The ins and outs of paragoge and apocope in Japanese-English interphonology - Steven Ross,
We chose items with a minimum frequency of 14, that is, those appearing at least once a year. The test items had a frequency of 0—4; the maximum frequency for the database is 69, In choosing the words, effort was made to avoid items whose first half constitutes a loanword. This is because, if the first half sounded like a loan, Japanese speakers may perceive that part as an independent word.
- Japanese/Transcribing English to Japanese - Wikibooks, open books for an open world;
- essays on wordsworths poems;
- Result Filters.
For example, if they hear topnotch , they may think that they heard top , which is a high-frequency loanword in Japanese. However, due to difficulty finding appropriate items, the first halves of eight Loan-A items buckshot, gasket, tuckshop, sackrace, suckling, bumkin, lugsail , and hemline potentially sounded like loanwords. Therefore, we conducted a supplementary analysis examining the vowel epenthesis rates for these items versus the other 12 Loan-A items, 5 which indicated that the effect of such items was probably minimal, if not zero.
Eighty-four filler items were prepared. Their familiarity to Japanese speakers varied. The other 42 did not e. The test words and fillers were recorded by a male native speaker of British English who was trained in phonetics, using a Zoom H4n digital recorder. He was asked to pronounce each word naturally while ensuring that word-medial stops were clearly released. After the recording session, the sound files were downsampled at 22, Hz with a depth of 16 bits. As discussed in the Overall design section, a modified version of the Mora Detection Task Nomura et al.
The participants were tested individually at a computer, wearing headphones. They pressed a specified key to indicate whether a target mora on the screen e. In the case of native speaker participants, target moras were presented using International Phonetic Alphabet IPA symbols e. The experiment was conducted using SuperLab 4. After a fixation cross and a beep, a test or filler item was presented.
Immediately after the end of the sound file, a target mora appeared on the screen using a mora-based Japanese katakana character or IPA symbols for native speakers. The participants pressed a specified key as quickly as possible to indicate whether the target mora was present or absent in the word that they heard. The responses and RTs were stored in the computer. Typically, the experiment took 7—10 minutes.
There was a training phase prior to the experiment that consisted of an instruction followed by three short practice sessions. In the previous studies that used nonwords, relatively short instructions and practice sessions seem to have sufficed. For example, in Dupoux et al.
"epenthesis" in Czech
In Dupoux et al. On the other hand, our use of real English words made the instructions and practice sessions especially important. The results vowel epenthesis rates might have shown a ceiling effect if the participants had responded based solely on L1 representations, whether activated from loanword representations in the lexicon or generated using L1 phonotactics. In the instruction session prior to the practice sessions, non-native speakers were given a description of the differences in syllable structure between English and Japanese, using words familiar to them.
The description started with the word Boston written on a piece of paper. They were then told that we would like them to answer no in this case, because English allows sequences of consonants while a consonant is usually followed by a vowel in Japanese. Romanized loanwords and their original English words were utilized for comparison by participants. We then provided a few other examples. After the description of the English and Japanese syllable structures or the IPA symbols, the participants went on to the practice sessions.
The first of the three practice sessions was a written one. When they were finished with all the words, they were given feedback to ensure that they understood what was required in the task. Others were due to their not being sure about the actual pronunciation.
In either case, they understood the feedback provided by the experimenter. The second and third practice sessions were given using a computer and the same 14 words. The warning did not appear even if they responded slowly so they were able to use as much time as they needed before providing their responses. After the third practice session, participants were asked whether they were ready to go on to the main experiment or wanted to practice again.
They typically answered that they were ready for the experiment. Figure 1.
- Definition and Examples of Epenthesis.
- The ins and outs of paragoge and apocope in Japanese-English interphonology.
- Most Downloaded Articles?
The vowel epenthesis rate constitutes a ratio scale that comes from five two-choice trials per condition for each participant. Then, a three-way analysis of variance ANOVA was conducted using Proficiency introductory, intermediate, and native as the between-subjects variable, and Frequency and Voice as the within-subject variables.
Post-hoc Bonferroni pairwise comparisons revealed the following significant differences. The native speaker group was significantly different from either of the learner groups for all conditions. Figure 2. Vowel epenthesis rates by condition for the three groups. The intermediate group did not show significant differences between any two voicing patterns for Loan-P items. Again, no significant differences were observed for the native speaker group. Since Matthews and Brown pointed out the possibility that vowel epenthesis processes are reflected in some way in RTs, we conducted a one-way ANOVA with the overall RTs for the three groups introductory learners: ms; intermediate learners: ms; native speakers: ms.
Figure 3 shows the scatter plots for the three groups. The crosses correspond to individual participants. Figure 3.
Mean reaction times RTs and vowel epenthesis rates for the three groups. Each cross corresponds to a participant. By examining Proficiency, Loanword Representation, and Voice, the present study aimed to determine the following: a the extent to which L1 phonology and phonotactics i. With regard to the effects of the L1, we manipulated the presence or absence of loanword representations in the L1 lexicon and voicing patterns for consonant clusters.
Both factors yielded results. We interpret this result as a demonstration of the enhancement of epenthesis by loanword representations.
This is discussed further below. The restoration of vowels devoiced through Japanese high vowel devoicing processes seems to be another robust factor that has enhancing effects on vowel epenthesis; the Loan-A items mainly showed different epenthesis rates according to voicing patterns. The lack of effects for Loan-P items is not surprising, because L1 loanword representations may make it difficult to observe the effect of voicing alone. The reason for the presence of somewhat similar patterns between the Loan-P and Loan-A items in the introductory learner group, but not in the intermediate learner group, is unclear.
We conjecture that links between some English words and their loanword counterparts are weaker in introductory learners than in intermediate learners. However, this requires further investigation, possibly by asking each learner to rate the loanword familiarity for the test items after the experiment. However, our preliminary study Nomura et al.
One possible explanation is related to the observation that a devoiced vowel often results in an aspiration-like period or lengthening of a fricative e. We speculate that a process different from the restoration of a devoiced vowel may be at work. For example, the continuous voicing quality across the two consonants could lead to the perception of an illusory vowel. However, we do not, at this point, have a clear explanation for this result.
"epenthesis" in Czech
To summarize the findings so far, despite some unexpected results, they generally show the enhancing effects of loanword representations and vowel restoration on perceptual vowel epenthesis. The loanword effect provides evidence for the notion that it is difficult for learners to maintain native-like representations i. The other effect, vowel restoration, confirms the higher epenthesis rates for voiceless-voiceless clusters observed in Dupoux et al. Apparently, Dupoux et al. However, by replicating their results using a larger number of items, our study implies that Japanese speakers exhibit a different process: they tend to restore vowels when they hear voiceless clusters, regardless of whether the language they hear is Japanese or English.
In our results, proficiency did have significant effects on vowel epenthesis rates. Therefore, higher proficiency does seem to lead to lower epenthesis rates in these conditions. Two questions have arisen regarding these results.
Related epenthesis japanese english
Copyright 2019 - All Right Reserved