languages

Article
Visible Vowels as a Tool for the Study of Language Transfer
Wilbert Heeringa 1, * and Hans Van de Velde 1,2, *

1 Fryske Akademy, 8911 DX Leeuwarden, The Netherlands

2 Utrecht University, 3584 CS Utrecht, The Netherlands
* Correspondence: [email protected] (W.H.); [email protected] (H.V.d.V.)

Abstract: In this paper, we demonstrate the use of Visible Vowels to detect formant and durational
differences between L2 and L1 speakers. We used a dataset that contains vowel measures from L1
speakers of French and from L2 learners of French, with Italian, Spanish and English as L1. We found
that vowels that are not part of the L1 phonological system are often pronounced differently by L2
speakers. Inspired by the Native Language Magnet Theory which was introduced by Patricia Kuhl
in 2000, we introduced magnet plots that relate vowels shared by the French phonological system
and the learners’ phonological system—the magnet vowels—to the vowels found only in the French
phonological system. At a glance, it can be seen which vowels are attracted to the magnets and which
vowels become further away from the magnets. When comparing vowel spaces, we found that the
shape of the French vowel space of the English learners differed most from the shape of L1 speakers’
vowel space. Finally, it was found that the vowel durations of the L2 speakers are greater than that of
the L1 speakers of French, especially those of the English learners of French.

Keywords: French; second language acquisition; vowels; formants; duration; Native Language
Magnet Theory; vowel normalization; Visible Vowels; language transfer

1. Introduction
1.1. Acquisition of Sounds
The influence of a person’s first language on the learning of a foreign language is a
Citation: Heeringa, Wilbert, and Hans classic topic in applied linguistics and second language learning. In particular, the degree
Van de Velde. 2024. Visible Vowels as of correspondence between the phonological systems of the languages was found to be an
a Tool for the Study of Language important factor that determines the extent to which someone is successful in acquiring an
Transfer. Languages 9: 35. https:// L2 language. Flege (1995, p. 238) writes:
doi.org/10.3390/languages9020035
“During L1 acquisition, speech perception becomes attuned to the contrastive
Academic Editors: Paolo Mairano and phonic elements of the L1. Learners of an L2 may fail to discern the phonetic
Sandra Schwab differences between pairs of sounds in the L2, or between L2 and L1 sounds,
Received: 18 October 2023
either because phonetically distinct sounds in the L2 are “assimilated” to a single
Revised: 8 December 2023
category (see Best this volume), because the L1 phonology filters out features
Accepted: 8 December 2023 (or properties) of U sounds that are important phonetically but not phonologically,
Published: 23 January 2024 or both.”
According to the first hypothesis of the Speech Learning Model (SLM) that was
developed by Flege (1995) and his colleagues, “learners perceptually relate positional
allophones in the L2 to the closest positionally defined allophone (or “sound”) in the L1.”
Copyright: © 2024 by the authors.
(p. 238).
Licensee MDPI, Basel, Switzerland.
Best and Tyler (2007, p. 20) write: “Perceptual learning occurs for some L2 contrasts,
This article is an open access article
but seems to depend on their phonological and phonetic relationship to the L1, specifically
distributed under the terms and
on perceived similarities vs. dissimilarities to L1 phonemes.”
conditions of the Creative Commons
Kuhl (2000) introduced the Native Language Magnet Theory. This theory suggests that
Attribution (CC BY) license (https://
creativecommons.org/licenses/by/
L1 learners (babies) categorize the sounds they hear in their mind into phonetic categories.
4.0/).
Once a category is established, it will function as a magnet for sounds that are similar to the

Languages 2024, 9, 35. https://doi.org/10.3390/languages9020035 https://www.mdpi.com/journal/languages

Languages 2024, 9, 35 2 of 16

sound that is represented by that category. When learning an L2 language, yet-unknown

sound patterns will be attracted to the L1 categories as well. Kuhl (2000) writes:
“A model reflecting this developmental sequence from universal perception to
language-specific perception, called the Native Language Magnet model, pro-
poses that infants’ mapping of ambient language warps the acoustic dimensions
underlying speech, producing a complex network, or filter, through which lan-
guage is perceived (39, 40, 82). The language-specific filter alters the dimensions
of speech we attend to, stretching and shrinking acoustic space to highlight the
differences between language categories. Once formed, language-specific filters
make learning a second language much more difficult because the mapping ap-
propriate for one’s primary language is completely different from that required
by other languages.” (p. 11854)
Visualizing the effect an L1 language has on an L2 language can help guide an L2 learner
more effectively in acquiring or improving their pronunciation of the speech sounds. We are
not referring here to evaluating an L2 learner’s pronunciation, but rather to identifying the
exact differences between the L2 speaker’s pronunciation and the target pronunciation.
In this paper, we focus on the pronunciation of vowels. Differences in vowel pronunci-
ation are evaluated in formants and duration.

1.2. Existing Software for Vowel Visualization

For the visualization of formants, several programs are available. Without claiming to
be exhaustive, we mention the R packages vowels and phonR, and the programs NORM
and VOIS3D, VowelWorm, VowelCat and Vowel Viewer.
With the R package vowels, phonetic and sociophonetic vowel formant data can be
manipulated, normalized and plotted (Kendall and Thomas 2018). This package is also the
backend for the web app NORM. Using NORM, vowels can be plotted and the formant
measurements can be normalized using several normalization methods. Since NORM is
less flexible than the vowels package, the authors encourage users to use their R package
vowels rather than using NORM.
With VOIS3D, both formant frequencies and duration can be normalized. Spectral
overlap can be assessed by an analytic geometric solution. VOIS3D runs only on Windows
operating systems (Wassink 2006).
Another R package that can be used for the visualization of vowels is phonR (McCloy
2016). Trajectories can be visualized with an unlimited number of measure points. Addi-
tionally, IPA glyphs, confidence ellipses and convex hulls that mark the outline of a vowel
space can be drawn. The degree of encroachment or overlap between vowel categories can
be calculated and plotted by means of a heat map.
There are a few programs that record a user’s voice and display vowel plots in
real-time such as KlinkerMikken (developed by linguists of Leiden University in 2010),
VowelWorm1 (Frostel et al. 2011), VowelCat2 (at Ohio University in 2014) and Vowel
Viewer3 (Rehman 2021).
The use of R packages requires some knowledge of the programming language R, and
as such programs like NORM4 and VOIS3D5 are more user-friendly, but are limited in
their functionality and flexibility. The four programs that give real-time response are useful
for training purposes, but are even more limited in their visualizations and not useful for
visualizing existing vowel measurements or relating them to potential explanatory factors.

1.3. Visible Vowels

In this paper, we present Visible Vowels, a web application for the visualization of
vowel variation that aims to combine user friendliness with maximum flexibility and func-
tionality. Characteristic of this web app is the use of a live view: each time the user changes
something in the settings, the plot shown in the viewer is immediately adjusted accordingly.
Languages 2024, 9, 35 3 of 16

Visible Vowels can be used in several research fields, such as phonetics, sociolinguistics,
dialectology, forensic linguistics, speech pathology and language acquisition. It is freely
available at https://www.visiblesounds.org/ (accessed on 18 October 2023) where a tutorial
can be found as well.

1.4. Case Study

We focus on variation in the pronunciation of vowels spoken by L2 speakers and
how they relate to the vowels pronounced by L1 speakers. We use a data set, which was
compiled by Paolo Mairano, that contains vowel measurements of L1 speakers of French
and of three groups of L2 learners of French, with Italian, Spanish and English as L1 (see
Section 2.1). Using this data set, we demonstrate how Visible Vowels can be used as a tool
in L2 research. We do this by answering the following questions:
1. What are the differences in F1 and F2 between the French vowels of Italian, Spanish
and English L2 speakers and French L1 speakers?
2. Do the vowel spaces of Italian, Spanish and English L2 speakers of French differ from
the vowel space of French L1 speakers?
3. How do the vowel systems of the French L2 speaker groups relate to the vowel system of
the French L1 speaker group, and to each other, regarding the inter-vowel relationships?
4. What are the differences in duration between the French vowels of Italian, Spanish
and English L2 speakers and French L1 speakers?
French has a lot of nasal vowels, and a lot of L2 speakers have difficulties in acquiring
the nasal vowels. However, the automatic formant extraction methods used for the data
sets do not provide valid values for nasal vowels, and therefore nasal vowels had to be
excluded from the analysis.
In Section 2, the data set is described. We pay special attention to the removal of
outliers and the normalization of the vowel formant measurements. In Section 3, different
ways of visualizing vowel variation are shown and the four research questions are answered.
In Section 4 we close the paper with some concluding remarks.

2. Methodology
2.1. Data Set
The learners of French data set was compiled by Paolo Mairano and includes 25 Italian
L1 speakers from the ProSeg corpus (Delais-Roussarie et al. 2018), 15 Spanish L1 speakers
from the COREIL corpus (Delais-Roussarie and Yoo 2010), 10 English L1 speakers from
the AixOx corpus (Herment et al. 2014) and 10 French L1 speakers from the same AixOx
corpus. Table 1 shows the distribution of speakers, split up for language group and gender.

Table 1. Distribution of the speakers, split up for language group L2 and L1 speakers of French)
and gender.

L2 English L2 Italian L2 Spanish L1 French

(AixOx) (ProSeg) (COREIL) (AixOx)
male speakers 7 4 8 4
female speakers 3 21 6 6

The English L2 speakers of French are L1 speakers of southern British English recruited
at the University of Oxford. They had a self-reported proficiency ranging from B1 to B2. The
Italian L2 speakers are from the northern part of Italy and were recruited at the University
of Turin. Their self-reported proficiency ranged from B1 to C1. The Spanish L2 speakers
were students at the Autonomous University of Mexico (UNAM), having a self-reported
proficiency that ranged from A2 to B2. For more details about the speaker groups, see
Mairano et al. (2023).
Languages 2024, 9, 35 4 of 16

All participants read the same text. Mairano et al. (2023) extracted the target vowels
automatically from the recordings by means of forced alignment. WebMAUS (Kisler et al.
2017) was used for the recordings of the L2 English speakers, and Easyalign (Goldman 2011)
for the other recordings. Subsequently, the alignments and transcriptions were verified
by the respective authors of the data sets using Praat (Boersma and Weenink 2021). The
transcriptions were minimally edited in order to reflect the target sounds, rather than
actual realizations.
The set includes measurements of the vowels /i/, /y/, /u/, /e/, /ø/, o/, /@/, /E/,
/œ/, /O/ and /a/. Within the group of Spanish learners, one speaker did not pronounce
that /œ/.
For all target vowels, the authors measured F1, F2, F3 and duration using Praat. The
formants were measured automatically and were not manually verified (Delais-Roussarie
et al. 2018; Delais-Roussarie et al. 2015; Herment et al. 2014). The formants were extracted
from the midpoint of each vowel to minimize coarticulation effects. The Burg method was
used in a band lower than 5.5 kHz for women and 5 kHz for men.
In order to eliminate formant measurement errors, we applied the interquartile range
method to find outliers. This was done separately for F1, F2 and F3 and per language
group, and within each group per gender. The quartile range (IQR) is calculated as the
third quartile (Q3) minus the first quartile (Q1). Then, the lower fence is Q1 − 1.5 × IQR,
and the upper fence is Q3 + 1.5 × IQR. Cases where formant measurements—whether F1
and/or F2 and/or F3—were below the lower fence or above the upper fence were removed
from the data set.
Table 2 shows the number of realizations per vowel and language group both before
and after the outliers have been removed. In total, 198 outliers (8%), equally spread over
the vowels, were removed.

Table 2. Number of vowel tokens averaged over the speakers per language group. Incl. = including
outliers, excl. = after removal of outliers.

Total Total L1 French L2 English L2 Italian L2 Spanish

Incl. Excl. Incl. Excl. Incl. Excl. Incl. Excl. Incl. Excl.
a 524 480 128 123 135 110 145 137 116 110
e 343 311 87 82 86 70 86 81 84 78
@ 356 327 71 68 90 75 104 97 91 87
E 388 365 98 92 102 92 95 91 94 90
i 314 285 76 64 78 69 86 82 75 70
o 79 71 17 16 20 16 28 26 15 13
ø 55 51 13 13 16 14 14 13 12 11
œ 41 38 11 10 16 14 9 9 5 5
O 145 137 33 33 36 31 50 48 26 25
u 110 101 28 27 29 23 26 25 27 26
y 123 114 24 23 29 24 47 45 23 22
2478 2280

2.2. Scale Conversion and Normalization

Scale conversion methods aim to represent frequencies and frequency differences of
pitch and/or formants in accordance with the perception of these differences. In Visible
Vowels, five scales are available for formant measurements: Hz, bark (three versions),
ERB (three versions), ln and mel (two versions). Additionally, 19 speaker normalization
methods are available. Different speakers have varying vocal tract lengths and shapes,
which can affect the formant frequencies. Vowel formant normalization aims to remove
the effects of those differences, making it easier to compare and analyze formants across
different speakers.
Languages 2024, 9, 35 5 of 16

In order to find the best combination of a scale conversion method and a speaker
normalization method for the data set that is uploaded by the user, an evaluation tab is
included in Visible Vowels. Using this tab for all possible combinations of a scale conversion
method and a speaker normalization method, it can be determined how effectively they
(1) preserve phonemic information, (2) minimize anatomical/physiological information
and (3) preserve sociolinguistic information in the formant measurements of vowels. These
criteria were introduced by Adank (2003) and Adank et al. (2004).
The criteria were tested by two different approaches. In the first approach, it was tested
(1) how well the acoustic variables can predict the phoneme that they represent, (2) how
poorly they predict the anatomical differences and (3) how well they predict sociolinguistic
distinctions. To this end, linear discriminant analysis (LDA) was used. In the second
approach, it was measured (1) how well phonemic distinctions explain the variance in the
acoustic variables, (2) how poorly anatomical differences are explained by the acoustic
variables and (3) how well sociolinguistic variables explain the acoustic measurements. For
that purpose, multivariate analysis (MANOVA) was used.
The same criteria and approaches are used in the evaluation tab of Visible Vowels.
However, a proper use of (parametric) MANOVA would require checking its assumptions:
independence of observations, randomly sampled data, the dependent variables should be
normally distributed within groups and the population covariance matrices of each group
should be equal. It cannot be assumed that the data that are uploaded to Visible Vowels
by the users will always satisfy (all of) these assumptions, and checking them for each of
the large number of MANOVAs that are carried out in the evaluation tab would make
the procedure complex and cumbersome. Therefore, in Visible Vowels, non-parametric
MANOVA is used as implemented in the function adonis2 in the R package vegan.
When using the evaluation tab, vowels that are not found across all speakers are
automatically excluded in order to ensure that the procedures are run on the basis of the set
of vowels that are found across all speakers. A notice listing the excluded vowels is given.
In our case, the vowel /œ/ is excluded.
We submitted the learners of French data set to the evaluation tab twice, one time
without and one time including F3. We have two reasons for this. First of all, the set of
normalization methods that can also handle F3 is smaller than the set of normalization
methods that are suitable for normalizing F1 and F2. The second reason is that a nor-
malization method that is well evaluated for normalizing F1 and F2 measurements is not
necessarily well evaluated for normalizing F3 measurements as well. In Table 3, for each
criterion, the ‘winning method’ is given. For a detailed explanation of the methods, see
Voeten et al. (2022). For the method of Johnson, see Johnson (2018, 2020).

Table 3. Evaluation results of scale conversion methods and speaker normalization methods. The
winning combinations are shown.

F1 + F2 F1 + F2 + F3
Highest Explained Highest Explained
Best Prediction Best Prediction
Variance Variance
Lobanov
phonemic Lobanov Hz Johnson Hz Nearey I Hz
Hz/bark I/mel II
Heeringa & Van de Heeringa & Van de Velde Gerstman bark
anatomic Lobanov bark I
Velde II bark II II bark II I/ERB II
Heeringa & Van de
socioling. Nearey I Hz Nearey II Hz LABOV II Hz6
Velde II ln
Languages 2023, 8, x FOR PEER REVIEW 6 of 17

bark II Velde II bark II

Languages 2024, 9, 35 6 of 16
socioling. Heeringa & Van de Velde II ln Nearey I Hz Nearey II Hz LABOV II Hz6

In
In L2
L2 research,
research, we we are interested in
are interested how well
in how well L2 L2 learners
learners pronounce
pronounce the the vowels
vowels and
and
how well the vowels are distinguished from each other, especially
how well the vowels are distinguished from each other, especially in cases where in cases where L1 andL1
L2
andsystems
L2 systems do not match.
do not match.Therefore,
Therefore,thethefirst
firstcriterion
criterion‘phonemic’
‘phonemic’isisrelevant
relevant here.
here.
Furthermore,
Furthermore, we wehave
have to choose
to choose between
between ‘best prediction’
‘best prediction’ and ‘highest
and ‘highest explainedexplained
variance’.
variance’. While the first approach focuses more on measuring
While the first approach focuses more on measuring the quality of the normalization, the quality of thethe
normalization,
second approach therather
second approachthe
addresses rather addresses
question of how thethe
question
relevantofinformation
how the relevant
in the
information in the acoustic
acoustic measurements can bemeasurements can befrom
optimally separated optimally
noise, separated from noise,errors.
such as measurement such
as measurement errors. Because we would like to be able to detect vowel
Because we would like to be able to detect vowel distinctions as well as possible, we opted distinctions as
well as possible, we opted for the second approach. This means
for the second approach. This means that for visualizations based on F1 and F2 the raw that for visualizations
based on F1 and
measurements areF2normalized
the raw measurements
with ‘Johnsonare Hz’,normalized with ‘Johnson
and for visualizations thatHz’,
alsoand
usefor
F3
visualizations
the measurements that are
alsonormalized
use F3 the measurements
with ‘Nearey I are Hz.’normalized with ‘Nearey I Hz.’
The effect of normalizing
normalizing the the data
data is
is visualized
visualized in in Figure
Figure 1.1. In
In each
each of
of the
the two
two plots,
plots,
the convex hulls of the vowel spaces of all speakers are are shown
shown in in F1/F2
F1/F2 space. In In the plot
on the
theleft,
left,
thethe original
original raw measurements
raw measurements are used. areIn used.
the plotInonthe plot on
the right, the right,
measurements
measurements normalized with Johson’s method (Johnson 2018,
normalized with Johson’s method (Johnson 2018, 2020) are used. In this graph, there 2020) are used. In this
is a
graph,
higher there
degree is of
a higher
overlapdegree
between of overlap between the envelopes.
the envelopes.

Figure 1.
1. Superimposed
Superimposedconvex
convexhulls
hullsofofthe
the speakers
speakers of the
of the ‘learners
‘learners of French
of French data data set’
set’ on theonbasis
the
basis of unnormalized data (left) and on the basis of measurements normalized by Johnson’s
of unnormalized data (left) and on the basis of measurements normalized by Johnson’s normalization
normalization
method (right).method (right).
The hulls The hulls arefrom
are distinguished distinguished
each other from each other
by different bythese
colors, different
colorscolors,
were
these colors were randomly assigned
randomly assigned to the hulls. to the hulls.

3. Results
Results
3.1. What
3.1. What Are
Are the
the Differences
Differences in
in F1
F1 and
and F2
F2 between
between the
the French
French Vowels
Vowels of
of Italian,
Italian, Spanish
Spanish and
and
English L2
English L2 Speakers
Speakers and
and French
French L1
L1 Speakers?
Speakers?
Figure 2 gives an overview of the vowel systems of the L1 of the language groups:
Figure 2 gives an overview of the vowel systems of the L1 of the language groups:
(Parisian) French, British English, Italian and Spanish. In this section, we compare the
(Parisian) French, British English, Italian and Spanish. In this section, we compare the
vowels pronounced by the L2 speakers of French to those pronounced by the L1 speakers of
vowels pronounced by the L2 speakers of French to those pronounced by the L1
French. We expect that, in particular, French vowels that are not found in the L1 languages
speakers of French. We expect that, in particular, French vowels that are not found in the
of the learners are pronounced differently compared to the pronunciation of the L1 speakers
L1 languages of the learners are pronounced differently compared to the pronunciation
of French. The fewer vowels there are in the learners’ L1, the more difficult it may be for
of the L1 speakers of French. The fewer vowels there are in the learners’ L1, the more
the learners to pronounce the French vowels correctly. On the other hand, in English there
difficult it may be for the learners to pronounce the French vowels correctly. On the
are vowels which are not found in French, which may also influence the pronunciation of
other hand, in English there are vowels which are not found in French, which may also
the French vowels by the English learners. In Figure 2, the seven vowels of Italian seem
influence the pronunciation of the French vowels by the English learners. In Figure 2, the
to be the best match with the corresponding French vowels, but five oral (and four nasal)
seven vowels of Italian seem to be the best match with the corresponding French vowels,
vowels still need to be acquired.
but five oral (and
In Section four
3.1.1, wenasal) vowels
compare still need
the vowel plotstoofbethe
acquired.
English, Italian and Spanish learners
to the vowel plot of the L1 speakers of French and try to detect noticeable differences. In
Section 3.1.2, we investigate whether the vowels that occur in both French and the learners’
L1 act as magnets that attract the vowels that do not occur in their L1.
Languages 2024, 9, 35 7 of 16

(a) (b)

(c) (d)

Figure 2. Vowel charts of Parisian French, British English, North Italian and Spanish. (a) Vowels of
French (Image taken from https://commons.wikimedia.org/wiki/File:French_oral_vowel_chart.png,
IPA vowel chart for French vowels, 18 January 2008, public domain). From Fougeron and Smith
(1999, p. 79); (b) Vowels of British English (Image taken from https://www.wikiwand.com/en/
Received_Pronunciation#Media/File:RP_English_monophthongs_chart.svg, Monophthongs of a
fairly conservative variety of RP, 18 April 2009, licensed under the Creative Commons Attribution-
Share Alike 3.0 license, see: https://creativecommons.org/licenses/by-sa/3.0/). From Roach (2004,
p. 240); (c) Vowels of Italian (Image taken from https://commons.wikimedia.org/wiki/File:Italian_
vowel_chart.svg, IPA vowel Chart for Italian in SVG format, 17 April 2009, 16:35 (UTC), licensed under
the Creative Commons Attribution-Share Alike 3.0 Unported license, see: https://creativecommons.
org/licenses/by-sa/3.0/deed.en. The position of some vowel labels has been adapted). From Rogers
and d’Arcangeli (2004, p. 119); (d) Vowels of Spanish (Image taken from https://commons.wikimedia.
org/wiki/File:Spanish_vowel_chart.svg, Spanish vowel chart, 17 November 2017, licensed under the
Creative Commons Attribution-Share Alike 4.0 International license, see https://creativecommons.
org/licenses/by-sa/4.0/deed.en). From Ladefoged and Johnson (2010, p. 227).

3.1.1. Comparing Vowel Plots

For each group of L2 speakers—L1 speakers of English, Italian and Spanish—the
vowels are plotted in F1/F2 space, together with the vowels of the L1 speakers of French.
Thus, larger deviations can easily be found. The plots are visualized in Figure 3. In each
plot, and for each vowel, the formant values are averaged over the speakers.
In Figure 3, the L1 speakers of English are compared to the L1 speakers of French. In
1the plot, larger deviations are found for the vowels /œ/ and /O/. In Figure 2b we can see
that /œ/ and /O/ are not found in the phonological system of British English. Sounds
relatively close to /œ/ are English /@/ and /3:/. However, the L2 speakers pronounce the
/œ/ close to French /E/. Sounds relatively close to /O/ are English /O:/ and /6/. The L1
speakers of English pronounce the /O/ even higher than English /O:/, somewhere between
French /o/ and /u/.
In Figure 3, the French vowels of the Italian L1 speakers are plotted together with the
vowels pronounced by the L1 speakers of French. Larger deviations are found for /œ/ and
/@/, two phonemes that are not found in the phonological system of Italian (see Figure 2c).
Languages 2024, 9, 35 8 of 16

The Italian L1 speakers pronounce /œ/ much higher. The vowel /@/ is pronounced lower
and more to the front. Interestingly, a similar deviation is found for /ø/. Consequently,
French /@/ is close to French /ø/, and Italian /@/ is close to Italian /ø/.
In Figure 3, the French vowels pronounced by the L1 speakers of Spanish are plotted
in relation to the vowels pronounced by the L1 French speakers. The largest differences are
found for the vowels /o/, /@/ and /y/.
The vowel /o/ is also found in the Spanish phonological system, but is pronounced
somewhere between /o/ and /O/, which may explain why the Spanish L1 speakers
Languages 2023, 8, x FOR PEER REVIEW 8 of 17
pronounced French /o/ lower than the L1 speakers of French do.

Figure 3. French vowels pronounced by L1 speakers of French and by English, Italian and Spanish
Figure 3. French vowels pronounced by L1 speakers of French and by English, Italian and Spanish L2
L2 speakers of French. The formants are averaged over the speakers per group. The labels refer to
speakers of French. The formants are averaged over the speakers per group. The labels refer to the
the native languages of the speakers.
native languages of the speakers.
In Figure 3, the L1 speakers of English are compared to the L1 speakers of French.
The
In the vowels
plot, larger/@/ and /y/
deviations areare not for
found foundthe in the Spanish
vowels /œ/ and phonological
/ɔ/. In Figure 2b system.
we canThe
unstressed vowel /@/ is pronounced more to the front and is hardly
see that /œ/ and /ɔ/ are not found in the phonological system of British English. Sounds distinguished from the
Spanish L1 speaker’s pronunciation of /e/. The vowel /y/ is pronounced
relatively close to /œ/ are English /ə/ and /ɜː/. However, the L2 speakers pronounce the more backwards,
between
/œ/ close/i/ to and /u/.
French /ɛ/. Sounds relatively close to /ɔ/ are English /ɔː/ and /ɒ/. The L1
speakers of English pronounce the /ɔ/ even higher than English /ɔː/, somewhere between
3.1.2. Detecting Magnet Vowels
French /o/ and /u/.
In
In Section
Figure 3,1,thetheFrench
Nativevowels
Languageof theMagnet
Italian Theory
L1 speakersof Kuhl (2000) (Kuhl
are plotted togetheret al.
with2008)
was
the mentioned. The vowel
vowels pronounced categories
by the of a learner’s
L1 speakers of French. L1Larger
may ‘attract’
deviationsthe are
vowelsfound in for
the L2
language thattwo
/œ/ and /ə/, arephonemes
not found that in the
arelearner’s
not found L1.inInthe
order to find outsystem
phonological whether of this is reflected
Italian (see
in our groups
Figure 2c). Theof L2 speakers,
Italian we first determine
L1 speakers pronounce which /œ/ muchvowels are the
higher. Themagnets.
vowel /ə/ Sinceis we
do not have vowel
pronounced lower measurements
and more to theoffront. the L1Interestingly,
languages ofa the learners,
similar we consult
deviation is found thefor
plots
in Figure
/ø/. 2. For each
Consequently, FrenchL2 group, wetoselect
/ə/ is close Frenchthe/ø/,
vowels that are
and Italian /ə/ isfound both
close to in the
Italian /ø/.plot of
the mother
In Figuretongue of the
3, the learners
French vowels(either Figure 2bbyorthe
pronounced Figure 2c or Figure
L1 speakers 2d) andare
of Spanish in the
plotted in relation to the vowels pronounced by the L1 French
plot with the French vowels (Figure 2a). Accordingly, the ‘magnet vowels’ for English are speakers. The largest
differences
/i:/, /u:/, /e/,are found
/@/ and for/O:/,
the vowels /o/, /ə/
for Italian areand /y/.
/i/, /u/, /e/, /o/, /E/, /O/ and /a/, and for
SpanishTheare
vowel
/i/,/o/
/u/,is also
/e/,found
/o/ andin the
/a/.Spanish phonological system, but is pronounced
somewhere
Now, webetweenhave to/o/ and /ɔ/, whether
investigate which may theexplain
French whyvowels thethatSpanish
do notL1coincide
speakerswith
pronounced
the French /o/
‘magnet vowels’ arelower than by
attracted thethe
L1 ‘magnet
speakers vowels’.
of FrenchFor do. each L2 group and for each
The vowels /ə/ and /y/ are not found in the Spanish phonological system. The
unstressed vowel /ə/ is pronounced more to the front and is hardly distinguished from
the Spanish L1 speaker’s pronunciation of /e/. The vowel /y/ is pronounced more
backwards, between /i/ and /u/.

3.1.2. Detecting Magnet Vowels

 consult the plots in Figure 2. For each L2 group, we select the vowels that are found both
in the plot of the mother tongue of the learners (either Figure 2b or Figure 2c or Figure
2d) and in the plot with the French vowels (Figure 2a). Accordingly, the ‘magnet vowels’
for English are /iː/, /uː/, /e/, /ə/ and /ɔː/, for Italian are /i/, /u/, /e/, /o/, /ɛ/, /ɔ/ and /a/, and
Languages 2024, 9, 35 for Spanish are /i/, /u/, /e/, /o/ and /a/. 9 of 16
Now, we have to investigate whether the French vowels that do not coincide with
the ‘magnet vowels’ are attracted by the ‘magnet vowels’. For each L2 group and for
each ‘magnet
‘magnet vowel’,vowel’,
we measure we measure
the distance the todistance
the vowelsto thethatvowels
are notthat are vowels’.
‘magnet not ‘magnet We
measure
vowels’. this
We distance
measure twice; namely,twice;
this distance on thenamely,
basis of onthe the
measurements
basis of theofmeasurements
the L1 speakers of
of
theFrench (d1) andofon
L1 speakers the basis
French (d1)ofandthe on
measurements
the basis of of thethe L2 speakers (d2).
measurements If d2
of the L2isspeakers
smaller
(d2). d1,
than then
If d2 is we assume
smaller than thatd1,thethen
magnet vowel has
we assume thatattracted a vowel
the magnet that has
vowel is not found ina
attracted
the L1 language of the learners in the L2 group. We calculate d1
vowel that is not found in the L1 language of the learners in the L2 group. We calculate − d2, which gives positive
and negative values. The positive values represent attraction,
d1−d2, which gives positive and negative values. The positive values represent attraction, and the negative values
represent repulsion.
and the negative The distance
values represent between
repulsion.a pair
The of distance
vowels isbetween
calculated as the
a pair ofEuclidean
vowels is
distance
calculated between
as the the F1/F2 values
Euclidean distance of between
the vowels, thei.e., the values
F1/F2 square ofroot
theofvowels,
the sumi.e., of the
the
squared
square rootF1 and F2 sum
of the differences.
of the squared F1 and F2 differences.
We
We nownowmake make a few
a few comments
comments on our on approach.
our approach. First, the potential
First, ‘magnet ‘magnet
the potential vowels’
of Italian and Spanish are a subset of the set of French vowels,
vowels’ of Italian and Spanish are a subset of the set of French vowels, but English also but English also has
‘magnet vowels’ which are not found in the set of French
has ‘magnet vowels’ which are not found in the set of French vowels. Second, we vowels. Second, we excluded
the nasal vowels.
excluded the nasal Third,
vowels.the English
Third, the ‘magnet
English vowels’
‘magnet are vowels’
long (except for the
are long schwa),
(except for but
the
we still match
schwa), but we them
stillwith
match their
themshort French
with theircounterparts.
short FrenchFourth, since we
counterparts. do notsince
Fourth, have we L1
measurements of the three groups of learners, we assume
do not have L1 measurements of the three groups of learners, we assume that the that the location of the vowels
in their L2
location of plot corresponds
the vowels in theirtoL2 theplot
location of the vowels
corresponds in theirofL1the
to the location plot.
vowelsThese four
in their
points may be cause for concern. We therefore present the results
L1 plot. These four points may be cause for concern. We therefore present the results in in this section with some
reservations.
this section with The results are shown in Figures
some reservations. The results4–6. are
In each
shownplot,inthe possible
Figures magnet
4–6. In eachvowels
plot,
are
the possible magnet vowels are found on the x axis and they are compared to thein
found on the x axis and they are compared to the other vowels that are not found the
other
L1 of the learners. For each combination of vowels, a colored dot is shown. The redder the
vowels that are not found in the L1 of the learners. For each combination of vowels, a
dot, the more the ‘other vowel’ is attracted by the ‘magnet vowel’, and the bluer the dot,
colored dot is shown. The redder the dot, the more the ‘other vowel’ is attracted by the
the more repulsion.
‘magnet vowel’, and the bluer the dot, the more repulsion.

Figure 4.
Figure
Languages 2023, 8, x FOR PEER REVIEW 4. Magnet
Magnet vowels
vowels of
of the
the English
English learners
learners versus
versus French
French vowels
vowels not
not found
foundin
inEnglish.
English. The
10The
of 17
redder the dot, the more attraction, and the bluer the dot, the more repulsion.
redder the dot, the more attraction, and the bluer the dot, the more repulsion.

Figure5.5.Magnet
Figure Magnet vowels
vowels of the
of the Italian
Italian learners
learners versus
versus French
French vowels
vowels not in
not found found in The
Italian. Italian. The
redder
redder the dot, the more attraction, and the bluer the dot, the more removal.
the dot, the more attraction, and the bluer the dot, the more removal.
 Languages 2023, 8, x FOR PEER REVIEW 10 of 17

Languages 2024, 9, 35 10 of 16

For the English learners (Figure 4), we find that /y/ and /E/ are attracted by /@/.
For Italian and Spanish learners, most ‘other vowels’ seem to be attracted by multiple
‘magnet vowels’. This happens when the magnet vowels are located relatively close to each
other in the vowel space. In that case, we consider the ‘magnet vowel’ with the strongest
attraction as the real magnet. As such, looking in Figure 5 (Italian learners), we find that
/y/ is attracted by /u/, /@/ and /ø/ are attracted by /e/, and /œ/ is attracted by /u/. In
Figure 6 (Spanish learners), we find that /y/ and /ø/ are attracted by /u/, /@/ and /O/
Figure 5. Magnet vowels of the Italian learners versus French vowels not found in Italian. The
are attracted
redder bythe
the dot, /e/, and
more /E/ andand
attraction, /œ/
theare attracted
bluer the dot,by
the/o/.
more removal.

Figure6.6.Magnet
Figure Magnetvowels
vowelsof
ofthe
theSpanish
Spanishlearners
learnersversus
versusFrench
Frenchvowels
vowelsnot
notfound
foundininSpanish.
Spanish.The
The
redderthe
redder thedot,
dot,the
themore
moreattraction,
attraction,and
andthe
thebluer
bluerthe
thedot,
dot,the
themore
moreremoval.
removal.

For theall
Almost English learners
potential magnet (Figure
vowels 4),also
we act
findasthat /y/ and
magnets /ɛ/ are
in the plotsattracted by /ə/.and
of the Italian For
Italian learners,
Spanish and Spanish but inlearners,
the plot of most ‘other vowels’
the English learners, seem to be attracted
most magnet vowels doby notmultiple
attract
‘magnet
any other vowels’.
vowels. This may happens when theby
be explained magnet vowels
the fact that allare located
vowels of,relatively
respectively, closetheto
each other
Italian in the vowel
and Spanish vowelspace.systems In are
thatpotential
case, wemagnet
consider the ‘magnet
vowels and are vowel’
simply with subsetsthe
strongest
of the set ofattraction as the real
French vowels (see magnet.
Figure 2). AsHowever,
such, looking in Figure
English has also 5 (Italian
vowels learners),
that are not we
find that
found /y/set
in the is attracted
of Frenchby /u/, /ə/namely
vowels, and /ø//I/,
are/U/,
attracted
/3:/, by/2//e/, and
and /œ/These
/6/. is attracted
vowelsbymay /u/.
act
In as magnets
Figure as well,learners),
6 (Spanish but we cannotwe finddetermine
that /y/ this
and because we have by
/ø/ are attracted no /u/,
measurements
/ə/ and /ɔ/ are of
these vowels,
attracted norand
by /e/, are /ɛ/
these
andvowels
/œ/ areshared with
attracted byFrench.
/o/. Furthermore, it could play a role
that the English
Almost allmagnets
potentialthat we included
magnet vowels in theact
also plotashave phonological
magnets in the plots length while
of the the
Italian
French counterparts
and Spanish learners,dobut
not.in the plot of the English learners, most magnet vowels do not
attract any other vowels. This may be explained by the fact that all vowels of,
3.2. Do the Vowel
respectively, theSpaces
ItalianofandItalian, Spanish
Spanish voweland systems
English L2 areSpeakers
potentialof French
magnet Differ
vowelsfromandthe are
Vowel Space of French L1 Speakers?
simply subsets of the set of French vowels (see Figure 2). However, English has also
In this
vowels thatsection,
are not wefound
compare the shapes
in the and sizes
set of French of the vowel
vowels, namelyspaces of the
/ɪ/, /ʊ/, /ɜː/,L2
/ʌ/speakers
and /ɒ/.
with those of the L1 speakers.
These vowels may act as magnets as well, but we cannot determine this because we have
no First, as was doneofin Section
measurements 3.1, the measurements
these vowels, nor are these of multiple
vowels realizations
shared with of theFrench.
same
vowel are averaged per speaker and the speaker averages are averaged
Furthermore, it could play a role that the English magnets that we included in the plot for each vowel.
Then, for each vowel group, the convex hull is determined. A convex hull is the smallest
have phonological length while the French counterparts do not.
possible hull that encloses all points in a two-dimensional space. Assume we represent
vowels
3.2. Doas thenails
Vowelthat are hammered
Spaces in a wooden
of Italian, Spanish surfaceL2
and English correctly
Speakersrepresenting
of French Differtheirfrom
acoustic
the
relationships. Then, if we stretch a rubber band around the nails, this forms the edge of the
Vowel Space of French L1 Speakers?
convex hull (Wikipedia Contributors 2023).
In this section,
Additionally, we compare
the centroid of thethe shapes
vowels and sizes and
is determined of the
linesvowel spacesare
(or spokes) of drawn
the L2
speakers
from with those
this center of the
to each L1 vowels.
of the speakers.
The results are shown in Figure 7. For each language group, convex hulls and spokes
are shown for both male speakers and female speakers. The French vowel space of the
English learners looks more deviant from the L1 speakers’ vowel space than the shapes of
the French vowel spaces of the Italian and Spanish learners.
 Additionally, the centroid of the vowels is determined and lines (or spokes) are
drawn from this center to each of the vowels.
The results are shown in Figure 7. For each language group, convex hulls and
spokes are shown for both male speakers and female speakers. The French vowel space
Languages 2024, 9, 35 11 of 16
of the English learners looks more deviant from the L1 speakers’ vowel space than the
shapes of the French vowel spaces of the Italian and Spanish learners.

Figure7. 7.
Figure Convex
Convex hulls,
hulls, centroids
centroids and spokes
and spokes for thefor the groups
groups of L1
of L2 and L2speakers
and L1 of
speakers of French
French obtained
obtained on the basis of measurements that are normalized by Johnson’s method. The labels refer
on the basis of measurements that are normalized by Johnson’s method. The labels refer to the L1 of
to the L1 of the speakers.
the speakers.

3.3.How
3.3. HowDo Dothe
theVowel
VowelSystems
Systemsofofthe
theFrench
FrenchL2L2 Speaker
Speaker Groups
Groups Relate
Relate to to
thethe Vowel
Vowel System
System of
of the
French L1 Speaker
the French GroupGroup
L1 Speaker and toand
Each OtherOther
to Each Regarding the Inter-Vowel
Regarding Relationships?
the Inter-Vowel Relationships?
Huckvale
Huckvale(2004)(2004)introduced
introducedACCDIST
ACCDIST(accent (accentdistances),
distances),a ametric
metricwhere
wherea aspeaker’s
speaker’s
vowel
vowelsystems
systemsare arecompared
comparedby bycorrelating
correlatingthe theinter-vowel
inter-vowelsegment
segmentdistances
distances(see(seealso
also
Huckvale 2007). He used his metric for the accent classification of speakers
Huckvale 2007). He used his metric for the accent classification of speakers into 14 into 14 English
regional
Englishaccents
regionalof accents
the British Isles.
of the Huckvale
British Isles.developed
Huckvale his method with
developed speech technology
his method with speech
intechnology
mind. He in writes:
mind.“Thus speech“Thus
He writes: technology
speechcould benefitcould
technology from benefit
modeling fromtechniques
modeling
which are sensitive
techniques which are to the particular
sensitive to thecharacter of accent
particular variation.
character Better
of accent modeling
variation. of
Better
accents would
modeling allow recognition
of accents would allow systems to accommodate
recognition systems tospeakers
accommodatefrom aspeakers
wide range from of a
accents, including second language speakers”. On the other hand,
wide range of accents, including second language speakers”. On the other hand, he also he also thinks about
sociolinguistics when he writes:when
thinks about sociolinguistics “. . . better definitions
he writes: of accent
“… better groups could
definitions lead togroups
of accent new
sociolinguistic insights into how groups form and change”.
could lead to new sociolinguistic insights into how groups form and change”.
We use the ACCDIST measure as another way to quantify the differences between
the pronunciation of the French vowels pronounced by the French L1 speakers and the
English, Italian and Spanish learners. From the Native Language Magnet Theory, it may
be expected that the inter-vowel segment distances among the vowels of L2 speakers are
affected by their L1, as the vowels in their L1 tend to attract the vowels of the L2 language.
This causes the relationships between vowels to differ between L1 speakers and L2 speakers
and between different L2 speaker groups.
The ACCDIST measure is available in Visible Vowels. The inter-vowel segment
distances are calculated as Euclidean distances on the basis of their formant values (F1
and/or F2 and/or F3). The distance between any pair of speakers is 1 minus the correlation
of their respective inter-vowel segment distances.
In order to be able to use this method, for each speaker the same set of vowels should
be available. However, in the learners of French data set, the vowel /œ/ is missing for one
speaker (see Section 2.1). Therefore, that vowel was excluded in this analysis to obtain the
same set of vowels for each speaker.
 and/or F2 and/or F3). The distance between any pair of speakers is 1 minus the
correlation of their respective inter-vowel segment distances.
In order to be able to use this method, for each speaker the same set of vowels
should be available. However, in the learners of French data set, the vowel /œ/ is missing
Languages 2024, 9, 35 for one speaker (see Section 2.1). Therefore, that vowel was excluded in this analysis to
12 of 16
obtain the same set of vowels for each speaker.
In Visible Vowels, it is possible to calculate distances between groups. Assume a
group with speakers
In Visible Vowels, A,
it isBpossible
and C, to
and anotherdistances
calculate group with
betweenspeakers X and
groups. Y. Then,
Assume a groupthe
distance
with between
speakers A, B the
andtwo groups
C, and is calculated
another group with asspeakers
the averageX and distance
Y. Then,ofthe
thedistance
speaker
pairs AX,
between theAY,
twoBX,
groupsBY, isCX and CY.asTherefore,
calculated the averagewe can determine
distance and pairs
of the speaker visualize
AX, AY,the
relationships
BX, BY, CX andamong the four we
CY. Therefore, language groups and visualize the relationships among the
can determine
Once the groups
four language distances are calculated, the speakers (or speaker groups) can be classified
usingOnce the distances
cluster analysis ofaremultidimensional
calculated, the speakers
scaling.(or speaker
With groups) can bescaling,
multidimensional classified
the
using cluster
speakers are analysis
projectedofinmultidimensional
a two-dimensional scaling.
space With multidimensional
such that scaling, the
the distances between the
speakers
speakers are
are projected in a two-dimensional
proportionally reflected as closelyspace such that
as possible the distances
(Torgerson 1952, between
1958). the
speakers are 8proportionally
Figure reflected as closely
shows a multidimensional as possible
scaling plot that(Torgerson 1952,on
was obtained 1958).
the basis of
Figure distances
ACCDIST 8 shows a among
multidimensional
the speakersscaling plotlearners
of the that wasof obtained
French on the set.
data basisTheof
ACCDIST
Euclideandistances
distancesamong the speakers
were calculated on of
thethe learners
basis of F1,ofF2French
and F3 data set. The Euclidean
measurements. Those
distances were calculated
formant measurements on the
were basis of F1,
normalized withF2the
andNearey
F3 measurements.
I normalization Those formant
method (see
measurements
Section 2.2). were normalized with the Nearey I normalization method (see Section 2.2).

Figure8.8.Projection
Figure Projection of and
of L2 L2 and L1 speakers
L1 speakers of French
of French in a two-dimensional
in a two-dimensional space by space by Kruska’s
applying applying
Kruska’s non-metric multidimensional scaling to the ACCDIST distances that were
non-metric multidimensional scaling to the ACCDIST distances that were measured among the measured
among the speakers. Nearey I normalization was applied to the original measurements in Hz. The
speakers. Nearey I normalization was applied to the original measurements in Hz. The labels refer to
labels refer to the L1 of the speakers. The distances among the speakers in two-dimensional space
the L1 of the speakers. The distances among the speakers in two-dimensional space explain 81.6% of
explain 81.6% of the variance in the original ACCDIST distances.
the variance in the original ACCDIST distances.
InVisible
In VisibleVowels,
Vowels, four
four different
different kinds
kinds of multidimensional
of multidimensional scaling
scaling can becan be classi-
used: used:
classical multidimensional scaling, Kruskal’s non-metric multidimensional
cal multidimensional scaling, Kruskal’s non-metric multidimensional scaling, Sammon’s scaling,
Sammon’smapping
non-linear non-linear mapping
and and t-distributed
t-distributed stochastic
stochastic neighbor neighbor (t-SNE).
embedding embedding
The (t-SNE).
quality
The
of qualityto of
a scaling a scaling to
n dimensions (in n
ourdimensions
case: n = 2) (in
can our case: n by
be assessed = determining
2) can be assessed
how much by
variance in the original distances (in our case the ACCDIST distances) is explained by
the distances in the n-dimensional space. In our case, most of the variance is explained
using Kruskal’s non-metric multidimensional scaling (Kruskal and Wish 1978), so we use
this method.
The colored dots in Figure 8 represent the speakers of the four language groups: the
English, Italian and Spanish learners of French and the L1 speakers of French. Some of the
Italian L2 speakers cluster clearly with the L1 speakers. Although the groups of Italian
and Spanish learners and the French L1 speakers can be more or less recognized, they are
not sharply distinguished. Striking are the English learners who do not form a coherent
group, but it should be noted that within each L2 group there are large differences. Further
analyses could reveal whether this is linked to differences in French language skills.
A clearer picture is obtained by visualizing the relationships between the four groups,
as can be seen in Figure 9. The group of Italian learners of French is relatively close to the
group of L1 French speakers, and only differs on the 2nd dimension. The groups of Spanish
and English learners are more distant to the group of L1 French speakers, and differ on both
dimensions. The Spanish and English L2 speakers differ strongly on the 1st dimension.
 A clearer picture is obtained by visualizing the relationships between the four
groups, as can be seen in Figure 9. The group of Italian learners of French is relatively
close to the group of L1 French speakers, and only differs on the 2nd dimension. The
groups of Spanish and English learners are more distant to the group of L1 French
Languages 2024, 9, 35
speakers, and differ on both dimensions. The Spanish and English L2 speakers13differof 16
strongly on the 1st dimension.

Figure9.9. Projection
Figure Projection of
of the
the groups
groups of
ofL2
L2and
andL1 L1speakers
speakersofofFrench
Frenchinina atwo-dimensional space
two-dimensional by
space
applying Kruska’s non-metric multidimensional scaling to the ACCDIST distances that were
by applying Kruska’s non-metric multidimensional scaling to the ACCDIST distances that were
measured among the groups. Nearey I normalization was applied to the original measurements in
measured among the groups. Nearey I normalization was applied to the original measurements
Hz. The labels refer to the L1 languages of the speakers. The distances among the groups in two-
in Hz. The labels
dimensional spacerefer to the
explain L1 languages
97.1% of the
of the variance in speakers.
the originalThe distances
ACCDIST among the groups in
distances.
two-dimensional space explain 97.1% of the variance in the original ACCDIST distances.
3.4.What
3.4. WhatAre
Arethe
theDifferences
Differencesin inDURATION
DURATIONbetween betweenthe
theFrench
FrenchVowels
VowelsofofItalian,
Italian,Spanish
Spanishand
and
English L2 Speakers and French L1 Speakers?
English L2 Speakers and French L1 Speakers?
Visible Vowels
Visible Vowels includes
includesaatab tabforforvisualizing
visualizingvowel
voweldurations.
durations. In In Figure
Figure 10,
10, the
the
durationsofofthe
durations thevowels
vowels areare visualized
visualized for for
eacheach of four
of the the language
four language
groups. groups.
As wasAsdone
was
done
for
Languages 2023, 8, x FOR PEER REVIEW thefor the formants,
formants, first thefirst the durations
durations of multipleof multiple realizations
realizations of vowel
of the same the sameare vowel
14 of are
averaged 17
averaged
for for eachThen,
each speaker. speaker. Then,
in the plot,inthe
themean
plot,speaker
the mean speakerare
averages averages
shown are withshown with
their 95%
their 95% confidence
confidence intervals. intervals.

Figure 10.
Figure Averaged durations
10. Averaged durations inin milliseconds
milliseconds ofof vowels
vowels pronounced
pronounced by
by L2
L2 and
and L1
L1 speakers
speakers of
of
French. The labels refer to
French. The labels refer to the L1 languages of the speakers.

The plot
The plot shows
showsthat
thatthe L2L2
the speakers have
speakers longer
have vowels
longer than the
vowels thanL1the
speakers of French.
L1 speakers of
The largest durations were found for the L1 English speakers. The longer
French. The largest durations were found for the L1 English speakers. The longer vowels vowels of the
learners
of are very
the learners arelikely related
very likely to a slower
related speech
to a slower rate (Derwing
speech and Munro
rate (Derwing 1997)1997)
and Munro and
hyperarticulation.
and The largest
hyperarticulation. vowelvowel
The largest durations of the of
durations L1 the
English speakersspeakers
L1 English may indicate
may
that pronouncing
indicate the French
that pronouncing thewords
Frenchrequires
words the greatest
requires theeffort fromeffort
greatest them.from
In particular,
them. In
the durations
particular, theofdurations
the Frenchofvowels /ø/ and
the French /œ/,/ø/
vowels which
and do notwhich
/œ/, occur doin the
notEnglish
occur invowel
the
system, are much larger than the durations of the same vowels by the L1
English vowel system, are much larger than the durations of the same vowels by the L1speakers. The fact
that FrenchThe
speakers. vowels
fact have similar durations
that French vowels haveto English
similarshort vowelsto(Krzonowski
durations English shortet al. 2018)
vowels
supports our explanation.
(Krzonowski et al. 2018) supports our explanation.

4. Concluding Remarks
In Section 3.1, we compared the vowel plots of the English, Italian and Spanish
learners of French to the vowel plot of the L1 speakers of French. Examining the plots is
an easy way to detect vowels that are pronounced differently by the L2 speakers
Languages 2024, 9, 35 14 of 16

4. Concluding Remarks
In Section 3.1, we compared the vowel plots of the English, Italian and Spanish learners
of French to the vowel plot of the L1 speakers of French. Examining the plots is an easy
way to detect vowels that are pronounced differently by the L2 speakers compared to the
pronunciation of the L1 speakers, and to explore where these differences can be found.
Vowels that are not part of the L1 phonological system are often pronounced differently
by L2 speakers. Inspired by the Native Language Magnet Theory of Kuhl (2000), we
introduced magnet plots that relate vowels shared by the French phonological system and
the phonological system of the learners—the magnet vowels—to the vowels that are only
found in the phonological system of French. At a glance, it can be seen which vowels are
attracted to the magnets and which vowels become further away from the magnets. This
approach works best when the magnet vowels of the L1 of the learners of French are simply
a subset of the full set of the French vowels, as is the case for Italian and Spanish, but not
for English. This is only an exploratory analysis. Further research is necessary.
In Section 3.2, we compared the vowel spaces of the L1 and L2 speakers of French.
The shape of the French vowel space of the English learners differed most from the shape
of the L1 speakers’ vowel space.
In Section 3.3, the vowel systems of the four language groups were related to each
other regarding the inter-vowel relationships. We found that the group of Italian learners of
French is relatively close to the group of L1 French speakers, and that the groups of Spanish
and English learners are more distant to the L1 speakers. An explanation may be that the
Italian vowel system is most similar to the French vowel system (see Figure 2).
In Section 3.4, durations of vowels were considered. The vowel durations of the L2
speakers are larger than those of the L1 speakers of French. The longest durations were
found for the English learners of French. Particularly, the durations of the vowels /ø/
and /œ/ were found to be much larger than the durations of the same vowels by the
L1 speakers. This durational difference can be related by a slower speaking rate and/or
hyperarticulation, linked to the proficiency level of the speakers.
The English L2 speakers of French consistently differ most from the native French
speakers with respect to vowel spaces (Section 3.2), vowel systems (Section 3.3) and vowel
duration (Section 3.4). In addition to inherent differences between, on the one hand,
English (stress-timed, less overlapping and more vowels) and, on the other hand, Italian
and Spanish (syllable-timed, more similar systems), the differences in self-reported L2
proficiency level between language groups might play a role in the observed differences
between the English learners and the Italian and Mexican Spanish learners. With the data
at our disposal, we cannot determine the precise role of proficiency level, and we would
like to suggest this as an axe for future research to the authors of the respective corpora
(ProSeg, COREIL and AixOx), also taking into account possible cultural differences in
self-reporting and trying to relate self-reported proficiency with proficiency levels based on
actual language production.
By answering the four research questions, we demonstrated features of Visible Vowels
by which differences between L2 and L1 speakers can be visualized. We focused on formant
measurements, speaker normalization, vowel spaces, and comparison of vowel systems
and vowel duration, utilizing an existing data set of learners of French.
The figures generated by Visible Vowels in this paper can be a tool for speech therapists
and language teachers to identify which vowels differ in pronunciation from the intended
pronunciation, and in what respect they differ. The magnet plots in particular not only
show the deviation in pronunciation, but also provide an explanation for the deviation.
However, the feedback provided through these figures cannot be provided in real time,
but only after a series of words that include both the magnet vowels and the target vowel
multiple times have been spoken and processed.
We used Visible Vowels to compare L2 speakers of French with native speakers of
French. However, within a language there can also be regional accents. The formants
and duration can be compared between those accents, and the results explained by soci-
Languages 2024, 9, 35 15 of 16

olinguistic variables such as age, gender, education, income and urbanicity. Further, the
relationships between the vowel systems of accents can be measured and represented with
the ACCDIST measure, as Huckvale (2004) did for accents of English spoken in the British
Isles. One can then also try to find the explanatory variables for these relationships.
In addition to the functionality that we showed in this paper, f0 (pitch) contours, vowel
formant trajectories and vowel dynamics can be visualized. We are also developing Visible
Consonants, which can visualize variation in consonants. Although this program is still
under development, a first version is already available at https://www.visiblesounds.org/
(accessed on 18 October 2023).

Author Contributions: Conceptualization, W.H. and H.V.d.V.; methodology, W.H. and H.V.d.V.;
software development, W.H.; validation, H.V.d.V.; formal analysis, W.H.; investigation, W.H. and
H.V.d.V.; data curation, W.H.; writing—original draft preparation, W.H. and H.V.d.V.; writing—
review and editing, W.H. and H.V.d.V.; visualization, W.H. and H.V.d.V.; supervision, H.V.d.V.; project
administration, H.V.d.V. All authors have read and agreed to the published version of the manuscript.
Funding: This research received no external funding.
Institutional Review Board Statement: Not applicable.
Informed Consent Statement: Informed consent was obtained from all subjects involved in the study
by the authors of each corpus.
Data Availability Statement: French L1 and L2 data are part of three various corpora, some of which
made be made available by authors (see Section 2.1 for details about each corpus).
Acknowledgments: We thank Paolo Mairano for providing the learners of French dataset. We thank
the anonymous reviewers and the editors for their useful comments.
Conflicts of Interest: The authors declare no conflict of interest.

Notes
1 Available at: https://github.com/BYU-ODH/apeworm (accessed on 18 October 2023).
2 Available at: https://github.com/BYU-ODH/VowelCat (accessed on 18 October 2023).
3 See: https://www.ivanarehman.com/l2-tech-portfolio (accessed on 18 October 2023).
4 See: http://lingtools.uoregon.edu/norm/norm1.php (accessed on 18 October 2023).
5 See: https://depts.washington.edu/sociolab/VOIS3D/ (accessed on 18 October 2023).
6 In the formant tab of Visible Vowels this method can be found as ‘Labov et al. (2006) log-geomean II’. In the evaluation tab the
method is labeled as ‘LABOV II’. The implementation is described in Voeten et al. (2022).

References
Adank, Patricia Martine. 2003. Vowel Normalization: A Perceptual-Acoustic Study of Dutch Vowels. Nijmegen: Catholic University
of Nijmegen.
Adank, Patricia Martine, Roel Smits, and Roeland Van Hout. 2004. A comparison of vowel normalization procedures for language
variation research. The Journal of the Acoustical Society of America 116: 3099–107. [CrossRef] [PubMed]
Best, Catherine T., and Michael D. Tyler. 2007. Nonnative and second-language speech perception: Commonalities and complementari-
ties. In Language Experience in Second Language Speech Learning; in Honour of James Emil Flege. Edited by Ocke-Schwen Bohn and
Murray J. Munro. Amsterdam: John Benjamins Publishing Company, pp. 13–34.
Boersma, Paul, and David Weenink. 2021. Praat: Doing Phonetics by Computer [Computer Program]. Version 6.4.01. Available online:
http://www.praat.org/ (accessed on 18 October 2023).
Delais-Roussarie, Elisabeth, and Hiyon Yoo. 2010. The COREIL Corpus: A Learner Corpus Designed for Studying Phrasal Phonology
and Intonation. Paper presented at 6th New Sounds, Poznan, Poland, May 1–3, pp. 100–5.
Delais-Roussarie, Elisabeth, Fabián Santiago, and Hi-Yon Yoo. 2015. The extended COREIL corpus: First outcomes and methodological
issues. Paper presented at Workshop on Phonetic Learner Corpora, Glasgow, UK, August 12, pp. 57–59.
Delais-Roussarie, Elisabeth, Tanja Kupisch, Paolo Mairano, Fabian Santiago, and Frida Splendido. 2018. ProSeg: A Comporable Corpus
of Spoken L2 French. Paper presented at EuroSLA 2018, Münster, Germany, September 5–8.
Derwing, Tracey M., and Murray J. Munro. 1997. Accent, intelligibility, and comprehensibility: Evidence from four L1s. Studies in
Second Language Acquisition 19: 1–16. [CrossRef]
Languages 2024, 9, 35 16 of 16

Flege, James E. 1995. Second language speech learning theory, findings, and problems. In Speech Perception and Linguistic Experience:
Issues in Cross-Language Research. Edited by Winifred Strange. Timonium: York Press, pp. 233–77.
Fougeron, Cécile, and Caroline L. Smith. 1999. French. In Handbook of the International Phonetic Association: A Guide to the Use of the
International Phonetic Alphabet. Cambridge: Cambridge University Press, pp. 78–81.
Frostel, Harald, Andreas Arzt, and Gerhard Widmer. 2011. The vowel worm: Real-time mapping and visualisation of sung vowels in
music. Paper presented at 8th Sound and Music Computing Conference, Padova, Italy, July 6–9; Padova: Padova University
Press, pp. 214–19.
Goldman, Jean-Philippe. 2011. EasyAlign: An automatic phonetic alignment tool under Praat. Paper presented at 12th Annual
Conference of the International Speech Communication Association 2011 (INTERSPEECH 2011), Florence, Italy, August 27–31,
pp. 3233–36. [CrossRef]
Herment, Sophie, Anne Tortel, Brigitte Bigi, Daniel J. Hirst, and Anastassia Loukina. 2014. AixOx, a Multi-Layered Learners Corpus:
Automatic Annotation. In Specialisation and Variation in Language Corpora. Edited by Francisco Javier Diaz-Pérez and Ana
Díaz-Negrillo. Bern: Peter Lang, pp. 41–76.
Huckvale, Mark. 2004. ACCDIST: A Metric for Comparing Speakers’ Accents. Paper presented at International Conference on Spoken
Language Processing, Jeju, Republic of Korea, October 4–8, pp. 29–32.
Huckvale, Mark. 2007. ACCDIST: An accent similarity metric for accent recognition and diagnosis. In Speaker Classification II: Selected
Projects. Berlin and Heidelberg: Springer, pp. 258–75.
Johnson, Keith. 2018. Vocal tract length normalization. In UC Berkeley PhonLab Annual Report. Berkeley: UC Berkeley PhonLab, vol. 14.
[CrossRef]
Johnson, Keith. 2020. The ∆F method of vocal tract length normalization for vowels. Laboratory Phonology 11: 10. [CrossRef]
Kendall, Tyler, and Erik R. Thomas. 2018. Vowels: Vowel Manipulation, Normalization, and Plotting. R Package Version 1.2-2.
Available online: https://CRAN.R-project.org/package=vowels (accessed on 18 October 2023).
Kisler, T., U. Reichel, and F. Schiel. 2017. Multilingual processing of speech via web services. Computer Speech & Language 45: 326–47.
Kruskal, Joseph B., and Myron Wish. 1978. Multidimensional Scaling. In Sage University Paper Series on Quantitative Applications in the
Social Sciences. Newbury Park: Sage Publications, pp. 7–11.
Krzonowski, J., F. Pellegrino, and E. Ferragne. 2018. Étude acoustique de la production de voyelles de l’anglais par des apprenants
francophones. Paper presented at Actes des XXXIIe Journées d’Études sur la Parole, Aix-en-Provence, France, June 4–8, pp. 525–31.
Kuhl, Patricia K. 2000. A new view of language acquisition. Proceedings of the National Academy of Sciences 97: 11850–57. [CrossRef]
Kuhl, Patricia K., Barbara T. Conboy, Sharon Coffey-Corina, Denise Padden, Maritza Rivera-Gaxiola, and Tobey Nelson. 2008. Phonetic
learning as a pathway to language: New data and native language magnet theory expanded (NLM-e). Philosophical Transactions of
the Royal Society B: Biological Sciences 363: 979–1000. [CrossRef] [PubMed]
Labov, William, Sharon Ash, and Charles Boberg. 2006. The Atlas of North American English: Phonetics, Phonology and Sound Change.
Berlin: Mouton de Gruyter.
Ladefoged, Peter, and Keith Johnson. 2010. A Course in Phonetics, 6th ed. Boston: Wadsworth Publishing.
Mairano, Paolo, Fabián Santiago, and Leonardo Contreras-Roa. 2023. Can L2 pronunciation be evaluated without reference to a native
model? Pillai scores for the intrinsic evaluation of L2 vowels. Languages 8: 280. [CrossRef]
McCloy, Daniel R. 2016. phonR: Tools for Phoneticians and Phonologists. R package Version 1.0-7. Available online: https://CRAN.R-
project.org/package=phonR (accessed on 18 October 2023).
Rehman, Ivana. 2021. Real-Time Formant Extraction for Second Language Vowel Production Training. Ph.D. dissertation, Iowa State
University, Ames, IA, USA.
Roach, Peter. 2004. British English: Received Pronunciation. Journal of the International Phonetic Association 34: 239–45.
Rogers, Derek, and Luciana d’Arcangeli. 2004. Italian. Journal of the International Phonetic Association 34: 117–21. [CrossRef]
Torgerson, Warren S. 1952. Multidimensional scaling. I. Theory and method. Psychometrika 17: 401–19. [CrossRef]
Torgerson, Warren S. 1958. Theory and Methods of Scaling. New York: Wiley.
Voeten, Cesko, Wilbert Heeringa, and Hans Van de Velde. 2022. Normalization of nonlinearly time-dynamic vowels. The Journal of the
Acoustical Society of America 152: 2692–710. [CrossRef] [PubMed]
Wassink, Alicia Beckford. 2006. A geometric representation of spectral and temporal vowel features: Quantification of vowel overlap
in three linguistic varieties. The Journal of the Acoustical Society of America 119: 2334–50. [CrossRef] [PubMed]
Wikipedia Contributors. 2023. Convex Hull—Wikipedia, the Free Encyclopedia. Available online: https://en.wikipedia.org/wiki/
Convex_hull (accessed on 4 August 2023).

Disclaimer/Publisher’s Note: The statements, opinions and data contained in all publications are solely those of the individual
author(s) and contributor(s) and not of MDPI and/or the editor(s). MDPI and/or the editor(s) disclaim responsibility for any injury to
people or property resulting from any ideas, methods, instructions or products referred to in the content.