak in the 14th century, when large parts of the Low Countries fell under rule of the House of Burgundy. During this period, we can clearly see a steep

increase of French loan words in Dutch.

In the 17th century, French became the lingua franca among the elite in Europe, and remained the elite language for more than three centuries. This was also true for the Netherlands, where many loan words and loan translations (such as kleinzoon and kleindochter (‘grandson and -daughter’, from French petit-fils and petite-fille)) were introduced to the language of the lower classes via the elite. The turn of the 19th century saw another sharp increase in loan words, during the French period in the Low Countries under Napoleon. It is not coincidental that a large part of these words were administrative terms like kabinet (‘cabinet’) and unaniem (‘unanimous’). After the French period in the Low Countries ended, the language stayed the main language of the Dutch elite. However, the influence of French, and therefore the amount of new French loan words, decreased.

The origin of most loan words in Dutch after their independence from the French Empire has been quite diverse. However, there have been two languages that saw a steep incline of loaned words over a period of time. The first of these languages was Yiddish, whose first loan word occurred in Dutch in the 17th century, even though Dutch acquired most of its Yiddish loan words from 1850 to 1950, when it adopted hundreds of Yiddish words for everyday objects and ideas. Some of these words were obtained directly into common Dutch and are still used in everyday Dutch language, such as mazzel (‘luck’), jennen (‘to tease/irritate’) and stiekem (‘secretly’), but most of these Yiddish loan words were adopted into Bargoens, the variant of Dutch that was spoken in by the lower class at the time. The other language that has seen a surge in loaned words is English, which has been a primary provider of loan words to Dutch since the year 1800.

English loan words in Dutch

The known history of English loan words in Dutch spans almost 800 years, even though it is believed that the actual history of English loan words in the Dutch language spans almost 1200 years. This is the result of a difference between the introduction of a word and the first recording of a word. Although the first recorded English loan word in Dutch, ootmoed (‘humility’), was written down for the first time in 1265 (Van der Sijs, 2001), it is suspected that this word was already loaned from the English language during the Christianization of the Low Countries by Boniface and Willibrord in the 8th century AD, together with the word delgen (‘to void (a debt)’). Despite the fact that it was encountered for the first time in written form in 1477, it must have been introduced before the 10th century, when verdelgen (‘to exterminate’) was recorded. The root delgen had to have been introduced before the common prefix ver- was added.

(8)

8 After this period of English missionaries in the Low Countries, the influence of English on the Dutch language halted. In fact, it has been speculated by Blok (1891) that the amount of English influence on the Dutch language was nearly insignificant from the start of the second millennium up until the 1800s. A proposed reason by Blok for this was the lack of a land border between the Netherlands and England, and therefore, between the Dutch and English languages. Van der Sijs (2001) has confirmed this speculation. From the Middle Ages until the year 1800, she only found 81 words which were certain to be of English origin, mostly exotic animal names, such as albatros, alligator and jaguar, and early shipping and trading words, such as praaien (‘communicate at sea’) and pakketboot (‘ferryboat’). Shipping words formed a large part of the early influence of English on Dutch, for the reason that while there may not have been a land border between the Low Countries and England, the Dutch coastal cities did trade with English sailors, which is still visible in the fact that the dialects of these cities contain proportionally more English loan words than the dialects of inland cities (Van der Sijs, 2005)

After the independence of the Low Countries from the French Empire in 1815, the influence of French on the Dutch language halted, and therefore the number of French loan words introduced to Dutch started to decline. The English language seems to have filled its spot as the main source of loan words, starting with a

multitude of words associated with industry and the steam engine, as a result of the start of the Industrial Revolution in the beginning of the 19th century. Examples

include locomotief (‘locomotive’), rails and tunnel. Around the turn of the 20th century, three main groups of loans can be distinguished: fashion, where names for items of clothing such as shirt, petticoat and sweater were directly taken out of English, early vehicles as bus, tram and metro, and early sports and related terms, such as rugby, tennis and match. In the first decades of the 20th century, sport and games would become one of the main areas in which English loan words could be found, with the introduction of new sport names (badminton, volleyball), terms related to sports (ace, peddelen (‘paddle’)) and games such as bingo, bridge and kruiswoordpuzzel

(‘crossword puzzle’).

The fact that not all loan words remain in Dutch has become the most apparent during and shortly after the Second World War. During this period, the Dutch

language saw a lot of English influence from the English-speaking troops present in the Netherlands. Mostly military words but also words for everyday objects were borrowed into Dutch. De Vooys (1970) had already doubted the fate of these words in the Dutch language, writing “the use of which of these words is transitory and which of these words will remain in use, only time will tell” (206). His expectations were correct: while some of these have stood the test of time in Dutch (for example training, T-shirt and, to a lesser degree, headquarters, which was calqued into hoofdkwartier), others have completely disappeared from the Dutch vocabulary. Examples of this last category include words such as belt, ration, and permit. These words have since then been replaced by their Dutch equivalents riem, rantsoen and vergunning again.

(9)

9 In recent times, the overwhelming majority of new loan words in Dutch are of English origin. Research has shown that the number of English loan words increases every year, but on the other hand, that most of these recent English loan words are not used frequently in daily speech, with only a few recent loan words regarding new media and technology being used regularly (Van der Sijs, 2012). In contrast to loan words of Romance languages, whose percentage had stayed relatively the same, the percentage of loan words of English origin had nearly doubled from 2,3% in 1994 to 3,7% in 2012.1 Nowadays, many English loan words enter Dutch via the language that is spoken by younger speakers online, and their usage is very common on social media (Verheijen, 2016).

One of the questions that has been raised by researchers is if there is a specific manner that these loan words are pronounced and/or should be pronounced by speakers of Dutch. This mainly concerns the question if the words should be

pronounced more native-like or more similar to that of the language from which the word was loaned. Thus far, there has not been an agreed upon answer to this question of approaching foreign words, as the consensus seems to indicate that it should be the speaker themselves who should decide to either pronounce them in a Dutch fashion or pronounce them similarly to the original words (Posthumus, 1989). Vowel systems

Comparing the vowel systems of English and Dutch is not as straightforward as one might think. Both languages have multiple sound systems, with the key factor being the speaker’s region within the Sprachraum of the language. In Dutch, for example, two main standards can be found in Europe: Northern Standard Dutch, spoken in the Netherlands, and Belgian Dutch, spoken in the Belgian regions of Flanders and Brussels. Further differences can be found between the dialects within one of these two standards. English, being the de facto language in a plurality of nations around the world, has many standard forms, including Received Pronunciation (or “BBC English”), General American English, Canadian English, General Australian English, New Zealand English, etc., all of which are divided themselves into their own dialectal variants

To compare the vowel systems of the Dutch and English languages, we will lay our focus on the three variants that will be the most important for our study, in this case Northern Standard Dutch, as this study takes place in the northern part of the Netherlands, and both Received Pronunciation and General American English, as both seem to have a substantial influence on the Dutch language via for example television shows and social media.

1 One should take into account the fact that these percentages were the results of a comparison between one copy of the NRC Handelsblad in 1994 and one copy in 2012. The fact that only two copies of one newspaper were compared might have influenced the results.

(10)

10 Northern Dutch vowels

Northern Dutch has a set of 11 monophthongs, of which 2 are always long (/a:/ and the rare /œː/), 4 are always short (/ɪ/, /ʏ/, /ə/, and /ɑ/) and 5 have long and short variants (/i(:)/, /y(:)/, /u(:)/, /ɛ(:)/, and /ɔ(:)/). The three closed long vowels /i:/, /y:/ and /u:/ are only present as allophones of their shorter counterparts /i/, /y/ and /u/ before /r/ and, together with non-native long vowels /ɛː/, /œː/ and /ɔː/, in stressed syllables of (mostly French) loan words, such as centrifuge, cruise and oeuvre. Furthermore, the Northern Dutch vowel system contains 12 diphthongs, divided into four groups (Gussenhoven, 1992; Collins & Mees, 2003).

- Three phonemic diphthongs /ɛi/, /œy/ and /ɔu/ (or /ɑu/).

- Three diphthongs that are represented in Belgian Dutch by the long vowels /øː/, /e:/ and /o:/. These long vowels are not present in Dutch, and have the diphthongs /øʏ/, /ei/ and /ou/ (or /oʊ/) as equivalents.

- Three tense backing diphthongs that only occur in syllables ending in -uw: /e:u/, /yu/ en /iu/, which could also be considered to be realisations of the vowel-approximant clusters /iʋ/, /yʋ/ and /eːʋ/ (Booij, 1999).

- Three tense fronting diphthongs that only occur in syllables ending in a vowel + i: /o:i/, /a:i/ and /ui/ which could also be considered the realisations of the vowel-approximant clusters /o:j/, /a:j/ and /uj/.

An illustration of the place(s) of articulation of the Dutch monophthongs and non-vowel-approximant diphthongs in Northern Dutch can be found in figures 1 and 2. The formant values of its monophthongs can be found in table 1:

Figures 1 and 2: Vowel diagrams of the monophthongs (left) and diphthongs (right) present in Northern Standard Dutch (Gussenhoven, 1992)

(11)

11 Table 1: Average formant values of the monophthongs in Northern Standard Dutch (Adank, Van Hout & Smits, 2004; Koopmans-van Beinum, 1994)2 in Hz

F1 F2

Vowel Male Female Male Female

/i/ 278 294 2162 2524 /y/ 259 305 1734 1918 /u/ 259 286 805 938 /ɪ/ 361 399 1919 2276 /ʏ/ 366 417 1595 1813 /ɛ/ 475 535 1739 1990 /ə/ 405 approx. 4452 1439 approx. 16402 /ɔ/ 402 419 821 918 /a/ 670 912 1425 1572 /ɑ/ 758 578 1172 1280

Received Pronunciation and General American vowels

While the consonant systems of Received Pronunciation (RP) and General American (GA) are almost identical, their respective vowel systems contain the majority of the phonological differences between the two variants. Not only is a different set of vowels used by each variant, but the places in which certain vowels are used do not overlap either.

The vowel system present in Received Pronunciation has 12 monophthongs (with the vowel in words such as dress being pronounced as [e] in General Received

Pronunciation, mostly used by older speakers, and as [ɛ] in Advanced Received Pronunciation, mostly used by younger speakers (Schmitt, 2007)), whereas the system used for General American has 9 to 11. This is because /ɔ/ is only

distinguished by some speakers while falling together with /ɑ/ by others (Labov, Ash & Boberg, 2006), and /ɜ/ is only considered as a contrastive vowel for some speakers of General American, whereas the majority of the population does not distinguish the few minimal pairs that exist between /ɜ/ and the other central vowel /ə/. For example, the words foreword and forward are pronounced as [fɔ:wɜ:d] and [fɔ:wəd] in Received Pronunciation, but both are pronounced as [fɔɹwəɹd] in General American. This

means that the two central vowels /ɜ/ and /ə/ are considered by certain linguists to be allophones in General American English (Wells, 1982).

2 Note that the rare non-native vowel /œː/ was not included in the study by Adank, Van Hout & Smits. However, as this vowel is also not present in English, it is not relevant to our study. /ə/ was also not measured in the article; its formant values are retrieved from the article by Koopmans-van Beinum (1994). Unfortunately, the article by Koopmans-van Beinum only has data for male speakers, but since the F1 values for most vowels is about 20 to 60 Hz higher for female speakers and the F2 values for most vowels is between 150 and 250 Hz higher for females, the same is also expected for /ə/.

(12)

12 The entire set of vowels in General American English overlaps with the vowels in Advanced Received Pronunciation, with the latter also containing an extra mid-open back vowel /ɒ/. However, although the monophthongs between these two variants mostly overlap with regard to their height and backness, but not their length. General American does not contain long vowels, whereas Received Pronunciation has five: /u:/, /i:/, /ɔː/, /ɜː/ and /ɑː/. An illustration of the place of articulation of the

monophthongs in both variants can be seen in figures 3 and 4. Values of the F1 and F2 of all English monophthongs in both Received Pronunciation and General

American can be found in tables 2 and 3

Figures 3 and 4: Vowel diagrams of the monophthongs in Received Pronunciation (left) (Roach, 2004)3 and General American (right) (Wells, 1982)4

Table 2: Average F1 and F2 values of English monophthongs in Received

Pronunciation (Deterding, 1997) and General American (Hillenbrand et al., 1995) in Hz

Received Pronunciation General American

F1 F2 F1 F2

Vowel Male Female Male Female Male Female Male Female

/i/ 280 303 2249 2654 342 437 2322 2761 /u/ 316 328 1191 1437 378 459 997 1105 /ʊ/ 379 410 1173 1340 469 519 1122 1225 /ɪ/ 367 384 1757 2174 427 483 2034 2365 /e/ or /ɛ/ 494 719 1650 2063 528 634 1944 2294 /ə/ or /ɜ/ 478 606 1436 1695 474 523 1379 1588 /ɔ/ 415 389 828 888 652 781 997 1136 /æ/ 690 1018 1550 1799 588 669 1952 2349 /ʌ/ 644 914 1259 1459 623 753 1200 1426 /ɒ/ 558 751 1047 1215 - - - - /ɑ/ 646 910 1155 1316 768 936 1333 1551

3 This vowel diagram shows the vowels used in General Received Pronunciation. As previously stated, in Advanced Received Pronunciation, /e/ should be replaced by /ɛ/.

4 /e/ and /o/ are often not considered monophthongs in General American, as they are allophones that belong to the same phonemes as the diphthongs /eɪ/ and /oʊ/ respectively, which will be discussed in another paragraph.

(13)

13 Nevertheless, the aspect in which the monophthongs in the two vowel systems of the variants differs the most is not in the amount of vowels in both systems, or the

specific vowels that are present. The largest difference lies in the fact that not all corresponding vowels are always used in the same words, especially with regard to open vowels and mid-open back vowels. To illustrate, an example would be /æ/, which is used in both variants in the word trap. However, the same vowel is also used in bath in General American, where it is replaced by /ɑː/ in Received Pronunciation. The shorter version /ɑ/ can be used in General American in words where Received Pronunciation would use /ɑː/ (such as in calm), /ɒ/ (as in pot) or /ɔː/ (as in caught). By some speakers of General American, words in the last and some in the second-to-last group can also be pronounced with /ɔ/, while others prefer /ɑ/. This is due to the so-called “cot-caught merger”, a phenomenon that in the United States mostly occurs west of the Mississippi river and in the north-eastern part of the country (Labov, Ash & Boberg, 2006)

In contrast to the monophthongs, the two variants do generally agree on both the diphthongs that are used and the words in which the same or corresponding diphthongs are present, as the four diphthongs /eɪ/, /aɪ/, /ɔɪ/, and /aʊ/ are used by both variants in the same words. However, Received Pronunciation contains 4 diphthongs that are either different or not present in General American:

- The General American diphthong /oʊ/ in words such as boat is equivalent to a central-starting /əʊ/ in Received Pronunciation

- Received Pronunciation contains 3 centring diphthongs ending in /ə/, which is caused by the fact that /r/ is not pronounced at the end of words in Received Pronunciation. Whereas the long vowels /ɔː/, /ɑː/ and /ɜː/ are not affected by the loss of the /r/, the shorter vowels /e/, /ʊ/ and /ɪ/ are changed to /eə/, /ʊə/ and /ɪə/ in words such as bear, pure and fear (Roach, 2004). In General American, this loss of the /r/ at word endings does not occur.

A diagram of the diphthongs in both variants can be found in figures 5 and 6

Figures 5 and 6: Vowel diagrams showing the diphthongs in Received Pronunciation (left) (Roach, 2004) and General American (right) (Wells, 1982)

(14)

14 When we compare the vowel systems of the two languages, we are able to get an insight into the vowels, mostly monophthongs, that might cause trouble not only in L2 pronunciation, but also in English loan words in Dutch. For example, when we

compare monophthongs of Northern Standard Dutch with those that are present in Advanced Received Pronunciation, four monophthongs in Received Pronunciation that are not present in Northern Dutch can be identified, namely /æ/, /ʊ/, /ʌ/ and /ɒ/. Some of these vowels are already known to cause production problems for speakers of English as a foreign language (EFL). In particular /æ/ (in the DRESS-TRAP

contrast) and /ʊ/ (in the GOOSE-FOOT contrast) are proven to create difficulties in their distinctive pronunciation vis-à-vis /ɛ/ and /u/ respectively (Šimáčková &

Podlipský, 2018). The fact that these four vowels are absent in the most used form of Dutch might not only indicate that these vowels could cause the most trouble for Dutch learners of English regarding their pronunciation of English, but it might also raise a dilemma in the Dutch language over English loan words that contain these vowels. Should it preserve the original vowels, or is it a more viable option to replace the vowels that are non-native to Dutch and replace them with vowels that are

already present in Dutch?

Pronunciation of loan words and L2 pronunciation

The pronunciation of loan words is related to second language pronunciation,

because in both instances, a speaker might struggle with sounds that are not present in their native sound system. This is true for both articulation and perception. A

speaker might not be able to pronounce a foreign sound and might not be able to differentiate two similar sounding sounds, only one of which is present in the native tongue. It has been proven that one of the main factors in accents in a second

language is the interference caused by differences in sounds between the L2 and the L1 (Weinreich, 1953). This interference caused by the difference between the two languages is also called “transfer”. However, a speaker’s struggle with L2

pronunciation differs from sound to sound. Flege (1995) states that an L2 speaker does not show difficulties when acquiring L2-sounds that are either identical to a sound in their native sound system, or sounds that are completely dissimilar to any sound in their native tongue, as they would not have any difficulties differentiating it from the other sounds in the L2. Instead, he suggests that the sounds that are the most challenging for an L2 speaker are those that are similar to sounds in the native sound system, but not identical. This is due to the fact that learners perceive and classify them as the equivalent sound in their own language.

(15)

15 Nevertheless, transfer of the L1 is not the only factor that causes differences in

pronunciation. It has been suggested that identity might also play a key role in the pronunciation of a second language (Rindal 2010, Zuengler 1988), and therefore, might also be also be a key factor in the pronunciation of loan words. This might be especially true for Dutch, given the fact that, as earlier stated, there is not a uniform manner to pronounce English loan words in Dutch (Posthumus, 1989). This could indicate that one group uses the Dutch pronunciation as a part of their identity, while the other group uses the pronunciation that is more similar to the English version of the loan word. One interesting divide to look at in the case of English loan words in Dutch would be the one regarding younger speakers versus older speakers, not only because of wanting to establish one’s identity as a part within or outside a certain group in society, but also because of the difference in exposure to the English

language. Considering that the youth in the Netherlands is exposed to a lot of English via social media, television and films, one might suspect that a they would pronounce loan words more similarly to English than the older generations.

Statement of purpose

The main research question in this thesis will be: to what extent does the

pronunciation of foreign English vowels change in words that were loaned into Dutch, and, if it changes, what has determined this difference? Sub-questions that are

contained within this main research question include:

- Is there a difference in the pronunciation of English loan words by younger speakers and older speakers?

- Is there a difference between the pronunciation of newer loan words as opposed to the pronunciation of more “established” loan words?

- Is there a difference between different sounds? (Do some sounds show no difference between the variables, while others do?)

The hypotheses for these sub-questions are as follows: it is expected that English loan words will be pronounced more similarly to English by younger generations, as they get more exposure to the English language and are more familiar with these words, since a great number of English loan words are linked to technology, with which the younger generation is more familiar. Furthermore, we predict that a pronunciation more similar to Dutch will be found for words that have already been implemented in the Dutch for at least a few decades, whereas the pronunciation of newer loan words will be more comparable to English. This is due to the assumption that older words have had more time to assimilate to the sounds in the Dutch vowel system, while this acclimatization has not yet occurred for newer words.

Lastly, it is also predicted that we will find different results for different English sounds. More precisely, we expect the differences between the different speaker ages and different word ages to be the least distinct for sounds that have a similar counterpart in Dutch but are not identical. This is in line with the Flege’s theory in L2 pronunciation (1995), which states that L2 learners experience the most troubles with sounds that are similar to their own language, but do not have an identical

counterpart in his/her native vowel system. It is predicted that these vowels with similar Dutch counterparts will still be affected by speaker age and word age, but that the difference between different age groups will be less distinct.

(16)

16 Methodology

This study will examine the dependent variable of the pronunciation of English loan words in Dutch, and more specifically, the formant values of the foreign vowels that are contained within these words. To be able to test which variables have an effect on the pronunciation, three main independent variables will be tested:

- Speaker age: a division will be made between young and old speakers of Dutch.

- Year of introduction or word age: a distinction will be made between loan words that were borrowed before 2000 and after 2000.

- Vowel: the results of 4 foreign vowels will be compared.

To test these variables, we have created a wordlist that was to be recorded by group of young speakers and a group of older speakers. Furthermore, a short questionnaire with questions on the participants’ use of English and their stances on English loan words in Dutch was filled in by all participants after they had recorded themselves. Participants

In total, 20 participants, 10 males (6 young and 4 old) and 10 females (4 young and 6 old), were recorded while pronouncing the loan words. The recordings of all

participants were made in April 2020. Unfortunately, during this period, measures were in place to reduce the spread of the SARS-CoV-2 virus, creating an effective lockdown of all non-necessary meetings, and therefore all contact between

researchers and participants for research purposes. This meant that we were forced to kindly ask the participants to record themselves, as we were not allowed to record them ourselves. This also meant that we were not able to record the participants in a recording room as they had to stay at home and had to find a quiet place to record themselves in their own houses.

The participants that were recorded for this study fall into two categories: a younger group and an older group. The younger group was around 20 years old at the time of recording, (N=10, M=20.9), with a minimum age of 20 and a maximum of 22 at the time of recording. They all had finished a HAVO- or VWO-level at high school5 and were currently either in the last year of their HBO or university bachelor studies, or in the first year of their master studies. The ages of the older group were more diverse, with a spread from 48 to 59 years of age at the time of recording (N=10, M=53.5), and had various highest levels of education, ranging from the LTS (lagere technische school, litt. ‘lower technical school’, a predecessor to the technical path of the VMBO-level in Dutch high schools) to an HBO bachelor’s degree. The majority of the

participants in both groups were from the same municipality in the province of Overijssel in the Netherlands.6

5 Equivalent to a GSCE-level (marks from A to C) + two to three advanced subjects in the United Kingdom. 6 Due to the outbreak of the SARS-CoV-2 virus, and the effective lockdown in most European countries, including the Netherlands, it was not possible to find and record people from various parts of the Netherlands, as we had to find our own participants among family, friends and acquaintances. This means that the

participant groups mostly consisted of people who had been in contact with the researchers before and persons recommended by other participants. While this does reduce a possible effect of dialects on our results, it might also mean that the results of this study are not representative for Dutch speakers in general.

(17)

17 Materials

As it was not allowed to record the participants in person, the research participants were sent a document containing some English loan words, with the task to record themselves while pronouncing these loan words. In total, 32 loan words were used, picked from an online list of English loan words and anglicisms in Dutch (Vos, 2014), whose dates of first appearance were subsequently checked in the Chronologisch Woordenboek (Van der Sijs, 2001) or the Online Van Dale dictionary (Van Dale Uitgevers, 2017). Of these words, 8 were filler items and 24 contained one or more of the vowels that we were planning to measure, divided over 4 groups of 6, one for each vowel that was tested. The tested vowels are the monophthongs that are present in Advanced Received Pronunciation but not in Northern Dutch. This group includes the following monophthongs: the near-open front unrounded vowel /æ/, the near-close back rounded vowel /ʊ/, the open-mid back unrounded vowel /ʌ/ and the open back rounded vowel /ɒ/. Within each group, another subdivision was made between older and younger loan words. In this research article “old” loan words will signify any English loan word that has been introduced into Dutch before the year 2000. We have also opted to not include any words that have a recorded use in Dutch before the end of the Second World War, because almost all of these loan words have been completely incorporated into the Dutch language, to the point where they are not recognisable as loan words anymore. The oldest old loan words in the assignment were bulldozer, hotdog and flashback, all of which were introduced in the 1950s (1950, 1951 and 1953, respectively). The youngest loan words in the older group of the task were content, lookalike and muffin, having been borrowed as late as 1994, 1992 and 1989, respectively. In order to not have too much discrepancy of word age between the vowels, all groups contain one old loan word that was

introduced in the 1950s, at least one word that was introduced in the 1980s or 1990s and one word that was introduced at some point between the first and last words in their respective groups.

The term “young” loan words in this article will signify all English loan words that were introduced into the Dutch language after the year 2000. These will mostly be

technological and online terms, the oldest of which present in the set is the filler word youtuben, which was loaned in 2007. The oldest young loan word present in any of the tested groups is goodiebag, which was introduced in 2009. The youngest loan word present in the task is pubquiz, which was only added to the Van Dale dictionary in 2017, three years before the publication of this thesis. All words that were tested, as well as their phonetic transcriptions, are shown in table 3. In this table, the years of introduction of all words are also shown, which, unless stated otherwise, are all mentioned in the Chronologisch Woordenboek (Van der Sijs, 2001)

(18)

18 Table 3: The words containing the vowels measured in the study, their introduction years and phonetic transcription, categorized by foreign vowel

/æ/:

Loan word Year of introduction Phonetic transcription of the borrowed English word (part)7 Flashback 1953 /ˈflæʃ.bæk/ Management 1970 /ˈmæ.nɪdʒ.mənt/ (RP) or /ˈmæ.nədʒ.mənt/ (GA) Cliffhanger 1981 /klɪf.ˈhæŋ.ə(ɹ)/ Hashtag 2011 /ˈhæʃ.tæɡ/

Whatsappen 2012 Blend of what’s /wɑts/ and

app /æp/

Capslocktoets* 2013 /ˈkæps.lɒk/

/ʊ/:

Loan word Year of introduction Phonetic transcription of the borrowed English word (part) Bulldozer 1950 /ˈbʊl.doʊ.zɚ/ Output 1970 /ˈaʊt.pʊt/ Lookalike 1992 /ˈlʊk.ə.laɪk/ Goodiebag* 2009 /ˈɡʊ.di.bæg/ E-book 2010 /ˈi.bʊk/ Pushbericht 2014 /pʊʃ/ /ʌ/:

Loan word Year of introduction Phonetic transcription of the borrowed English word (part) Brunch 1957 /bɹʌn(t)ʃ/ Multiplechoicevraag** 1961 /ˈmʌl.tɪ.pl̩ tʃɔɪs/ Muffin 1989 /ˈmʌ.fɪn/ Fundraiser 2012 /ˈfʌnd.ɹeɪ.zə/ Bucketlist 2015 /ˈbʌ.kɪt lɪst/ Pubquiz 2017 /pʌb kwɪz/

7 This column shows the phonetic transcription of the parts in every loan word that were borrowed directly from English. Most of the words that were tested were entirely loaned from English, in which case the phonetic transcription of the entire word is shown, but some of them have added Dutch suffixes or are compounds of English and Dutch roots. These parts include the Dutch infinitive suffix -en in the words whatsappen, flossen,

copy-pasten, youtuben and levelen and the words toets (‘key’) in capslocktoets, bericht (‘notification/message’)

in pushbericht and vraag (‘question’) in multiplechoicevraag. These non-English parts are not shown in the transcription and their vowels will not be measured.

(19)

19

/ɒ/:

Loan word Year of introduction Phonetic transcription of the borrowed English word (part) Hotdog 1951 /ˈhɒt.dɒɡ/ Flossen 1984 /flɒs/ Content 1994 /ˈkɒn.tɛnt/ Copy-pasten 2010 /ˈkɒ.pi.peɪst/ Smartwatch 2014 /ˈsmɑɹt.wɒtʃ/ Singalong 2016 /ˈsɪŋ.ə.lɒŋ/

*These words also contain another of the foreign vowels in this study, which also be measured (/ɒ/ in the case of

capslocktoets and /æ/ in the case of goodiebag)

** This word was not present in the Chronologisch woordenboek, but the Ngram viewer of the Digital Library for Dutch Literature (Digitale Bibliotheek der Nederlandse Letteren, 2015) shows that the word multiplechoicevraag was used in 1961. This means that this was the latest date this word could have been introduced.

Additionally, the following words in table 4 were used as filler items. In total, 8 words were used as filler items, also being divided in an older group and a younger group, with each of these two groups containing four words. The oldest of these filler words was fifty-fifty. We do not have an exact year of introduction for this word, but the first time that it was recorded in the Ngram viewer of the Digital Library for Dutch

Literature (Digitale Bibliotheek der Nederlandse Letteren, 2015) was in 1955, therefore we can be sure that it had to have been introduced before this date. The youngest of the filler words was levelen (from “to level”), which was included in the Van Dale dictionary for the first time as recent as 2017. The phonetic transcription of these words is also shown, for the purpose of showing that none of the four

phonemes measured for our study are present in these words. A full version of the all words in the random order as they were given to the participants, together with the given explanation of the task, can be found in appendix 1.

Table 4: The words that were used as filler items, their introduction years and phonetic transcription

Loan word Year of introduction Phonetic transcription of the borrowed English word (part) Fifty-fifty 1955 /ˈfɪf.ti.fɪf.ti/ Marshmellow 1968 /ˈmɑɹʃ.mɛ.loʊ/8 Creditcard 1974 /ˈkrɛ.dɪt.kɑrt/ Blooper 1991 /ˈbluː.pə/ Youtuben 2007 /ˈjuː.tjuːb/ (RP) or /ˈju.tub/ (GA) Webcare 2014 /wɛb.ˈkɛə/ All-inclusive 2016 /ɔːl.ɪn.ˈkluː.sɪv/ Levelen 2017 /ˈlɛ.vəl/

8 Pronounced as /ˈmɑɹʃ.mæ.loʊ/ in some variants of English, which would contain an /æ/. The difference between this word and the words in the /æ/-group is the fact that this word does not always contain an /æ/, while the others do.

(20)

20 In addition to the recording task, a questionnaire was sent to the participants of the study. The questions asked about their use of English in daily life and their stance of (English) loan words in Dutch; the list included questions such as: “How often do you use English for your study and/or work (on a scale from often to never)?”, “What is your stance on the introduction of English loan words in Dutch (on a scale from very positive to very negative)?” and “Do you think Dutch alternatives should be offered for every loan word that is borrowed into Dutch?”. Furthermore, some personal

information of the participants, such as age, gender and highest finished level of education was asked. In order to avoid any subconscious influence of the

participants’ enunciation of the loan words as a result of filling in the questionnaire before performing the experiment, the questionnaire was sent to the participants after they had handed in their recordings of the loan words. The full list of questions can be found in appendix 2.

Procedures

All participants were asked to record themselves while pronouncing the 32 loan words.9 For this, some instructions were given in the document: they were

encouraged to read the words once before starting the recording, to become familiar with the words and reducing potential mishaps and/or stuttering as much as possible. Furthermore, they were advised to talk in their normal voice and in a calm tempo, in order to ease the analysis of their pronunciation. Lastly, it was recommended to them to record in a quiet place and, if possible, to avoid any background noises that could be present and audible in their recordings.

All participants were given the same document, and therefore the same order of the loan words. When recording, they were advised to record smaller chunks of the list, instead of one big recording of 32 words. This was encouraged because it would be easier to fix mistakes, as less words would have to be re-recorded, and the

recordings would be more convenient and clearer to analyse for the researchers. To make this easier and more consistent for the participants, the words in the

documents were divided into 4 columns of 8 words, and participants were specifically asked to make 4 recordings, one of each of the columns.

The questionnaire was sent to the participants right after they had handed in their recordings. After they had sent their answers, they were thanked for their

participation and were asked to not share the questions of the questionnaire with any of the other participants of the study. This had to be done, as more than half of the participants of the younger group are part of a close-knit friend group, and questions could easily be distributed among them. This had to be prevented, as circulation of the questions to participants who had not yet recorded themselves could influence their pronunciation of the English loan words.

9 Originally, it was planned to record all participants in person, so that the same microphone could be used for all recordings. However, due to the outbreak of the SARS-CoV-2 virus and the social distancing measures that were taken in the Netherlands as a result of the pandemic, it was neither possible nor allowed to do so. This means that we were dependent on the presence (and if so: the quality) of the microphones of the participants, which might have had a slight effect on the results.

(21)

21 Design and Analyses

Due to the fact that the participants had to record themselves, there was not a uniform application that was used for the recordings. However, most of the participants have notified the researchers that they had used the standard voice recording application that is available on Windows computers. To answer the

question if this application could be used for an accurate analysis of vowels, we refer to De Decker & Nycz (2011), who, in their study of determining which recording devices can be used in sociophonetic studies, state that the voice recorders on computers and mobile devices can be used in the basic study of vowels. This is because they do not show any significant differences in the measurements of F1 and F2-values when compared to the measurements of these values for the same vowels that were recorded on professional voice recorders. As this study will focus on the measurements of the F1 and F2-values only, recordings that were made on a computer or mobile phone do not raise a problem. Any background noise in the recordings was, if possible, maximally reduced in Audacity (Team Audacity, 2019) and they were further analysed in Praat (Boersma & Weenink, 2020). This analysis included the calculation of the vowel formants, which were calculated as an average value over the entire duration of the vowel. After this calculation, the formant values were illustrated using Visible Vowels (Heeringa & Van de Velde, 2018), where a depiction of the measured values of the formants of each of the vowels can be created, and possible differences between age groups and years of introduction of the loan words can be shown. Further statistical analysis was conducted in Rstudio (Rstudio team, 2020). The alpha-level for this analysis will be α = 0.05

The plots of the F1 and F2-values that will follow as a result from the described analysis will be used to test the following statistical hypotheses:

- The formant values of the vowels that have a similar vowel that is present in Dutch will show a pattern that is less clear than the vowels that do not have a similar vowel in Dutch.10

- The pronunciation of the loan words by the younger speakers will be more akin to English than the pronunciation of the older participants

- The pronunciation loan words that were borrowed before the year 2000 will be less similar to English than the pronunciation of those that were borrowed after 2000.

Results

In total, 580 vowels were measured: 20 speakers times 29 vowels.11 All vowels were measured in Praat and their formants were visualised using Visible Vowels. The results of this analysis of all the vowels are shown in figure 7:

10 Taking the IPA vowel chart into account, we consider /æ/ and /ʊ/ to be vowels that have a similar

counterpart in Dutch, since both vowels only have to slightly change 1 or 2 of their characteristics. For /æ/, its near-openness changes to mid-openness, while /ʊ/ would change its near-backness and near-closeness to completely back and completely close. /ʌ/ and /ɒ/ do not have similar counterparts, since /ʌ/ would have to completely change its roundedness or its openness from mid-open to fully open and /ɒ/ would also have to change its roundedness or change from a fully open vowel to a mid-open vowel.

11 The division between the four different vowels is /ʌ/ - 6, /ʊ/ - 6, /ɒ/ - 8 and /æ/ - 9. While it is true that all four vowels had the same amount of words (3 new and 3 old), there were some words which contained more than 1 of the vowels. This includes the earlier mentioned goodiebag (which in addition to /ʊ/ also contains an /æ/) and capslocktoets (which contains both an /æ/ and an /ɒ/) and three words which contain the same vowel twice: flashback, hashtag and hotdog.

(22)

22 Figure 7: The formants of all measured vowels, separated for male speakers (left) and female speakers (right)

3 groups of the vowel measurements are mostly around their two respective expected positions, either around their formant values in English or the formant values of the closest Dutch vowel, for /æ/ this ranges from a mid-open to near-open front vowel, for /ʊ/ from a near-close to close back vowel, and for /ɒ/ from a mid-open to open back vowel. However, we can see that one of the vowels, /ʌ/, is not in its expected position of a mid-open back vowel, but has shifted towards a middle vowel, with some notable exceptions of measurements with a lower F1 value or either a higher or a lower F2 values than the other measurements in their groups.

To a lesser degree, a noticeable result can also be distinguished for /æ/, which

normally is a front vowel, but as we can see, four of the measurements for this vowel, all of which are from the more recent loanwords, have a remarkable low F2 values below 1500 Hz, which would lead to the conclusion that they could be considered back vowels. A better visualisation of this disparity can be seen in figure 8, where all measurements are plotted separately per vowel in order to analyse differences within the groups. An explanation for this discrepancy will be given in the discussion

(23)

23 Figure 8: The formant values of each vowel separated, with /æ/ on the top left, /ʊ/ on the top right, /ʌ/ on the bottom left, and /ɒ/ on the right. The green dot represents the average formant values of each of the vowels in English.12

As we can see in the separated visualisations, two divisions were made for all vowels: a division between the vowels pronounced by the younger participants and those pronounced by the older participants, and a difference between the vowels that were present in the older loan words versus those that were present in the more recent loan words. The mean values of all four combinations of these variables per vowel can be found in table 5 and are visualised in figure 9:

12 These values were taken from table 2, as an average of both male and female values of Received Pronunciaition and General American (where possible).

(24)

24 Table 5: The average vowel formants per vowel per combination of speaker age group and word age group

Average F1 values:

Vowel Younger speakers Older speakers

Pre-2000 words Post-2000 words Pre-2000 words Post-2000 words

Mean SD Mean SD Mean SD Mean SD

/æ/ 611 89 582 67 581 96 595 70

/ʊ/ 400 51 377 56 386 70 372 55

/ʌ/ 491 83 504 74 478 77 507 84

/ɒ/ 556 104 570 94 553 85 574 72

Average F2 values:

Vowel Younger speakers Older speakers

Pre-2000 words Post-2000 words Pre-2000 words Post-2000 words

Mean SD Mean SD Mean SD Mean SD

/æ/ 1756 154 1715 200 1834 134 1789 174

/ʊ/ 1102 223 1173 216 1053 209 998 157

/ʌ/ 1406 235 1453 164 1395 146 1399 148

(25)

25 Figure 9: The mean formant values for all combinations of the variables speaker age and word age per vowel, with /æ/ on the top left, /ʊ/ on the top right, /ʌ/ on the bottom left, and /ɒ/ on the right

When we examine the results of both all separate measurements of the vowels and their mean values, we can see that the differences for each group are quite small, with the discrepancy between the mean values of the groups consisting of at most 30 Hz for the first formant and 175 Hz for the second formant, which is only the case for /ʊ/. The differences in second formant values for the three other vowels does not exceed 100 Hz.

(26)

26 We will now proceed to briefly discuss the noticeable differences between the groups per vowel, and the results of the statistical tests that were conducted, namely a Two-way ANOVA for both the F1 and F2 values per vowel. Among the vowels that were tested, the clearest difference in terms of formants values can be seen by /ʊ/. A clear distinction can be made here between the older speakers with lower values for both the first and second formants, and the younger speakers which show higher F2 values for both word ages and also higher F1 values for older loan words, but not for more recent loan words. In the statistical analysis, most of these findings were also identified as a significant medium-sized effect of speaker age on the F2 (F(1, 116) = 9.15, p = 0.003. ω2 = 0.064) and also a small interaction between the two variables

regarding the F2 values (F(1, 116) = 2.95, p = 0.089. ω2 = 0.016), meaning that the

F2 was higher for younger speakers when pronouncing more recent loan words, while the opposite is true for older speakers. However, this interaction was not significant, as the p-value is just above our alpha level of 0.05. Furthermore, another non-significant result was found in the form of a small word age effect on the first formant value (F(1, 116) = 3.06, p = 0.083. ω2 = 0.017).

A clear distinction between the influence of our variables can be seen in the measured formants of /ɒ/, where we can clearly conclude that the time since the introduction of the loan word has a larger influence on the pronunciation than the age of the speaker, as both speaker ages lay close together for each word age, while the two word ages are further separated in both F1 and F2 values. However, a two-way ANOVA showed that neither of these effects was significant, with the effect on the F1 not even being of small size (F(1, 156) = 1.20, p = 0.27. ω2 = 0.004) and the effect on

the F2 being of small to medium size, but not significant: (F(1, 156) = 2.47, p = 0.12. ω2 = 0.051) As expected from the visualisations, no effect was found for speaker age

(for both formants ω2 < 0.01 and p > 0.05).

When examining the measurements for /æ/, we see that the older participants have a higher average F2 for both word ages, which was also found in the two-way ANOVA as a significant small result (F(1, 176) = 8.92, p = 0.003. ω2 = 0.042). The older

participants also have a lower average F1, albeit only slightly (around 5 Hz). This is because of a split in the pronunciation of younger participants: they pronounce older words with a higher F1, but their F1 for younger words is similar to that of older participants. Therefore, a small interaction between the two variables (F(1, 176) = 3.00, p = 0.085. ω2 = 0.011) was found during the analysis, meaning that the F1 for

older people is higher when pronouncing older loan words, and higher for younger people when pronouncing recent loan words. However, this interaction is just non-significant, with the p-value of this interaction also being slightly higher than our alpha-level of 0.05. Furthermore, a non-significant small effect of word age on the F2 was discovered: F(1, 176) = 2.86, p = 0.092. ω2 = 0.010.

(27)

27 Lastly, the average values of /ʌ/ show us a difference for both variables, although one is more evident than the other. The word age shows an expected difference, with more recent words showing an average F1 of more than 500 Hz while older words for both speaker ages have a lower F1. Moreover, older participants showed a lower F2 rating than younger participants for both word ages, but this difference is only

obvious for newer loan words. The F2 for older loan words pronounced by younger participants is quite similar to that of older participants, with both having values around 1400. However, when tested with a two-way ANOVA, neither the word age effect on the F1 (F(1, 116) = 1.94, p = 0.16. ω2 = 0.008) or the effect of speaker age

on the F2 (F(1, 116) = 1.00, p = 0.32. ω2 = 0.000) showed a significant result of any

size. In fact, /ʌ/ was the only vowel for which no statistical results of any size were found for any variables.

Along with the words that had to be recorded, we also asked the 20 participants about their use of English in their study and/or work environments, their assessed use of English loan words when speaking their mother tongue and their stances on the introduction of English words to the Dutch language. When examining the answers given on these questions, we can observe a clear difference between the younger and the older participants.

The younger participants all answered that they either regularly or often use English for their studies or their job. A more diverse spread is seen within the older participant group, where around half of the participants indicated that they sometimes use

English, some said they seldom use the language, and one participant even acknowledged to never using the language. The same difference applies to the assessed use of English loan words in Dutch as well.

Almost all the younger participants all indicated that they regularly or often use English loan words while they are speaking Dutch, while none of the older participants thought they often used loan words. The most given answers were ‘regularly’ and ‘sometimes’.

However, whereas the use of English and English loan words differs between the two groups, both groups generally agree on the borrowing of English loan words in

Dutch. Both groups generally do not have a particular opinion on the introduction of new loan words. Most of the participants have indicated that they do not have a positive nor a negative stance on the introduction of new loan words, with one young participant indicating they think this is a positive change, and two older participants thinking this is a negative change.

Not coincidentally, the two participants who expressed that they have a negative point of view towards the growing number of English loan words in Dutch, are also the only participants who indicated that they believe that alternative Dutch forms for English loan words should be offered alongside the newly borrowed loan words. All other 18 participants answered that they do not think Dutch alternatives should be introduced. The results of the questionnaire are shown in full in figures 10, 11, 12 and 13.

(28)

28 How often do you use English for your study and/or work?

What is your stance on the introduction of English loan words into Dutch?

(29)

29 Should Dutch alternatives be offered for all new loan words that are introduced into the Dutch language?

Figures 10, 11, 12 and 13: The results of the four questions on English use and introduction of loan words, with the younger participants on the left and the older participants on the right

Discussion

From the visualisations of the measurements it has become clear that each vowel differs from the others with regard to the distinction between speaker ages and times of introduction of the loan words. However, not one of the vowels behaves in the exact way that we predicted in our hypotheses: the pronunciation of older speakers enunciating older loan words being more similar to the vowel existent in Dutch while younger speakers pronounce more recent loan words more akin to the original English word, with the formant values of the two mixed categories of older speaker-recent loan word and younger speaker-old loan word lying somewhere in between the two extremes. In order to find out to what extent the vowels have followed these predictions, we will first have to decide which Dutch vowel is the most equal to the English vowel, after which we compare the formants of both vowels (and the differences between them) to the results from our analysis.

Referenties

GERELATEERDE DOCUMENTEN

However, this so-called interlanguage speech intelligibility benefit (hence ISIB) has not yet been shown to exist between learners of a foreign language who

Mail ze dan naar Aduis (info@aduis.nl) en wij plaatsen deze dan als downlaod op onze web site. Memory

(1) Far from rejecting the lowering of final -u to -o before an ini- tial nasal consonant, he observes that the reflex of the rounded nasal vowel is -u if the vowel of the

He is the author of various publications, including Religion, Science and Naturalism (Cambridge: Cambridge University Press, 1996), and Creation: From Nothing until Now (London and

He started with the observation that &#34;a connection between the Turkic and Tocharian words does indeed seem likely: both Old Turkish kn and Tocharian B kaum, A kom occur

We use a touchscreen version for our Infant and Child waves (i.e., R3 and R6: touch screen is Dell S2240T 23.5&#34;; operating system: macOS High Sierra), in which children touch

We hypothesized that there is regional variation in the pronunciation of /s/ within the Dutch language area. a more [ʃ]-like pronunciation of /s/) than speakers from other

Also, at the lower edge of the vowel space there is little differentiation between more centralized (half) open lax vowels and peripheral open tense vowels as the Dutch ESL speakers