Download as pdf or txt
Download as pdf or txt
You are on page 1of 25

The acquisition of rhotics

by child L2 and L3 learners

Alexandra Morales Reyes1, Begoña Arechabaleta-Regulez2 and


Silvina Montrul2
1University of Puerto Rico at Mayagüez / 2University of Illinois at Urbana-
Champaign

This study investigated how previous linguistic experience and universal strate-
gies guide the acquisition of phonology in the classroom. More specifically, it fo-
cused on the possible advantages that bilingual children have over monolingual
children. Thirty-four children completed a picture-naming task: 9 Spanish native
speakers, 19 English native speakers acquiring Spanish as L2 and 6 Korean-
English bilinguals acquiring Spanish as L3. Results indicated that in general the
children rapidly acquired native-like pronunciation of the Spanish rhotics, but
the Korean-English bilinguals outperformed the English-speaking children. We
propose that although previous linguistic knowledge plays a role in L2 and L3
acquisition, children are able to overcome transfer errors because they are guided
by universal developmental strategies from the initial stages of acquisition. We
suggest that if L3 learners have an advantage over L2 learners, this may be due to
their complex linguistic knowledge and higher metalinguistic competence.

Keywords: child language acquisition, second language, third language, rhotics,


Spanish

1. Introduction

Mastery of the sound system of another language is one of the greatest chal-
lenges that second language (L2) learners face in acquiring the target language.
Several studies have found that the younger the learner, the more native-like the
accent (Asher & Garcia, 1969; Moyer, 1999; Yeni-Komshian, Flege & Liu, 2000).
Although late learners who develop a native-like accent (Moyer, 2014) and early
learners with non-native accents do exist (Moyer, 2004; Moyer, 2007), in general,
adult L2 learners tend to exhibit a persistent foreign accent regardless of the length

Journal of Second Language Pronunciation 3:2 (2017), 242–266. doi 10.1075/jslp.3.2.04rey


issn 2215–1931 / e-issn 2215–194X © John Benjamins Publishing Company
The acquisition of rhotics by child L2 and L3 learners 243

of exposure. Overall, L2 learners who began learning the L2 in childhood are usu-
ally more successful at achieving a native or near-native accent (Abrahamsson,
2012). Many of these studies were about adults who acquired the L2 as children
and were tested in adulthood. However, there is also research that tested children
learning an L2 while they were learning the language. This research on child L2
acquisition suggests that although influence of the L1 is initially apparent in pho-
nology (Goldstein, 2004; Oyama, 1976), after a short period of time of exposure
to the target language, children’s foreign accent rapidly diminishes and their pro-
duction closely resembles native norms (Winitz, Gillespie, & Starcev, 1995). While
many factors may influence L2 phonology acquisition, age of acquisition has been
repeatedly shown to be the most significant factor predicting the degree of native-
like pronunciation (Flege et al., 1995a, Flege et al., 1995b; Granena & Long, 2013;
Piske, MacKay & Flege, 2001).
A highly influential theoretical model that has largely informed and guided
studies on the effects of age of acquisition in the L2 acquisition of pronunciation is
Flege’s (1995) Speech Learning Model (SLM). This model assumes that the mecha-
nisms available for learning a phonetic system remain available over the speaker’s
lifespan. The SLM accounts for the differences in pronunciation between child
and adult L2 learners by taking into account how the learners perceive the sounds
of the L2. The SLM links the likelihood of establishing new phonetic categories to
two variables: (1) cross-language perceived phonetic distance and (2) age of ac-
quisition. Sounds perceived as being distant from those of the L1 are expected to
be easier to acquire. L2 learners are more successful at establishing new phonetic
categories for these sounds. In contrast, the sounds perceived as being similar to
those in the L1 are predicted to be harder to acquire. L2 learners will need more
experience in the L2 to differentiate the L2 sound from the L1 sound and to create
a new phonetic category for it. The ability to perceive the sounds is also related to
age of acquisition. Young learners are hypothesized to discriminate the phonetic
differences between the L1 and L2 sounds better than late learners, even when the
sounds are very similar. Late learners are predicted to have difficulties discrimi-
nating sounds that are similar in the L1 and the L2, because their L1 phonologi-
cal system is more entrenched than in younger learners. While the SLM makes
specific predictions about what differences should be expected in early and late
L2 learners, most of its evidence is based on data from L2 learners with significant
L2 exposure and difficulties in perception. It is still an open question whether the
SLM would also be applicable to L2 learners at incipient stages of L2 development,
and whether difficulties with pronunciation of the L2 sound system are primarily
due to difficulties with perception.
When learning an L2, learners are constrained by their L1 at initial stages of
development, which often leads to transfer errors. At the same time, interlanguage
244 Alexandra Morales Reyes et al.

grammars also exhibit developmental errors, which cannot be traced back to the
learners’ L1. Developmental errors are consistently made by nearly all L2 learners
regardless of their L1 and are also found in native speakers of the language (e.g.,
overregularization errors with morphology such as take-taked). Developmental
errors in both L1 and L2 acquisition stem from universal cognitive or linguistic
mechanisms. The interplay of L1 and developmental patterns in the process of
acquiring the phonological system of a new language has been reported in early
studies (Ferguson & Debose, 1977; Hecht & Mulford, 1982; Wode, 1976), which
suggests neither of these factors alone explains L2 acquisition in children. The
Ontogeny Model (OM) (Major, 1987, 2001) argues that adult L2 learners rely on
the phonological systems of their L1 at early stages of acquisition and later resort
to universal developmental strategies. Transfer errors are frequent at early stages
of L2 acquisition but decrease with greater exposure to the L2. Developmental er-
rors, on the other hand, are initially rare but increase at later stages of L2 acquisi-
tion before possibly decreasing after intensive exposure to the L2.
The ways in which the L1 and universal strategies affect L2 phonology acquisi-
tion vary depending on the similarity and markedness of the sounds. For example,
the Spanish tap [ɾ] (e.g. hora ‘hour’) is similar to the American English alveolar flap.
Major (2001) proposed that when acquiring similar sounds, learners are affected
by the L1 but not by universal strategies. In contrast, marked sounds are sounds
that are complex and less common in human language. For example, the Spanish
trill [r] is considered to be a marked sound. When acquiring a marked sound,
universal strategies affect the acquisition of that sound more than the learner’s L1.
This study investigates the acquisition of rhotics in Spanish by school-age chil-
dren learning Spanish as a foreign language in a classroom setting in the United
States. Due to early exposure to the L2 phonology, children may be able to use
mechanisms similar to those used by child L1 learners and rely on universal de-
velopmental strategies from the beginning. At the same time, because L2 children
already have knowledge of a phonologically fully developed L1, they can also show
transfer effects similar to adult L2 learners. One of the aims of this study was to
evaluate how and to what extent previous linguistic knowledge and developmental
errors affect young L2 learners.
A unique feature of our study is that it tested both L2 learners and L3 learners
of Spanish. Previous studies have mostly examined children learning L2 English
(Amastae, 1977; Anderson, 2004; Goldstein, 2004). The second aim of our study
was to investigate the development of L3 phonology, which is a relatively unex-
plored area of study (Cabrelli-Amaro, 2012). Extending our investigation to child
L3 acquisition provides valuable new information about how bilingual children
acquire additional languages. Since children with knowledge of two languages
have more complex previous linguistic knowledge than children with knowledge
The acquisition of rhotics by child L2 and L3 learners 245

of only one language (Kopečková, 2014), we may be able to determine whether


the complexity of previous linguistic knowledge facilitates or impedes the acquisi-
tion of an L3 at a young age and with a limited amount of exposure to the target
language in a classroom context. Although content-based classes in a foreign lan-
guage program may emulate a naturalistic learning setting, the two settings are
not strictly comparable. They differ in both hours of daily exposure and nature
of input (e.g., teachers’ language proficiency). Therefore, our findings and con-
clusions are limited to child L2 and L3 acquisition in a foreign language context.
Before discussing the details of our experiment and the results, we begin with a
description of the rhotic systems of Spanish, English and Korean, the other native
language of the L3 children.

2. The rhotic systems of Spanish, English and Korean

Spanish has two rhotic sounds, the tap [ɾ] and the trill [r]. The tap “is produced
with a single rapid contact of the tip of the tongue against the alveolar region”
(Hualde, 2005, p. 44). The trill is commonly described as a sound produced with
two or three occlusions (Hualde, 2005; Lipski, 1990). In Spanish, both the tap and
the trill hold phonemic status in intervocalic context (e.g., perro ‘dog’ vs. pero
‘but’). In word-beginning position, the rhotic is produced as a trill (e.g., rana ‘frog’,
rosa ‘rose’) but in word final position, the distinction between the two sounds may
be neutralized. In this study, we only analyze the intervocalic context because it is
shared by both sounds and the contrast is phonemic.
English has a single tap that has allophonic status in intervocalic position, and
it is in complementary distribution with alveolar stops /t/ and /d/ (e.g., ‘better’,
‘muddy’). In English, the grapheme r is not pronounced as it is in Spanish, but is
articulated as an approximant [ɹ]. In a word like caramel, the r is produced as [ɹ]
whereas in Spanish caramelo ‘caramel’ the r is articulated as a tap. When produc-
ing the English approximant, there is no contact between the tongue and the roof
of the mouth (Ladefoged, 2006; Ladefoged & Maddieson, 1996).
Korean also has a single tap, but it is an allophone of the lateral liquid /l/. In
Korean there is one phonemic liquid that has two phonologically conditioned al-
lophones (Ingram & Park, 1998). In the coda position, the lateral liquid is realized
as a dental lateral [l̪], which has the same manner of articulation as the English ‘l’
(e.g. [tal] ‘moon’). In intervocalic position and in word initial position, the lateral
liquid is realized as a flap [ɾ] that occurs with a single tap of the tip of the tongue
against the alveolar region (e.g. [taɾi] ‘leg’). Table 1 summarizes the phonological
status of rhotics in the three languages under consideration in this study: Spanish,
English, and Korean.
246 Alexandra Morales Reyes et al.

Table 1. Status of Rhotics Intervocalic Position in Spanish, English and Korean


Spanish English Korean
[ɾ] phonemic allophonic /t/ and /d/ allophonic /l/
complementary distribution complementary distribution
[r] phonemic

When either English or Korean speakers learn Spanish, they need to create a new
phonemic category for the trill (marked sound) and reassign the tap (similar
sound) to phonemic status. Acquiring both of these sounds (i.e., tap and trill) is
very challenging for both child L1 learners of Spanish and adult L2 learners.

3. The acquisition of rhotics in Spanish

For child native speakers of Spanish, the rhotics are one of the last sounds to be
mastered. Children display many difficulties producing these sounds in a target-
like manner, especially the trills (Melgar de González, 1976; Serra, 1979). At
the ages of 2;4 and 2;10, the tap and the trill occur less than 1% of the time in
children’s speech. These sounds are consistently omitted and when they are pro-
duced, children are not very accurate. For example, Anderson and Smith (1987)
tested 2-year-old Spanish monolingual children and found that their participants
only produced the correct rhotic sound 17% of the time. At this stage, children
commonly substitute [l] and [ʔ] for the tap (e.g.,/cala/ or /caʔa/ for /caɾa/ ‘face’),
whereas [h] and [l] are frequently substituted for the trill (e.g., /peho/ or /pelo/ for
/pero/ ‘dog’). Gómez-Fernández (2004) reported similar results and found that
rhotics are practically absent in children’s early speech (e.g., 12 to 18 months of
age), and they start appearing more frequently when children reach the age of
2. Approximation to the adult norm is very gradual and by preschool age, dif-
ficulties with these sounds decrease. However, children still make substitution er-
rors, producing [l] instead of the tap or the trill. After the age of 4, the child’s
speech is closer to that of an adult, but it is not until approximately age 7 that they
show mastery of the rhotics, producing less than 10% of non-target articulations
(Bosch, 1983). In places where Spanish is in contact with another language (e.g.
with English in the US), the tap and trill phonemic distinction in child heritage
speakers is often maintained in duration only, rather than in manner of articula-
tion (e.g. Henriksen, 2015).
Spanish rhotics are also very challenging for L2 learners of Spanish. Studies
of L1-English L2-Spanish adults at early stages of acquisition suggest that adults
are overly influenced by their L1s and produce both Spanish rhotics as the English
The acquisition of rhotics by child L2 and L3 learners 247

[ɹ] (Olsen, 2012). L2 learners require a great deal of exposure to Spanish in order
for these transfer errors to decrease. As in L1 acquisition, research also shows that
the tap and the trill are not acquired in the same manner (Face, 2006; Olsen, 2012;
Rose, 2010a). In the case of the tap, beginner level learners produce the American
voiced alveolar approximant [ɹ] in phonological environments where the tap is
needed (Major, 1987; Rose, 2010a). Learners at an intermediate level start pro-
ducing substitutions (not L1-related) when trying to produce the tap: they either
produce [r], glide, trill, and lateral, or simply delete the sound (Face, 2006; Major,
1987). Only learners that have been exposed to Spanish for a long period of time
are capable of producing the Spanish tap in a native-like manner.
Beginner adult L2 learners tend to also produce the American voiced alveolar
approximant [ɹ] in phonological environments where the trill is needed (Rose,
2010a). However, adult L2 speakers are also affected by universal strategies and
produce more developmental errors with the trill than with the tap. They over-
generalize the production of the tap, which they have already acquired, and use it
where the trill is expected. This is similar to what native speakers of Spanish do,
who also overgeneralize the tap to the trill, and go through this stage in the acqui-
sition of rhotics (Carballo & Mendoza, 2000). These results support Major’s OM,
which hypothesizes that learners will rely more on universal strategies for the trill
because it is a marked sound.
The combined findings from all these studies also suggest that L2 learners
initially filter the L2 through the L1 phonological system, but as they develop a
greater competence in the L2, L1 transfer gradually becomes less influential. That
is why, at first, native English speakers produce the American voiced alveolar ap-
proximant in the context where the trill and tap are expected. Then, they gener-
alize [ɾ] to all rhotic contexts and, finally, as the learners’ proficiency improves,
their ability to differentiate between [ɾ] and [r] increase. These findings suggest
that difficulties with the trill are greater than those observed for the tap at all
levels of proficiency (Face, 2006; Major, 1987; Olsen, 2012; Reeder, 1998; Rose,
2010a). Learners’ problems with the trill may be due to the articulatory demands
involved in producing this sound. The trill is one of the last sounds to be mastered
by L1 children and adult native speakers also display variation in trill productions
(Blecua, 2001; Lewis, 2004; Solé, 2002; Widdison, 1998).
Most of the studies that have evaluated the production of the Spanish rhotics
have measured speakers’ accuracy. They have assumed that target-like productions
of the tap are the only productions in which a single occlusion is involved, whereas
the trill is produced with two or more occlusions. As pointed out by Rose (2010a)
it is not fair to only take into consideration those two variants because not even
native speakers produce the rhotics consistently. Rose questioned this methodol-
ogy and identified seven phonetic categories that account for the possible variants
248 Alexandra Morales Reyes et al.

of both the tap and the trill produced by native and non-native speakers. Four
variants are associated with the tap (i.e., tap, perceptual tap, lenited tap, alveolar
approximant), and three variants to the trill (trill, tap+ and assibilated variant).
Variants of the Phonological Tap (Rose, 2010a)
– Tap: This is the prototype sound in which a single occlusion is clearly observed
in the spectrogram.
– Perceptual Tap: This refers to the realization in which the place of the tap can
be identified in the spectrogram but there is no visible closure because the
distinction between the tap sound and the vowels surrounding it is unclear.
– Lenited Tap: This does not have any closure and is phonetically closer to the
American approximant than to the Spanish tap.
– Alveolar Approximant: This is the American English approximant sound,
which is different from the Spanish rhotics.
Variants of the Phonological Trill
– Trill: This is the prototypic category in which two or more closures are ob-
served in the spectrogram.
– Tap+: This refers to the allophone with an initial tap followed by a frication
realization.
– Assibilated Variant: This displays frication without any moments of total clo-
sure.
Rose found that for the Spanish taps, the phonetic tap and the perceptual tap are
the most commonly produced variants by native speakers and that for the Spanish
trill, productions are within the trill and the tap+. Rose added that native speak-
ers of Spanish make a clear distinction between the tap and trill sounds and do
not produce any variant of a tap in a context where a variant of a trill is needed
and vice versa.
While the studies discussed above have all focused on the acquisition of the
Spanish rhotics by English speakers (Face, 2006; Major, 1987; Olsen, 2012; Reeder,
1998; Rose, 2010a, 2010b) we know of no study that has looked at the acquisition
of these sounds by adult Korean learners or by child L2 learners.
In L3 acquisition, speakers may be affected not only by their L1 and develop-
mental effects but also by their L2. While in L2 speakers cross-linguistic influence
can only happen between two languages, in L3 speakers there are three linguistic
systems that can interact and different models of cross-linguistic influence have
been advanced. Many studies have argued for an L2 status factor (e.g. Tremblay,
2007; Wrembel, 2010). According to these studies, when acquiring an L3, learners
are primarily affected both negatively and/or positively by the L2 phonological
The acquisition of rhotics by child L2 and L3 learners 249

system (Cabrelli-Amaro, 2012). However, there are also studies with mixed re-
sults, suggesting a combined influence from both native and non-native languages
(Wunder, 2010) and from any previously acquired languages, that is a combined
effect (De Angelis, 2007). It appears that for adults, the early stages of acquisition
are mainly influenced by cross-linguistic influence rather than by universal strate-
gies, but we are not aware of any study on L3 rhotics by adults.
Kopečková (2014) is the only study that has explored the acquisition of Spanish
rhotics in child L3 learners. The children who participated in this study were
German native speakers who had English as an L2 and were acquiring Spanish as
an L3. The participants were 12–14 years old. The objectives of the study were (1)
to analyze whether learners were able to differentiate the rhotics, (2) to look for the
source of transfer in the L3; and (3) to test whether learners displayed L3 regressive
transfer. Results showed that the participants could differentiate the Spanish rhotics
and most of them were able to produce the tap while they were still in the process
of correctly producing the trill. Results also suggested that transfer does not come
exclusively from learners’ L1 or L2 but instead, as explained by Kopečková (2014)
“L3 phonological systems consist of a balance between all the language in contact
and the universal properties of language” (p. 220). Finally, evidence of regressive
influence was not observed. The L3 (i.e., Spanish) did not appear to have an effect
in the production of the rhotics of the L2 (i.e., English) or the L1 (i.e., German).
One reason for the success obtained by the participants in Kopečková’s study
could be L3 phonological awareness. Bilingual children who are acquiring a third
language have already gone through stages of learning an L2. When compared
to monolingual children, they have a better understanding of the process of lan-
guage learning and have developed metalinguistic skills that monolingual children
may not have. Phonological awareness, the ability to identify sounds that comprise
words in language, is one such metalinguistic skill. Research has shown that bilin-
gual children acquire new sounds fasters than monolingual children due to their
more developed and greater phonological awareness (Chen et al., 2004; Georgia,
2007). Kopečková suggests that participants’ multilingual skills combined with the
teaching methodology that the participants received could explain their success
on acquiring the Spanish phonological system so quickly.
While Kopečková’s study introduces relevant information for the study of L3
phonology acquisition in general, it is doubtful whether the results fully describe
what happens in child L3 acquisition. The participants in the study were classified
as child L3 learners, but they were 11 or 12 years old when they were first exposed
to Spanish. The exact age at which child L2 acquisition and L3 acquisition starts
and ends is still a matter of debate; however, most researchers agree that child lan-
guage acquisition happens sometime after birth but before puberty (Lennenberg,
1967; Meisel, 2008; Schwartz 2003; Unsworth, 2005, 2008). Finally, Kopečková’s
250 Alexandra Morales Reyes et al.

study only described the very early stage of L3 phonological acquisition, and the
path followed by multilingual learners and the different stages they go through
during phonological acquisition remains to be investigated.

4. The study

The main objective of this study was to determine whether and how school-age
children learning Spanish as L2 and L3 in an instructed foreign language con-
text acquire Spanish rhotics. The study also considered the possible effects of the
number of previously acquired languages on the learning of a new language. The
following research questions were formulated for the present study:
1. How close to native norms can child L2 and L3 learners get at the ages of 4–6
years old?
2. How and to what extent are child L2 and child L3 learners affected by previ-
ous linguistic experience and/or developmental factors when acquiring the
Spanish tap and the trill?
3. Do child L3 learners have an advantage over child L2 learners in the acquisi-
tion of Spanish rhotics?
The following hypotheses were formulated:
1. We hypothesized that children will be able to reach native-like norms.
Phonological attainment is strongly affected by learners’ age, and young
learners tend to be more successful than adult L2 learners at mastering
the L2 phonology.
2. We expect L2 and L3 children who do not yet pronounce rhotics in Spanish at
native levels to be affected by both transfer and developmental strategies from
the beginning of the process. Due to their early age of acquisition we predict
L2 and L3 children to be more affected by universals than by their previous
linguistic experience. Furthermore, we expect no difference in the role played
by transfer and universals in either the tap or the trill. We expect both sounds
to follow a similar developmental path, contrary to what the Ontogeny Model
states. According to the OM transfer errors only appear at early stages and
developmental errors only increase at later stages of L2 acquisition.
3. Because multilinguals have already experienced learning a non-native lan-
guage, we assume (based on previous research) that they have more devel-
oped metalinguistic knowledge than monolinguals and bilinguals. Therefore,
we expect the L3 children to show target-like production earlier than
the L2 children.
The acquisition of rhotics by child L2 and L3 learners 251

4.1 Participants

In total, we tested 34 children between the ages of 4 and 8 with a mean age of 5.5,
25 of whom were English native speakers acquiring Spanish as a L2 (19 children)
or Korean-English bilingual speakers acquiring Spanish as a L3 (6 children). There
were also 9 native Spanish speakers serving as a comparison. The children acquir-
ing Spanish as a L2 or L3 were all learning Spanish at a private school in the United
States. They had been exposed to Spanish at an early age (Range: 4 to 6 years old
Mean: 4.7). The monolingual children were speakers of Peruvian Spanish and were
recorded in Lima, Peru. Table 2 provides more information about the participants.1

Table 2. Participants’ linguistic background


Native language N Mean age Age of Acquistion
(AoA) Spanish
Korean/English 6 5.6 (4–7) 4.8 (4–6)
English 19 5.4 (4–7) 4.6 (4–6)
Spanish 9 5.5 (4–8) –

The learners were exposed to Spanish as a foreign language four days a week for
one to two hours per day in school. Children were always taught by the same two
teachers: one teacher was a native speaker of Spanish who had been in the U.S.
for 6 years. The other teacher was born in the U.S and started learning Spanish at
the age of 6. Classes were taught entirely in Spanish and teachers followed a the-
matic instruction. Every week, the teacher chose a particular theme (professions,
music, etc.) and prepared activities integrating language, content, and culture
topics. Although the children did not receive explicit explanations of the Spanish
phonology, the Spanish alphabet was taught to the students through songs. The
learners’ amount of exposure measured in hours varied from low exposure (e.g.,
140 hours of exposure), intermediate exposure (e.g., 280 hours of exposure) to
high exposure (e.g., +500 hours of exposure). See Table 3. Therefore, some of the
participants were more advanced than others. Amount of exposure was taken as
a proxy for proficiency, since these two variables are positively correlated in child
L2 acquisition research: The greater the exposure to the target language, the higher
the accuracy displayed by the learner (Blom, Paradis, & Sorenson Duncan, 2012;
Paradis, 2011).

1. Thanks to Paloma Pinillo for helping with the data collection in Lima, Peru.
252 Alexandra Morales Reyes et al.

Table 3. Amount of exposure


Amount of exposure N Mean age AoA Spanish
low 11 5.3 (4–7) 4.8 (4–6)
intermediate 8 5.2 (4–7) 4.4 (4–6)
high 6 5.7 (4–7) 4.4 (4–5)

4.2 Materials and procedure

To elicit the use of the rhotics, children completed a picture-naming task. They
were asked to name pictures of objects/animals (e.g., a picture of a dog ‘perro’)
that elicited the intervocalic taps and trills. There were 11 words in total (6 with
an intervocalic tap and 5 with an intervocalic trill).2 Participants met with the
researcher individually in a quiet room and were asked to name the picture in
Spanish twice. Most of the participants were able to produce the correct name of
the picture without a prompt because the words selected were familiar to them.
All the words to be named had been taught during the first semester of Spanish
(e.g. colors, animals, forms etc.). When the children could not remember the word
presented or did not know its name, the researcher would say the word and the
child would repeat it twice. In these cases, the second production was taken into
consideration in order to analyze the most spontaneous production.

4.3 Analysis

In total 370 words were analyzed: 202 words had an intervocalic tap and 168
words had an intervocalic trill. All of these occurrences were examined acous-
tically (Boersma, 2012). The utterances were classified following the analysis in
Rose (2010a) described in Section 2. Due to the small sample of participants, the
number of analyzed utterances is limited. Therefore, the results presented in the
next section are suggestive rather than definitive. A bigger sample size would be
needed to validate these findings.

4.4 Results

Two types of analyses were conducted. The first analysis was completed based
on the number of hours that children had been exposed to Spanish prior to the

2. The number of words with an intervocalic tap and with an intervocalic trill was not balanced
because the present study is part of a comprehensive project on Spanish L2 acquisition by chil-
dren. The words were selected to test not only the Spanish rhotics, but other Spanish sounds.
The acquisition of rhotics by child L2 and L3 learners 253

testing. The second analysis was based on the children’s language background.
Some of the participants were Korean-English bilingual children (i.e., bilinguals
learning Spanish), while others were English native speakers (i.e., monolinguals
learning Spanish). Having two groups of speakers with different language back-
grounds made it easier to assess whether universal strategies or the L1 played a
role in the first stage of phonological acquisition for children learning an L2. Also,
because of a comparison between a monolingual and bilingual group, results can
verify what previous research on L2 and L3 acquisiton children has stated: that
children learning an L3, due to a more developed metalinguistic awareness, learn
faster than monolingual children.

4.4.1 Analysis by time of exposure

4.4.1.1 Intervocalic tap production. Table 4 illustrates the total number of times
(N) each variant was used by the children to produce the intervocalic tap accord-
ing to the number of hours exposed to Spanish

Table 4. Intervocalic tap realizations


Low exposure Intermediate High exposure Native speakers
(n = 11) exposure (n = 8) (n = 6) (n = 9)
Variant N % N % N % N %
Tap 35 54.6 7 14.6 17 47.2 34 62.9
Perceptual Tap 9 14 10 20.8 11 30.5 14 25.9
Lenited Tap 10 15.6 11 22.9 1    2.7 – –
Alveolar App. 4    6.2 10 20.8 1    2.7 1    1.8
Trill 1    1.5 3    6.3 1    2.7 – –
Tap+ 3    4.6 3    6.3 2    5.5 1    1.8
Assibilated – – – – – – – –
Lateral [l] 1    1.5 1    2.1 – – – –
Occlusive [d] 0    0 1    2.1 – – – –
Other 1    1.5 2    4.2 3    8.3 4    7.4
Total 64 100 48 100 36 100 54 100

Table 4 shows that the Spanish-speaking children realized 88.8% of their phono-
logical taps as phonetic taps or perceptual taps, demonstrating that these are the
most common variants produced by young native speakers. Thus, these two vari-
ants were considered as the target-like variants for the production of the intervo-
calic tap. Among the non-native speakers, 68.6% of the productions for the group
254 Alexandra Morales Reyes et al.

with the lowest time of exposure, 35.4% of the productions for the group with the
intermediate time of exposure, and 77.7% of the productions for the group with
the highest time of exposure were realized as taps or perceptual taps. Children with
the highest time of exposure to Spanish performed as the most native-like group,
followed by the children with the lowest time of exposure of Spanish. Interestingly,
the children with the intermediate time of exposure to Spanish produced the few-
est native-like realizations of the intervocalic tap sound. The findings suggest that
time of exposure has an effect on the learner’ production. To find whether there
was a significant relation between time of exposure and the frequency of the vari-
ant produced, we ran a Chi-Square test and it confirmed that there was a signifi-
cant association χ2 (16) = 36.11, p < .05. Thus, there is an association between the
time of exposure and likelihood of producing a native-like variant.
Moreover, the results indicate that native-like production of intervocalic taps
does not include the production of continuant variants (i.e., lenited taps and al-
veolar approximants). Children with the highest time of exposure to Spanish seem
to have acquired this rule by avoiding continuant variants in their intervocalic tap
production. Children with the lowest time of exposure to Spanish appear to be
more native-like than children with intermediate exposure in this aspect too; how-
ever, the former seem to be in the process of acquiring this phonological rule, since
they still produced several lenited taps (15.6%), but fewer alveolar approximants
(6.2%). Nonetheless, 43.7% of the productions for the group with intermediate ex-
posure were realized as continuant variants (i.e., lenited tap and alveolar approxi-
mants), which demonstrates that this group’s production was the least native-like.
The results also show that native speakers clearly distinguish between the
Spanish taps and the Spanish trills because they know in which phonological con-
texts the tap is needed and in which phonological contexts the trill is needed.
Only one native speaker, and on just a single occasion, produced a trill variant in a
context where a tap variant was expected. This time, the group of learners with the
highest and the lowest time of exposure to Spanish behaved in the most target-like
manner by producing a trill variant in contexts were a tap was required for 7.9%
and 7.1% of their production consecutively. Only 12.9% of the productions for the
group with intermediate exposure were trill variants in contexts were a tap was
needed. Although this might be an indication that this group cannot differentiate
the tap and the trill contexts, the production of trill variants in tap contexts is a
positive manifestation of rhotics acquisition.

4.4.1.2 Intervocalic trill production. Table 5 illustrates the total number of times
(N) each variant was used by the children to produce the intervocalic trill accord-
ing to the time of exposure to Spanish.
The acquisition of rhotics by child L2 and L3 learners 255

Table 5. Intervocalic trill realizations


Low exposure Intermediate High exposure Native speak-
(n = 11) exposure (n = 8) (n = 6) ers (n = 9)
Variant N % N % N % N %
Tap 4    7.5 4 10.3 1    3.3 – –
Perceptual Tap – – – – – – 1    2.2
Lenited Tap – – – – – – – –
AlveolarApp. 1    1.9 3    7.7 – – – –
Trill 3    5.6 3    7.7 9 30 21 46.6
Tap+ 16 30.1 4 10.3 9 30 9 20
Assibilated 23 43.4 22 56.4 6 20 8 17.8
Lateral [l] 6 11.3 – – – 10 4    8.8
Occlusive [d] – – 2    5.1 1    3.3 – –
Other – – 2    5.1 1    3.3 2    4.4
Total 53 100 40 100 30 100 45 100

Table 5 shows that the two variants the native Spanish speakers produced the most
for the intervocalic trill were the phonetic trill and the tap+. They produced these
two variants in 86% of their productions. Among the non-native speakers, the
group with the highest amount of hours exposed to Spanish was the most native-
like with 56.7% of their productions being trills and tap+, followed by the group
with the lowest amount of exposure to Spanish with 35.7% of their productions
being trills and tap+. The group with the intermediate time of exposure to Spanish
performed the least native-like with 18% of their productions realized as trills or
tap+. Similar to the tap production, the statistical analysis also indicated that there
was a significant association between time of exposure and trill production χ2
(12) = 25.55, p < .05.
The native speakers also produced numerous assibilated variants (17.8%). The
fact that child native speakers are producing this variant is an indication that this
is part of the process of acquiring the intervocalic trill. Non-native participants
also produced assibilated variants: The group in the most developed stage of lan-
guage acquisition (i.e., 500 hours) produced the fewest assibilated variants (20%),
followed by the group exposed to Spanish for the lowest time (43.4%). The group
with intermediate time of exposure to Spanish produced the most assibilated
variants (56.4%)
The results reflect the complexity of the trill and the challenges of acquiring
and mastering it. The use of the [l] for the intervocalic tap is a developmental er-
ror and an indication of the acquisition process. The native speakers produced the
256 Alexandra Morales Reyes et al.

rhotic as a [l] 4 times (8.8%). This developmental error also appears in the produc-
tion of the children with the highest time of exposure to Spanish (10%) and in the
production of the children with the lowest time of exposure to Spanish (11.3%). It
did not in the children with intermediate exposure to Spanish, who produced more
often the variants related to the L1 (e.g., alveolar approximant and occlusive /d/).
Finally, another developmental error is the production of tap variants in pho-
nological trill contexts. Child native speakers in this study produced a single vari-
ant of a tap in 2.2% of their total productions in a phonological trill environment.
This is also true for the group with the highest time of exposure to Spanish. Only
3.3% of their productions were produced as a tap in a context where a trill was
expected. By comparison, 9.4% of productions for the group with the lowest expo-
sure were tap variants in contexts were a trill was required, while 18% of produc-
tions for the group with intermediate exposure were tap variants.
To summarize, the analysis by amount of input to Spanish shows that the pro-
duction by children exposed to Spanish for the greatest amount of hours is the
most native-like for both the tap and the trill sounds. This group appears to have
most fully acquired these sounds. Surprisingly, the group with the lowest time of
exposure to Spanish was more native-like than the group with intermediate ex-
posure to Spanish. The former group seems to be in the process of acquiring the
Spanish rhotics and follows a very similar path to that followed by young native
speakers. Most of their errors are developmental errors also found in native speak-
ers. The latter group is very much affected by the L1 evidenced by errors that are
not related to universal strategies.

4.4.2 Analysis by L1
Due to the fact that all of the Korean-English speaking children belonged to the
group with the lowest time of exposure to Spanish, for this analysis we compared
them to the English-speaking children who had been exposed to the L2 for the
same amount of time.

4.4.2.1 Intervocalic tap production. Table 6 illustrates the total number of times
(N) each variant was used by the children to produce the intervocalic tap accord-
ing to their language background.
The acquisition of rhotics by child L2 and L3 learners 257

Table 6. Total Number of times (N) each variant was used by Korean/English vs. English
children to produce the intervocalic tap
Korean/English (n = 6) English (n = 5) Spanish speakers (n = 9)
Variant N % N % N %
Tap 22 62.8 13 44.8 34 62.9
Perceptual Tap 5 14.3 4 13.8 14 25.9
Lenited Tap 1    2.9 9 31 – –
Alveolar Approximant 3    8.5 1    3.4 1    1.8
Trill – – 1    3.4 – –
Tap+ 2    5.7 1    3.4 1    1.8
Assibilated Variant – – – – – –
Lateral [l] 1    2.9 – – – –
Occlusive [d] – – – – – –
Other 1    2.9 – – 4    7.4
Total 35 100 29 100 54 100

Between these two groups of learners, Korean-English bilinguals outperformed


the English native speakers in the production of the Spanish rhotics. Korean-
English children learning Spanish as L3 produced the highest percentage of na-
tive-like taps with 77.1% of their realizations produced as taps and perceptual
taps. In contrast, 58.6% of the productions by English-speaking children learning
Spanish as L2 were taps and perceptual taps. A noteworthy difference between
these two groups was that the English native speakers produced a high number of
lenited taps (31%) compared to the Korean participants (2.9%). As explained by
Rose (2010a), this variant is usually produced by native speakers in spontaneous
speech and is the first target-like variant to emerge for learners, followed by the
tap and perceptual tap. The results for both groups indicate that the L2 and the L3
children are in the process of acquiring target-like sounds, but the Korean-English
L3 learners were in a more advanced stage in the process.
Another common developmental error is to produce trill variants when a tap
is needed: 5.7% of the productions by the L3 children and 6.8% of the productions
by the L2 children were trill variants in tap contexts. These low percentages sug-
gest that although these two sounds are not found in any of the children’s native
languages, after only 140 hours of exposure to Spanish, the children can already
differentiate between the tap and trill environments, an aspect that is late-acquired
by adult L2 learners (Face, 2006). Our results also show that both groups are more
affected by developmental errors than by the L1 participants. None of the groups
produced alveolar approximants, the English approximant. Korean-bilingual
258 Alexandra Morales Reyes et al.

speakers produced a single lateral liquid sound (2.9%). Because in Korean [l] and
[ɾ] are allophones (in complementary distribution), more errors of this type would
be expected from the L1 influenced Korean-English children at the early stages
of acquisition.
Overall the Korean-English children were found to be more accurate and clos-
er to the native norm than the English-speaking children. Nonetheless, despite
Korean-English children being more target-like than the English native speakers,
no significant association was found between language group and the tap variant
produced by the children χ2(7) = 12.71, p = .08.

4.4.2.2 Intervocalic trill production. Table 7 illustrates the total number of times
(N) each variant was used by the children to produce the intervocalic trill accord-
ing to their language background.

Table 7. Total number of times (N) each variant was used by Korean/English vs. English
children to produce the intervocalic trill
Korean/English (n = 6) English (n = 5) Spanish (n = 9)
Variant N % N % N %
Tap 2    6.6 2    8.3 – –
Perceptual Tap – – – – 1    2.2
Lenited Tap – – – – – –
Alveolar App. 1    3.3 1    4.1 – –
Trill 1    3.3 2    8.3 21 46.6
Tap+ 12 40 4 16.6 9 20
Assibilated 11 36.6 12 50 8 17.7
Lateral [l] 4 13.3 3 12.5 4    8.8
Occlusive [d] – – – – – –
Other – – – – 2    4.4
Total 30 100 24 100 45 100

The Korean-English children performed differently than the English children:


43.3% of the productions by Korean-English children were trills and tap+,
while 24.9% of English children’s realizations were produced as trills and tap+.
Surprisingly, no significant association was found between language group and the
variant produced χ2(4) = 3.55, p = .47. Although the productions of the Korean-
English children were closer to those of the native children, the language group
(i.e., L2 or L3) was not related to the variant they produced.
The acquisition of rhotics by child L2 and L3 learners 259

Yet again, both groups were more influenced by developmental errors than
by L1 transfer. The Korean-English children (36.6%) and the English children
(50%) produced assibilated variants, a variant also realized by native speakers
(17.7%). Moreover, both groups produced assibilated variants when intending to
produce the intervocalic trill: 13.3% of the Korean children’s realizations were lat-
eralized and 12.5% of the bilinguals’ productions were lateralized. The Spanish
native speakers also lateralized the Spanish trill (8.8%). All in all, the analysis of
participants by L1 shows that (1) the Korean children acquiring Spanish as L3
had a more target-like production than the children acquiring Spanish as L2 and
(2) both groups of learners were more affected by universal strategies than by L1
transfer at initial stages of acquisition.

5. Discussion

The purpose of our study was to investigate the L2 and L3 acquisition of Spanish
phonology by school-age children in a foreign language context. In particular, we
focused on the production of rhotics in Spanish to examine the effects of amount
of exposure, previous linguistic knowledge and the potential operation of univer-
sal learning principles in their acquisition.
Previous research on adult L2 learners has shown that the most advanced
learners always produce the most target-like realizations. In our study we used
amount of exposure in hours as a proxy for proficiency, and we obtained similar
results with child L2 learners. We found that the children with the highest time of
exposure to Spanish were the most native-like. Their performance was very simi-
lar to the performance by the child Spanish native speakers. We believe that these
results answer our first research question and corroborate our first hypothesis:
Child L2 and child L3 learners after an exposure of 500 hours or more to Spanish
approach native norms and are able to master sounds that are absent or different
from their L1.
The second research question asked whether child L2 and child L3 learners
are affected by previous linguistic experience and/or developmental factors when
acquiring the Spanish rhotics and if so, how. Our results seem to suggest that
children are affected by both previous linguistic experiences and developmental
factors. Our results also seem to show that the developmental stages followed by
adult L2 learners are different from the developmental stages followed by child L2
learners. Studies on adult L2 learners’ acquisition (Face, 2006; Rose, 2010a) tend to
show a positive relation between the amount of exposure and the proficiency level;
their production becomes more native-like as proficiency level increases. Adults
are greatly affected by their previous linguistic experience at first. Only with time
260 Alexandra Morales Reyes et al.

and exposure to the L2 do they move away from their L1 and reach a more native-
like pronunciation. However, the path seems to be different for child L2 learners.
Children with the lowest time of exposure to Spanish appear to be mostly guided
by universal processes while children with intermediate time of exposure seem
to be more affected by the L1. This could explain why the group with intermedi-
ate time of exposure to Spanish performed worse than the group with the least
amount of Spanish overall. However, because children are hypothesized to also
utilize universal processes from the initial stage of acquisition, they are able to
overcome transfer errors more quickly than adults. This developmental scenario
is illustrated in Figure 1. At first child L2 learners are affected by universal factors
and that is why they produce more native-like sounds, but are later on affected by
transfer and start producing more non-native-like sounds. If they continue to be
exposed to the new language, this developmental scenario suggests that they will
be again influenced by universal strategies and will produce native-like sounds.
Native-Like

Non-Native-Like

140 h. 280 h. +500 h.


Figure 1. Phonological stages in the acquisition of a new language by children

The SLM model predicts that young learners will be more successful than adults at
producing L2 sounds because the former are better at discriminating the phonetic
differences between the L1 and L2 sounds. Therefore, it could be argued that chil-
dren seem to be better at acquiring the Spanish rhotics simply because they have
less difficulty than adults discriminating new sounds. As a result they need less
time than adults to master these sounds. However, adult L2 speakers appear not
to have problems perceiving and differentiating the Spanish rhotics (Rose, 2010b),
so perception is unlikely to explain apparent differences between child and adult
L2 learners. Rose conducted a perception study with English speaking adults with
previous experience with Spanish and with English speaking adults with no previ-
ous Spanish experience and found that none of the groups had greater difficulties
The acquisition of rhotics by child L2 and L3 learners 261

perceiving the Spanish rhotics. It is unlikely that our results are due to children’s
ability to better perceive the sounds in the L2 than adults.
Our results can also shed some light on whether the acquisition of the tap and
the trill are guided by similar strategies or the influence of universal strategies and
transfer are different for each sound, as has also been proposed for adult L2 ac-
quisition by Major (2001). According to Major, the production of the tap (similar
sound) should be heavily influenced by L1 transfer, and the production of the trill
(marked sound) should be affected by developmental errors. Our findings do not
support Major’s proposal. Both tap and trill sounds exhibited the same U-shaped
developmental pattern, as shown in Figure 1. Children were relatively success-
ful producing both sounds at the initial stage of acquisition, which suggests that
children seem to have access to universal strategies and are less influenced by L1
transfer at early stages of acquisition.
With respect to previous linguistic knowledge, we hypothesized that the child
L3 learners would be more successful when acquiring the phonology of the new
language because of their greater linguistic experience and language awareness.
The descriptive results supported our hypothesis but the statistical analysis found
no significant association. However, because our sample size was small, the failure
to find a significant effect may be related to the limited number of observations in
each group. Our results seem to support previous studies stating that children who
have already learned a foreign language may have more developed metalinguis-
tic awareness than monolingual children (Bruck & Genesee, 1995; Yelland et al.,
1993). The Korean-English bilingual participants were already familiar with the
process of learning a different language. Therefore, when confronted with the task
of learning an L3, they were already prepared with strategies developed during L2
acquisition. According to Melhorn (2007), that is why language acquisition by L3
learners should be faster and more effortless than by learners who are experienc-
ing non-native language acquisition for the first time.
Furthermore, our analysis by linguistic background suggested that child L2
phonological acquisition process is mainly governed by universal principles and
less by L1 transfer at initial stages of language acquisition. Although the Korean-
English bilingual children performed better than the English native speakers, both
groups were highly accurate. If the initial stage of child L2 acquisition was gov-
erned by L1 transfer, both Korean-English bilingual speakers and English native
speakers’ performance would have been less accurate. Moreover, we would expect
each group to be affected by different phonological errors due to L1 transfer. For
example, Korean adult speakers have difficulties with the production of the tap
(Flege et al., 1995a; Sheldon & Winifred, 1982). They use [l] instead of a rhotic
sound because in Korean [l] and [ɾ] are allophones. The fact that lateralization
was also found in the native speakers’ speech suggests that they are developmental
262 Alexandra Morales Reyes et al.

errors due to the process of acquiring the Spanish rhotics. At the same time, if dur-
ing this early stage of acquisition, the L1 played a more important role than univer-
sals, the English-Korean bilinguals would have lateralized the Spanish trill more
than the English monolingual children. Finally, the English-Korean bilinguals and
the English native speakers would have produced more alveolar approximants, a
variant similar to the English approximant. Therefore, the L1 did not seem to play
a prominent role at the earliest stages of child L2 and L3 acquisition.

6. Conclusion

We propose that child L2 learners of Spanish as an L2 and L3 follow similar ac-


quisition strategies and exhibit the same U-shape developmental pattern for the
production of Spanish rhotics. We believe that children, contrary to adults, are
guided by universal processes from the initial stage of acquisition. This advantage
helps them acquire new sounds before L1 transfer begins to affect their produc-
tion. Our results are encouraging, and could serve to motivate the acquisition of
an L2 in childhood and earlier development of multilingualism. In the U.S. there
are many children who are growing up bilingual at home and whose parents may
hesitate to expose them to an L3 due to a popular belief that learning a third lan-
guage might confuse the children and negatively affect children’s lexical and mor-
phosyntactic knowledge. Not only do our results show that bilingual children do
not get confused, but they show that children can reach milestones in the language
faster than monolingual speakers by acquiring native-like accent features. Our re-
sults also have important implications for language teaching, since they suggest
that monolingual children acquiring an L2 might need more time and effort than
bilingual children learning an L3. Teachers should be aware of this and understand
that not all the students will acquire a language at the same speed even if they are
all exposed to the same amount and type of input. Moreover, our results suggest
that teachers can be made aware of the different stages followed by children when
acquiring a new language. Students with intermediate exposure to the target lan-
guage might seem as if they are starting to lose interest or making less effort to
learn; however, when teachers and parents understand that they are going through
a normal stage in development, they can continue to provide the instruction that
the children need to support their learning. Finally, although having children with
different abilities may be challenging, it can also be an advantage: the bilingual
children with more native-like pronunciation are a model and a source of input to
children with less native-like pronunciation.
The acquisition of rhotics by child L2 and L3 learners 263

References

Abrahamsson, N. (2012). Age of onset and ultimate attainment of L2 phonetic and gram-
matical intuition. In N. Abrahamsson & K. Hyltenstam (Eds.), High-level L2 acquisition,
learning and use. Thematic issue of Studies in Second Language Acquisition (pp. 177–214).
Cambridge: Cambridge University Press.
Amastae, J. (1977). The use of phonology in the study of bilingualism. In B. Hoffer & B. Lou
Dubois (Eds.), Southwest linguistics then and now (pp. 229–281). San Antonio, TX: Trinity
University Press.
Anderson, R. (2004). Phonological acquisition in preschoolers learning a second language via
immersion: A longitudinal study. Clinical Linguistics & Phonetics, 18, 183–210.​
doi: 10.1080/0269920042000193571
Anderson, R., & Smith, B. (1987). Phonological development of two-year-old monolingual
Puerto Rican Spanish-speaking children. Journal of Child Language, 14, 57–78.​
doi: 10.1017/S0305000900012733
Asher, J. J., & Garcia, R. (1969). The optimal age to learn a foreign language. Modern Language
Journal, 53, 334–341. ​doi: 10.1111/j.1540-4781.1969.tb04603.x
Blecua, B. (2001). Las vibrantes del español: Manifestaciones acústicas y procesos fonéticos.
Unpublished Doctoral dissertation. Retrieved from: https://ddd.uab.cat/pub/tesis/2001/
tdx-0111102-110913/bbf1de3.pdfhttps://ddd.uab.cat/pub/tesis/2001/tdx-0111102-
110913/bbf1de3.pdf
Blom, E., Paradis, J., & Sorenson Duncan, T. (2012). Effects of input properties, vocabulary
size, and L1 on the development of third person singular –s in child L2 English. Language
Learning, 62, 965–994. ​doi: 10.1111/j.1467-9922.2012.00715.x
Boersma, P. (2012). Praat, a system for doing phonetics by computer. Glot International, 5(9/10),
341–345.
Bosch, L. (1983). El desarrollo fonológico infantil: Una prueba para su evaluación. Anuario de
Psicología, 28, 85–114.
Bruck, M., & Genesse, F. (1995). Phonological awareness in young second language learners.
Journal of Child Language, 22, 307–324. ​doi: 10.1017/S0305000900009806
Cabrelli-Amaro, J. (2012). L3 Phonology: An understudied domain. In J. Cabrelli Amaro,
S. Flynn, & J. Rothman (Eds.), Third language acquisition in adulthood (pp. 33–60).
Amsterdam: John Benjamins. ​doi: 10.1075/sibil.46.05ama
Carballo, G., & Mendoza, E. (2000). Acoustic characteristics of trill productions by groups of
Spanish children. Clinical Linguistics & Phonetics, 14, 587–601.​
doi: 10.1080/026992000750048125
Chen, X., Anderson, R. C., Li, W., Hao, M., Wu, X., & Shu, H. (2004). Phonological aware-
ness of bilingual and monolingual Chinese children. Journal of Educational Psychology, 96,
142–151. ​doi: 10.1037/0022-0663.96.1.142
De Angelis, G. (2007). Third or additional language acquisition. Clevedon, UK: Multilingual
Matters.
Face, T. L. (2006). Intervocalic rhotic pronounciation by adult learners of Spanish as a second
language. In C. A. Klee & T. L. Face (Eds.), Selected Proceedings of the 7th Conference on the
Acquisition of Spanish and Portuguese as First and Second Language (pp. 47–58). Somerville,
MA: Cascadilla Proceedings Project.
264 Alexandra Morales Reyes et al.

Ferguson, C., & Debose, C. (1977). Simplified registers, broken language, and pidginization.
In A. Valdman (Ed.), Pidgin and creole linguistics (pp. 99–125). Bloomington, IN: Indiana
University Press.
Flege, James E. (1995). Second language learning: Theory, findings, and problems. In W. Strange
(Ed.), Speech perception and linguistic experience: Issues in cross-language research (pp. 233–
277). Timonium, MD: York Press.
Flege, J. E., Takagi, N., & Mann, V. (1995a). Japanese adults can learn to produce English /ɹ/ and
/l/ accurately. Language and Speech, 38, 25–55. ​doi: 10.1177/002383099503800102
Flege, J. E., Munro, M., & MacKay, I. (1995b). Factors affecting strength of perceived foreign ac-
cent in a second language. Journal of the Acoustical Society of America, 97, 3125–3134.​
doi: 10.1121/1.413041
Georgia, A. (2007). Phonological awareness in bilingual and trilingual school children. The
Linguistics Journal, 3, 8–15.
Goldstein, B. (2004). Bilingual language development and disorders in Spanish–English speakers.
Baltimore, MD: Paul H. Brookes.
Gómez-Fernández, D. (2004). La adquisición de las líquidas en los niños de Sevilla y su pro-
vincia: Desde los doce a los treinta y seis meses. Revista de Filología y su Didáctica, 27,
125–185.
Granena, G., & Long, M. H. (2013). Age of onset, length of residence, language aptitude, and
ultimate L2 attainment in three linguistic domains. Second Language Research, 29, 311–343. ​
doi: 10.1177/0267658312461497
Hualde, J. I. (2005). The sounds of Spanish. New York, NY: Cambridge University Press.
Hecht, B. F., & Mulford, R. (1982). The acquisition of a second language phonology: Interaction
of transfer and developmental factors. Applied Linguistics, 3, 313–328.
Henriksen, N. (2015). Acoustic analysis of the rhotic contrast in Chicagoland Spanish. Linguistic
Approaches to Bilingualism, 5, 285–321. ​doi: 10.1075/lab.5.3.01hen
Ingram, J. C. L., & Park, S.-G. (1998). Language, context, and speaker effects in the identification
and discrimination of English /r/ and /l/ by Japanese and Korean listeners. The Journal of
the Acoustical Society of America, 103, 1161–1174. ​doi: 10.1121/1.421225
Kopečková, R. (2014). Crosslinguistic Influence in instructed L3 child phonology acquisition.
In M. Pawlak & L. Aronin (Eds.), Essential topics in applied linguistics and multilingualism,
second language learning and teaching. Studies in honor of David Singleton (pp. 205–224).
Dordrecht: Springer. ​doi: 10.1007/978-3-319-01414-2_12
Ladefoged, P. (2006). A course in phonetics. Rosenberg: Thomson Wadsworth
Ladefoged, P., & Maddieson, I. (1996) The sounds of the world’s languages. Oxford: Blackwell.
Lenneberg, E. (1967). Biological foundations of language. New York, NY: Wiley & Sons.
Lewis, A. M. (2004). Coarticulatory effects on Spanish trill production. In A. Agwuele, W.
Warren, & S.-H. Park (Eds.), Proceedings of the 2003 Texas Linguistics Society Conference
(pp. 116–127). Somerville, MA: Cascadilla Press.
Lipski, J. (1990). Spanish taps and trills: phonological structure of an isolated opposition. Folia
Linguistica, 24, 153–74. ​doi: 10.1515/flin.1990.24.3-4.153
Major, R. C. (1987). A model for interlanguage phonology. In G. Ioup & S. L. Weinberger (Eds.),
Interlanguage phonology: The acquisition of a second language sound system (pp. 101–124).
Rowley, MA: Newbury House.
Major, R. C. (2001). Foreign accent: The ontogeny and phylogeny of second language phonology.
New York, NY: Routledge.
The acquisition of rhotics by child L2 and L3 learners 265

Meisel, J. M. (2008). Child second language acquisition or successive first language acquisition?
In B. Haznedar & E. Gavruseva (Eds.), Current trends in child second language acquisition
(pp. 55–80). Amsterdam: John Benjamins. ​doi: 10.1075/lald.46.04mei
Melgar de González, M. (1976). Cómo detectar al niño con problemas del habla. México: Trillas.
Melhorn, G. (2007). From Russian to Polish: Positive transfer in third language acquisition. 16th
International Congress of Phonetic Sciences, Saarbrücken (pp. 1745–1748).
Moyer, A. (1999). Ultimate attainment in L2 phonology. Studies in Second Language Acquisition,
21, 81–108. ​doi: 10.1017/S0272263199001035
Moyer, A. (2004). Age, accent and experience in second language acquisition: An integrated ap-
proach to critical period inquiry. Clevedon, UK: Multilingual Matters.
Moyer, A. (2007). Do language attitudes determine accent? A study of bilinguals in the USA.
Journal of Multilingual and Multicultural Development, 28, 1–17. ​doi: 10.2167/jmmd514.0
Moyer, A. (2014). Exceptional outcomes in L2 phonology: The critical factors of learner engage-
ment and self –regulation. Applied Linguistics, 35, 418–440. ​doi: 10.1093/applin/amu012
Olsen M. K. (2012). The L2 acquisition of Spanish rhotics by L1 English speakers: The effect of
L1 articulatory routines and phonetic context for allophonic variation. Hispania, 95, 65–82.
Oyama, S. (1976). A sensitive period for the acquisition of a nonnative phonological system.
Journal of Psycholinguistic Research, 5, 261–283. ​doi: 10.1007/BF01067377
Paradis, J. (2011). Individual differences in child English second language acquisition:
Comparing child-internal and child-external factors. Linguistic Approaches to Bilingualism,
1, 213–237. ​doi: 10.1075/lab.1.3.01par
Piske, T., MacKay, I. R. A., & Flege, J. E. (2001). Factor affecting degree of foreign accent in an
L2: A review. Journal of Phonetics, 29, 191–215. ​doi: 10.1006/jpho.2001.0134
Reeder, J. T. (1998). English speakers’ acquisition of voiceless stops and trills in L2 Spanish.
Texas Papers in Foreign Language Education, 3, 101–117.
Rose, M. (2010a). Intervocalic tap and trill production in the acquisition of Spanish as a second
language. Studies in Hispanic and Lusophone Linguistics, 3, 379–419.​
doi: 10.1515/shll-2010-1080
Rose, M. (2010b). Differences in discriminating L2 consonants: A comparison of Spanish taps
and trills. In M. T. Prior et al. (Eds.), Selected Proceedings of the 2008 Second Language
Research Forum (pp. 181–196). Somerville, MA: Cascadilla Proceedings Project.
Serra, M. (1979). Normas estadísticas para la población escolar de tres a siete años en el área
metropolitana de Barcelona. Revista de Logopedia, Foniatría y Audiología, 3, 232–235.
Sheldon, A., & Winifred, S. (1982). The acquisition of /r/ and /l/ by Japanese learners of English:
Evidence that speech production can precede speech recognition. Applied Psycholinguistics,
3, 243–261. ​doi: 10.1017/S0142716400001417
Solé, M. (2002). Aerodynamic characteristics of trills and phonological patterning. Journal of
Phonetics, 30, 655–688. ​doi: 10.1006/jpho.2002.0179
Schwartz, B. D. (2003). Why child L2 acquisition? In J. van Kampen & S. Baauw (Eds.),
Proceedings of GALA 2003, 1 [LOT Occasional Series] (pp. 47–66). Utrecht University.
Tremblay, M. C. (2007). L2 influence on L3 pronunciation: Nativelike VOT in the L3 Japanese
of EnglishFrench bilinguals. Paper presented at the Satellite Workshop of ICPhS XVI,
Freiburg, Germany, 3–4 August 2007.
Unsworth, S. (2005). Child L2, adult L2, child L1: Differences and similarities. A study on the
acquisition of direct object scrambling in Dutch. Unpublished doctoral dissertation, Utrecht
University.
266 Alexandra Morales Reyes et al.

Unsworth, S. (2008). Comparing child L2 development with adult L2 development: How to


measure L2 proficiency. In E. Gavruseva & B. Haznedar (Eds.), Current trends in child sec-
ond language acquisition (pp. 301–336). Amsterdam: John Benjamins.​
doi: 10.1075/lald.46.15uns
Widdison, K. A. (1998). Phonetic motivation for variation in Spanish trills. Orbis: Bulletin
International de Documentation Linguistique, 40, 51–61. ​doi: 10.2143/ORB.40.1.505038
Winitz, H., Gillespie, B., & Starcev, J. (1995). The development of English speech patterns of a
7-year-old Polish-speaking child. Journal of Psycholinguistic Research, 24, 117–143.​
doi: 10.1007/BF02143959
Wode, H. (1976). Developmental sequences in naturalistic L2 acquisition. Working papers in
Bilingualism, 11, 1–31.
Wrembel, M. (2010). L2 accented speech in L3 production. International Journal of
Multilingualism, 7, 75–90. ​doi: 10.1080/14790710902972263
Wunder, E. (2010). Phonological cross-linguistic influence in third or additional language ac-
quisition. Proceedings of 6th New Sounds 2010, Poznań, 566–571.
Yelland, G. W., Pollard, J., & Mercuri, A. (1993). The metalinguistic benefits of limited contact
with a second language. Applied Psycholinguistics, 14, 23–444.​
doi: 10.1017/S0142716400010687
Yeni-Komshian, G., Flege, J., & Liu, S. (2000). Pronunciation proficiency in first and second
languages of Korean–English bilinguals. Bilingualism: Language and Cognition, 3, 131–149. ​
doi: 10.1017/S1366728900000225

Authors’ addresses
Alexandra Morales Reyes Begoña Arechabaleta-Regulez
Recinto Universitario de Mayaguez Span Ital & Port/Linguistics
Universidad de Puerto Rico 4120 For Lang Bldg
PO BOX 9000 707 S Mathews
Mayaguez M/C 176
00681-9000 Urbana, IL 61801
Puerto Rico United States

Silvina Montrul
Span Ital & Port/Linguistics
4120 For Lang Bldg
707 S Mathews
M/C 176
Urbana, IL 61801
United States

You might also like