Perception of illusory vowels by Persian speakers in several consonant clusters of French loanwords
Introduction Previous research has highlighted the challenge listeners face in distinguishing between legitimate and non-native consonant sequences, presenting potential perceptual illusions (Berent, et al., 2007; Dupoux, et al., 1999). Some researchers argue that the difficulty in discerning these...
Saved in:
| Main Authors: | , , |
|---|---|
| Format: | Article |
| Language: | Persian |
| Published: |
Alzahra University
2025-02-01
|
| Series: | زبان پژوهی |
| Subjects: | |
| Online Access: | https://zabanpazhuhi.alzahra.ac.ir/article_8380_52ef4269cb9690dac4447f0f35590186.pdf |
| Tags: |
Add Tag
No Tags, Be the first to tag this record!
|
| Summary: | Introduction
Previous research has highlighted the challenge listeners face in distinguishing between legitimate and non-native consonant sequences, presenting potential perceptual illusions (Berent, et al., 2007; Dupoux, et al., 1999). Some researchers argue that the difficulty in discerning these pairs stems from a perceptual illusion induced by the phonological system, hindering the accurate perception of phono-tactically illegal consonant clusters in the L1. The relationship between speech perception and loanword adaptations has been extensively explored through experimental research and theoretical debates (Guevara-Rukoz, et al., 2021). Across various languages, listeners tend to perceive illusory vowels inserted between consonant clusters prohibited in their native language (Dupoux, et al., 1999). Notably, listeners may perceive non-existent illusory vowels when exposed to non-native consonant clusters (Dupoux, et al., 1999). In this paper, we aim to show that this illusory vowel perception occurs in Persian as well.
Materials and methods
Sixty-one monolingual or quasi-monolingual Persian speakers (age: 18-40), who had zero or low proficiency in a foreign language with legal consonant clusters, were recruited to participate online in the first and second perception experiment (31 participants for the identification task, and 30 participants for the discrimination task). None of the participants reported any hearing or articulation disorders.
The stimuli for the identification task had the structure C1C2V1C3, C1C2V1C3, and fillers were presented in two forms V1C1C2V2C3 and C1VC2V1C3. None of the stimuli were words in either Farsi, French, or English. English was included because most Iranians learn English at school, and even with the lowest language proficiency, they have some knowledge of English (this also applies to the discrimination task). The stimuli were natural recordings produced by a native speaker of French. This speaker was selected to avoid any overlap with the participants’ native or familiar dialects and to ensure accurate pronunciation of the non-words.
Participants were tested individually in a soundproof room. In each trial, they heard a stimulus and were asked to identify the vowel between the two consonants, if any. A Persian transcription of the item was provided on screen. Participants were informed that the stimuli were non-words with no meaning, and they were instructed to choose the option that best matched how they would write down the word they heard, including the short vowels. There were seven options for each question. These options were written based on no vowel [Ø] (the natural consonant cluster), or there are Farsi vowels in the form of C1VC2 from the list [e, a, i, o, u, ɑ]. The stimuli were concatenated in random order for each participant. In the discrimination task, the filler and the target non-word were in the ABX form; X was like one of them (A or B as the target word and the filler). Participants were tested individually in a quiet room. At each trial, they heard three stimuli and were asked to choose the first option if the third non-word was like the first one and the second option if it was like the second one. The stimuli were concatenated in random order, and in each question, the order of the non-words was random (ABB or ABA) with an ISI of 500 Ms.
Results and discussion
Research Question: Do Persian monolingual native speakers perceive an illusory vowel in consonant clusters, and if so, which vowel do they perceive?
Table 1.
Examination of differences in the frequency of perceived vowels in consonant clusters in the sample group and the hypothetical population for the identification task.
Frequency
Mean
Standard
eviation
Standard
Error Mean
t-value
Degrees of
Freedom
Two-tailed Significance
Level
Sample Group
31
70.68
30.49
5.48
4.037
30
0.008
The t-value (t = 4.037) and the significance level (p = 0.008) indicate that the observed difference between the mean frequency of perceived vowels in the sample group and the hypothetical population mean is statistically significant. Since the mean in the sample group (70.68) is greater than the hypothetical population mean (65), it can be concluded that both male and female participants perceived vowels in consonant clusters, and the frequency of heard vowels in the sample group is significantly higher than in the hypothetical population (see Table 1).
Table 2.
One-sample t-test for perceived vowels in consonant clusters in discrimination task
Frequency
Mean
Standard Deviation
Standard
Error Mean
t-value
Degrees of Freedom
Two-tailed Significance
Level
Sample Group
30
38.17
4.93
90/0
2.405
29
0.023
The t-value (t = 2.405) and the significance level (p = 0.023) indicate that the observed difference between the mean frequency of perceived vowels in the sample group and the hypothetical population mean is significant. It can be concluded that both male and female participants heard vowels in consonant clusters, and the frequency of heard vowels in the sample group was significantly higher than in the hypothetical population (see Table 2).
Research Question: Where do Persian monolingual native speakers perceive the illusory vowel?
Table 3.
One-way ANOVA test for assessing the differences in vowel perception frequency across different consonant clusters.
Significance
Level (p-value)
Variance Test
Statistic (F)
Mean
Squares
Degrees of
Freedom (df)
Sum of
Squares
Category
0.000
83.127
13514.057
2
27028.115
Between Groups
162.572
88
14306.325
Within Groups
90
41334.440
Total
Table 3 shows a significance level of 0.000. Since this is below the conventional threshold of 0.05, it can be inferred that there is a statistically significant difference in the frequency of heard vowels among initial, medial, and final consonant clusters. Considering that the mean frequency of perceived vowels in the initial cluster is higher, it can be concluded that the highest frequency of perceived vowels is in the initial consonant cluster, followed by the medial cluster. However, the mean frequency of perceptual vowels in the final cluster is close to zero, indicating a noticeable difference compared to the initial and medial clusters. This suggests that in the final consonant cluster, either no vowels are heard, or they are heard to a minimal extent, possibly due to chance or other errors.
Research Question: Which cluster has a higher illusory perception?
Table 4.
Summary of Friedman rank test calculations
Sample
Size
Chi-Square (χ²)
Degrees of
Freedom (DF)
Significance
Level (Sig)
30
115.234
24
0.000
Table 4 presents the results of the Friedman test, including sample size, Chi-Square value, degrees of freedom, and significance level. The test yielded a Chi-Square value of 115.234, with 24 degrees of freedom, and a significance level of 0.000. Since the significance level is below 0.05, it can be concluded that the differences in mean ranks among perception frequencies in different clusters are statistically significant.
Conclusion
In response to the research questions, we first examined whether Persian-speaking monolinguals perceive an illusory vowel in consonant clusters. We then investigated where the illusory vowel is mostly perceived (onset, middle, or coda). The results showed that Persian-speaking monolinguals perceive an illusory vowel in onset and medial consonant clusters. We also explored the types of consonant clusters in which participants perceived vowels and ranked them from highest to lowest frequency. The highest frequency was observed for the [gl-] initial cluster (rank 1), and the lowest in the final cluster [-str] (rank 25). |
|---|---|
| ISSN: | 2008-8833 2538-1989 |