International Journal of Bilingual Education and

Bilingualism

ISSN: (Print) (Online) Journal homepage: www.tandfonline.com/journals/rbeb20

Inflectional morphology: evidence for an

advantage of bilingualism in dyslexia

Maria Vender, Shenai Hu, Federica Mantione, Silvia Savazzi, Denis Delfitto &
Chiara Melloni

To cite this article: Maria Vender, Shenai Hu, Federica Mantione, Silvia Savazzi, Denis Delfitto
& Chiara Melloni (2021) Inflectional morphology: evidence for an advantage of bilingualism
in dyslexia, International Journal of Bilingual Education and Bilingualism, 24:2, 155-172, DOI:
10.1080/13670050.2018.1450355

To link to this article: https://doi.org/10.1080/13670050.2018.1450355

Published online: 15 Mar 2018.

Submit your article to this journal

Article views: 2390

View related articles

View Crossmark data

Citing articles: 5 View citing articles

Full Terms & Conditions of access and use can be found at

https://www.tandfonline.com/action/journalInformation?journalCode=rbeb20
INTERNATIONAL JOURNAL OF BILINGUAL EDUCATION AND BILINGUALISM
2021, VOL. 24, NO. 2, 155–172
https://doi.org/10.1080/13670050.2018.1450355

Inflectional morphology: evidence for an advantage of

bilingualism in dyslexia
Maria Vender a, Shenai Hub, Federica Mantionea,c, Silvia Savazzid, Denis Delfittoa and
Chiara Mellonia
a
Department of Cultures and Civilizations, University of Verona, Verona, Italy; bCollege of Foreign Languages and
Cultures, Xiamen University, Xiamen, People’s Republic of China; cDepartment of Psychology and Cognitive Science,
University of Trento, Trento, Italy; dDepartment of Neurosciences, Biomedicine and Movement Science, University of
Verona, Verona, Italy

ABSTRACT ARTICLE HISTORY

It has been shown that morphological skills are particularly enhanced in Received 23 October 2017
bilingual children, whereas they are compromised in dyslexics. The aim Accepted 28 February 2018
of this work is that of investigating how bilingualism interacts with
KEYWORDS
dyslexia in a task measuring the subject’s morphological abilities, to Bilingualism; developmental
verify if the advantage typically found in bilingualism arises also in dyslexia; inflectional
presence of a linguistic pathology such as dyslexia. We administered a morphology; benefits of
task assessing the ability to generate plural noun inflections of bilingualism; nonword
nonwords to 106 children: 24 Italian monolingual dyslexics (mean age pluralization
10;0 y.o.), 30 Italian monolingual typically developing children (10;1), 22
bilingual dyslexic children with Italian as L2 (10;4) and 30 bilingual
typically developing children with Italian as L2 (10;2). Results point to a
positive effect of bilingualism, which also extends to dyslexia, with
bilingual dyslexics performing consistently better than monolingual
dyslexics, approaching and even surpassing, as in the most difficult
conditions, the performance of monolingual unimpaired children.

1. Introduction
There is evidence that bilingualism can provide a positive boost to cognitive functioning, con-
veying important advantages in specific fields such as the domain of executive functions (EF),
where the use of two languages seems to favor the ability to focus on the relevant stimuli, inhi-
biting the irregular ones, and the sphere of metalinguistic awareness, where bilingualism
appears to enhance the subjects’ sensitivity to the structures of their languages (Bialystok and
Barac 2013; but see Paap in press for a critical assessment of previous studies). On the other
hand, it has not been investigated whether these advantages occur also in situations of atypical
bilingualism, where children suffer from linguistic or cognitive disorders. Trying to fill this gap,
our study aims at disentangling the relationship between bilingualism and developmental dys-
lexia, comparing the performance of monolingual and bilingual children, with and without a
diagnosis of dyslexia, in a task tapping the subjects’ metalinguistic awareness, and in particular
their morphological skills, in which bilinguals are generally reported to outperform monolin-
guals, whereas dyslexics tend to display a poorer performance in comparison to that of typically
developing children.

CONTACT Maria Vender [email protected]

© 2018 Informa UK Limited, trading as Taylor & Francis Group
156 M. VENDER ET AL.

1.1. Metalinguistic awareness and morphological skills in bilingualism

Metalinguistic awareness and morphological skills have been reported to be particularly enhanced in
bilingualism. As Vygotsky (1962) first underlined, bilinguals appear to be more familiar with the arbi-
trariness of the relationship between meaning and form in comparison to monolinguals, and there-
fore they are less reluctant to separate them. This consideration was confirmed by Bialystok (1986)
who found that bilingual children were more competent than monolinguals to accept grammatically
well-formed but anomalous sentences (e.g. Apples grow on noses), displaying therefore a better ability
to separate form and content, which was taken to attest their more sophisticated metalinguistic
awareness. This metalinguistic advantage has been explained by arguing that bilingualism seems
to guarantee a higher symbolic flexibility, allowing children to experience an accelerated separation
of meaning and form and to focus their attention on language form (Cummins 1978).
Another example of a metalinguistic task tapping morphological skills1 in which bilinguals have
been reported to show advantages in comparison to monolinguals is the Wug Test, which was
designed by Berko (1958) with the aim of assessing monolingual children’s ability to generate inflec-
tions in English by means of nonwords.
Barac and Bialystok (2012) administered the Wug Test and two tasks assessing language profi-
ciency (i.e. receptive vocabulary and grammatical ability) to a group of monolingual children and
three groups of age-matched bilinguals (Spanish-English, French-English and Chinese-English),
with the aim of comparing bilinguals and monolinguals, considering also other effects such as
language similarity (Spanish and French are more similar to English than Chinese) and language of
schooling (English for the vehicular language for Spanish L1 and Chinese L1 children, whereas
French was used for the French L1 children). The authors found that the Spanish-English children,
who did not differ in language proficiency from the monolinguals, outperformed them in the Wug
Test, showing higher morphological abilities. Conversely, the other two groups of bilinguals per-
formed worse than monolinguals in language proficiency, but their performance was better in the
Wug Test than in the tasks tapping their language competence more directly. The differences
between the groups of bilinguals were ascribed to the fact that only the Spanish-English bilinguals
could have benefitted from both the linguistic similarity between Spanish and English and the use
of English as language of instruction, whereas the other two groups could not take advantage of
at least one of these essential aspects.
Bialystok, Peets, and Moreno (2014) investigated whether a bilingual advantage arose also in
second language learners, by testing two groups of English children acquiring French in an immer-
sion education program and attending to the second and fifth grade. They found that all bilingual
children outperformed monolinguals in the Wug Test, whereas only the older bilingual children
were more accurate than the monolinguals in a sentence-judgment task. Results suggest that the
bilingual advantage emerges earlier and more clearly in the Wug Test, which can be considered
less complex than the other task, since there is no misleading information to be filtered out or any
need for particularly effortful processing (Bialystok et al. 2014).
Summarizing, the studies reviewed in this section suggest that bilinguals, having access to two
different linguistic systems, are prompted to develop a more sophisticated ability to focus and to
reflect on the structures of their languages in comparison to monolinguals.

1.2. Morphological skills in developmental dyslexia

Developmental Dyslexia is a genetic disorder characterized by a difficulty in properly acquiring
reading and spelling skills, despite adequate classroom exposure, in absence of cognitive, physical
or sensorimotor impairments and socio-economical or emotional problems (Vellutino 1979).
Although reading and spelling problems are the most evident manifestations of dyslexia, it has
been proposed that literacy problems constitute just the tip of the iceberg of the difficulties reported
in dyslexia, suggesting that this is a more complex disorder, characterized by linguistic and cognitive
 INTERNATIONAL JOURNAL OF BILINGUAL EDUCATION AND BILINGUALISM 157

deficits at different levels. An extensive body of research has determined that phonological deficits
are one of the most distinctive traits of dyslexia, persistent across different ages and languages
(Snowling 1995; Ramus and Szenkovits 2008). More recently, it has been demonstrated that other
domains of linguistic competence are compromised as well, such as lexical development (Scarbor-
ough 1990; Lyytinen et al. 2001; Snowling, Gallagher, and Frith 2003) and grammatical competence
(Byrne 1981; Bar-Shalom, Crain, and Shankweiler 1993; Waltzman and Cairns 2000; Robertson and
Joanisse 2010; Mantione 2016). Moreover, it has been suggested that the cognitive profile of dyslexics
is also characterized by poor working memory resources (Jeffries and Everatt 2004; McLoughlin,
Leather, and Stringer 2002; Vender 2017).
As for morphological competence, it has been shown that dyslexics display a different profile in
comparison to unimpaired children, performing worse than both chronological and reading age-
matched controls in tasks assessing the abilities to isolate and blend morphemes (Casalis, Colé,
and Sopo 2004). Deficits are also reported in tasks assessing gender and number agreement,
where dyslexics performed more poorly than both chronological and reading age-matched controls
(Jiménez et al. 2004; Rispens 2004). Moreover, as shown by Joanisse et al. (2000), dyslexics are less
skilled than controls in tasks tapping their mastery of inflectional morphology, and requiring the
application of past tense agreement and plural rules to both familiar words and nonwords.
This outcome has been confirmed by a more recent study conducted by Vender, Mantione,
Savazzi, Delfitto and Melloni (2017) with the aim of comparing the performance of dyslexics and
unimpaired children in a task assessing their inflectional morphology. Taking Berko’s original para-
digm as a basis, the authors adapted it to the more complex context of Italian inflection system, in
which plurals are typically obtained by modifying the phonological shape of the singular ending
in accordance with the declension class and gender feature of the stem. In their study, the
authors compared the performance of dyslexics and controls in five conditions, corresponding to
the different declension classes of Italian and characterized by distinct levels of complexity, as
briefly summarized in (1), with an existing noun for each of the relevant declension classes reported
in brackets.

(1). Conditions
(i) Condition 1: Feminine a > e, e.g. la muv-a > le muv-e (Class I, e.g. la port-a > le port-e, ‘door’)
(ii) Condition 2: Masculine o > i, e.g. il fol-o > i fol-i (Class II, e.g. il gall-o > i gall-i, ‘rooster’)
(iii) Condition 3: Masculine a > i, e.g. il tred-a > i tred-i (Class IV, e.g. il pirat-a > i pirat-i, ‘pirate’)
(iv) Condition 4: Masculine e > i, e.g. il dort-e > i dort-i (Class III masc., e.g. il pesc-e > i pesc-i, ‘fish’)
(v) Condition 5: Feminine e > i, e.g. la stab-e > le stab-i (Class III fem., e.g. la nav-e > le nav-i, ‘boat’).

Conditions 1 and 2, which correspond respectively to Class I and II of Italian nominal morphology,
are fully productive, predominant in the input and extremely salient from the acquisition perspective,
due to their high regularity and predictability (Dressler and Thornton 1996; among others). Moreover,
they show maximal discrepancy in gender and in phonological form (a > e – o > i). On the contrary,
Conditions 4 and 5 comprise items belonging to Class III, which is numerically consistent, but totally
unproductive and completely independent from gender specification. A noun ending in −e can be
either masculine or feminine and takes the corresponding plural ending in −i independently of
the stem’s gender. Therefore, differently from Class I and II, there is no systematic gender-based
rule or pattern for class determination. Finally, Condition 3, which corresponds to Class IV of Italian
morphology, is numerically less pervasive than Class III, but opaque as far as gender manifestation
is concerned. The majority of Italian nouns ending in −a are indeed feminine, whereas those in
this class are masculine: hence, the learner must pay attention to agreement patterns in the singular
to disentangle gender specification and devious phonological exponence, learning to produce the
target plural marker −i. Given the peculiarities of these declension classes of Italian, Conditions 3,
4 and 5 were expected to be more challenging than Conditions 1 and 2. Consistently, the authors
found that dyslexics performed significantly worse than controls, especially in Conditions 1, 2 and
158 M. VENDER ET AL.

3, suggesting that their morphological skills are less sophisticated than those of unimpaired children.
In Conditions 4 and 5, instead, the two groups of children showed a similar and particularly inaccurate
behavior, especially in the last condition, confirming thus that the relevant morphological rules were
more difficult to apply to nonwords.

2. The current study

As outlined above, bilingual children are reported to outperform monolinguals in metalinguistic and
morphological tasks, as in the Wug Test, whereas dyslexics tend to exhibit difficulties in this task. We
decided to administer the same task used by Vender et al. (2017) to both monolingual and bilingual
children, with and without a diagnosis of dyslexia, with a three-fold purpose: first, confirming or dis-
confirming the presence of an advantage of bilingualism in morphological skills; secondly, verifying if
morphological processing is compromised in dyslexia; and finally, exploring the relationship between
bilingualism and dyslexia in order to establish whether bilingualism confers an advantage also in
presence of dyslexia.

2.1. Methods
2.1.1. Participants
The experimental protocol was administered to 106 children divided in four groups: 24 Italian mono-
lingual dyslexic children (mean age 9;11 years old, SD = 1.3), 30 Italian monolingual typically devel-
oping children (mean age 10;1, SD = 1.0), 22 bilingual dyslexic children with Italian as L2 (mean
age 10;4, SD = 1.4) and 30 bilingual typically developing children with Italian as L2 (mean age 10;2,
SD = 1.2). A one-way ANOVA was carried out on these data, revealing that there were no significant
differences in the age of the subjects (F(3, 104) = .322, p = .809). A subset of the monolingual partici-
pants (all the monolingual dyslexic children and 28 out of 30 control children) were the same who
took part in Vender et al’s (2017) study.
Children with dyslexia were recruited from clinical speech centers or public schools in the area of
Trento and Verona; they were diagnosed with dyslexia on standard criteria (ICD-10, World Health
Organization, 2004) and they had no diagnosed or reported oral language problems or hearing dis-
order. All bilingual dyslexics (BD) had been diagnosed during or after their third year of school attend-
ance in Italy.
Control children were recruited in the same public schools as the dyslexics, and they had neither
diagnosed or referred cognitive deficits, nor language problems, hearing disorders or reading difficul-
ties. All children had normal or corrected to normal vision. The study was approved by the local ethics
committee and conducted in accordance with the standards specified in the 2013 Declaration of
Helsinki.
All bilingual children had Italian as their L2 and spoke another language at home. We have
decided to put no restrictions on the L1 spoken by the subjects, mainly due to the difficulty of
finding diagnosed bilinguals. The first languages spoken by the BD were Arabic (6 children), Albanian
(5), Romanian (4), Spanish (2), Indian (1), Turkish (1), Kosovan (1), Yoruba (1) and Senegalese (1). The
native languages of the bilingual controls (BC) were: Romanian (10 children), Arabic (8), Albanian (3),
Indian (2), Spanish (1), Ghanaian (1), Nigerian (1), Moldovan (1), Serbian (1), Polish (1) and
Macedonian (1).
We collected precise information about the exposure to the two languages of the bilinguals by
administering a questionnaire, the Bilingual Language Exposure Questionnaire, which had been
filled in by parents to analyze the current exposure to the two languages, the age of first exposure
(AFE), the cumulative and the traditional length of exposure (TLE) (see next section for more
details). As a cut-off, we included in our sample only children who had at least three consecutive
years of exposure to Italian, at least two years of school attendance in Italy and who still used
their mother language at home. Concerning this last aspect, we have found great heterogeneity,
 INTERNATIONAL JOURNAL OF BILINGUAL EDUCATION AND BILINGUALISM 159

with the majority of the children actually showing a quite balanced bilingualism and maintaining
their L1 with parents and siblings, whereas some of them tended to speak Italian also at home,
especially with their siblings.
To collect data concerning the socioeconomic status of the participants, we considered parental
education (PE) asking parents to provide information concerning their educational level. As for the
scoring system, one point was assigned for primary education (i.e. primary school and middle
school), two points for secondary education (i.e. high school) and three points for higher education
(i.e. university). The individual PE score of the children was calculated by using the average of the two
parents’ levels of education. A one-way ANOVA was carried out on these data revealing that there
were no significant differences in the PE score of the subjects (F(3, 104) = 1.499, p = .219).

2.1.2. Materials
2.1.2.1. Preliminary tasks
Nonverbal intelligence. To ensure that all subjects had a normal cognitive level, we administered the
Raven’s Colored Progressive Matrices test (Raven, Court, and Raven 1998). Results were calculated as
standard scores based on the Italian standardization (Belacchi et al. 2008).
Reading abilities. In order to take part in the study, children diagnosed with dyslexia had to score 2
SD below the mean for their age/class of education in speed or accuracy of word or pseudo-word
reading (as assessed by the test Prova di lettura di parole e di nonparole included in the Batteria
per la Valutazione della Dislessia e della Disortografia Evolutiva DDE-2 by Sartori, Job, and Tressoldi
2007). Conversely, controls had to score within the normal ranges in all reading tasks. To have a
measure of text reading, we administered the Prove di Lettura MT per la Scuola Elementare-2 (Cornoldi
and Colpo 1998), which are standardized for children attending the primary school, and the Nuove
prove di lettura MT per la scuola media inferiore (Cornoldi and Colpo 1995), which are standardized
for children attending the junior secondary school.
Receptive Vocabulary. To have a standardized measure for the subjects’ lexical abilities, we adminis-
tered the receptive vocabulary test Peabody – Test di Vocabolario Recettivo, the Italian version of The
Peabody Picture Vocabulary Test by Dunn and Dunn (2000).
Working Memory. We administered two tasks to assess the subjects’ working memory abilities: a
Forward Digit Span Task (FDS) and a Backward Digit Span Task (BDS), both taken from the
Working Memory Test Battery for Children (Pickering and Gathercole 2001) and adapted to Italian.
The FDS involved the spoken presentation of sequences of digits: the experimenter uttered a
sequence of digits of increasing length (starting from only one digit) and the child was asked to
repeat the digits exactly in the same order as they were presented. All digits were uttered in even
monotone at the rate of 1 per second. Each block was composed by 6 items; when the first four
trials of one block were correctly recalled, the fifth and the sixth trials were omitted and the child
was presented with trials of the subsequent block. Testing stopped when the child committed
three errors within the same block. As for the scoring system, we assigned 1 point for each item cor-
rectly recalled, including the omitted trials, whereas no points were given for wrong repetitions. In the
BDS, the experimenter uttered a sequence of digits of increasing length (from a minimum of 2 to a
maximum of 7) and the child had to recall the digits in the reverse order, starting from the last digit
heard and ending with the first. The administration of the task and the scoring system were the same
as in the FDS task.
Bilingual Language Exposure Questionnaire. Detailed information about bilingual children’s
exposure to their first language and to Italian was collected by administering a questionnaire that
we have developed on the basis of the Italian adaptation of the Utrecht Bilingual Language Exposure
Calculator (Unsworth et al. 2012; see Vender, Garraffa, Sorace and Guasti (2016) for a study deploying
the Italian version of the questionnaire). The questionnaire, filled in by parents, provides a detailed
description of children’s exposure to Italian. We collected information about the children’s AFE to
Italian, their current Quantity of Exposure (QE) to the L2, the TLE, which is calculated as the child’s
chronological age minus their AFE to Italian, and the Cumulative Length of Exposure (CLE), which
160 M. VENDER ET AL.

is a more precise measure that takes additional variables into consideration in order to determine the
actual exposure to the L2.

2.1.3. Pluralization task

Participants were administered the same test of inflectional morphology developed and adminis-
tered by Vender et al. (2017), with the aim of assessing each child’s ability to generate plural inflec-
tions of nonwords. In the first part of the task, we tested the subject’s capacity to apply proper plural
inflections to familiar words. In a typical example, the child was presented a target noun (e.g. il gallo
‘the rooster’) as part of a sentence, which was written and read aloud by the experimenter,
accompanied by a picture representing the relevant object. For instance, first, the experimenter
reads Questo è il gallo (‘This is the rooster’) and points to the picture of a rooster; then, she says
Adesso ce ne sono due (‘Now we have two of them’), and points to the picture of two identical roosters.
Finally, the experimenter reads Questi sono … (‘These are … ’) and pauses, inviting the child to com-
plete the sentence (e.g. target i galli ‘the roosters’).2
Testing proceeded similarly in the case of nonwords, where children were shown pictures of ficti-
tious creatures and told that the nonword was the name of that creature. We invented 15 nonwords,
equalized for phonological complexity to the words of the first part, namely 3 for each of the five
conditions and corresponding to four of the Italian declension classes, as exemplified in (2).

(2). Conditions
(i) Condition 1 (Class I): Feminine a > e (F a > e), e.g. la muv-a > le muv-e
(ii) Condition 2 (Class II): Masculine o > i (M o > i), e.g. il fol-o > i fol-i
(iii) Condition 3 (Class IV): Masculine a > i (M a > i), e.g. il tred-a > i tred-i
(iv) Condition 4 (Class III): Masculine e > i (M e > i), e.g. il dort-e > i dort-i
(v) Condition 5 (Class III): Feminine e > i (F e > i), e.g. la stab-e > le stab-i

As discussed above, these conditions correspond to different declension classes that can be distin-
guished according to their ‘type’ frequency and productivity in Italian: Conditions 1 and 2 are very
common and productive, whereas the other three are less common, especially Condition 3, and totally
unproductive. An additional complexity concerns the two gender-mixed conditions with the ending in
−e, namely Condition 4 and 5 which are ambiguous since they can be both masculine and feminine.
As for the scoring procedure, we attributed 1 point to each correct inflection including both the
determiner and the correct plural form (as in Vender et al. 2017). Especially in view of the phonolo-
gical deficits typically manifested by dyslexic subjects, we considered as correct also the items that
were wrongly repeated by the children but produced with the correct inflection (e.g. il treda > i
trebi), and the nonwords correctly inflected though with determiner omission (e.g. il treda > tredi).
Conversely, no points were assigned in case of a wrong nonword plural (e.g. il treda > i treda) and
of a wrong article (il treda > le trede).

3. Results
3.1. Preliminary measures
The preliminary measures comprised the Raven, the PPVT-R, the reading measures (accuracy and
speed in word reading, nonword reading and text reading), the FDS and the BDS. Means and stan-
dard deviations of the groups are reported in Table 1.
A series of one-way ANOVAs with Group as fixed factor were carried out to compare the perform-
ance of monolingual dyslexics (MD), BD, monolingual controls (MC) and BC.
Nonverbal intelligence. Results show that the four groups did not significantly differ for their non-
verbal cognitive level as measured by the Raven task (F (3, 102) = 1.731, p = .165).
 INTERNATIONAL JOURNAL OF BILINGUAL EDUCATION AND BILINGUALISM 161

Table 1. Mean score (standard deviation) for the preliminary measures.

MD BD MC BC
Raven 0.01 (0.75) 0.10 (0.56) 0.44 (0.75) 0.20 (0.83)
PPVT-R, raw score 98.04 (32.66) 86.45 (23.67) 101.66 (21.33) 95.73 (13.48)
Word reading – accuracy −2.20 (1.80) −2.64 (1.36) 0.47 (0.55) −0.01 (0.93)
Word reading – speed −3.74 (2.73) −1.97 (2.14) 0.38 (0.61) 0.27 (0.81)
Nonword reading – accuracy −2.13 (1.04) −2.15 (1.03) 0.48 (0.63) 0.13 (0.79)
Nonword reading – speed −2.87 (2.59) −0.81 (1.50) 0.36 (0.57) 0.63 (0.70)
Text reading – accuracy −0.42 (4.25) −1.98 (1.17) 0.59 (0.61) 0.29 (0.44)
Text reading – speed −1.84 (1.79) −1.20 (1.46) 0.29 (0.36) 0.14 (0.47)
Forward Digit Span 24.87 (3.68) 23.86 (3.46) 29.39 (4.90) 28.33 (4.23)
Backward Digit Span 9.50 (4.23) 10.14 (3.62) 13.50 (4.03) 13.87 (4.69)

Receptive vocabulary. No significant differences were found for what concerned the participants’
language proficiency, assessed by the PPVT-R (F (3, 102) = 1.906, p = .133). Since raw scores on the
test were age appropriate, according to the available standardized tests evaluating language profi-
ciency in Italian, we can conclude that the children participating in our study do not have any specific
vocabulary or language deficits. Moreover, given the absence of significant differences between
monolinguals and bilinguals, we can argue that our bilingual children’s competence in Italian was
good and, at least concerning their receptive vocabulary, comparable to that of their monolingual
peers.
Reading abilities. There were significant differences in the reading abilities of the four groups, with
dyslexics, both monolinguals and bilinguals, performing worse in comparison to controls in all
measures. As for word reading, group was significant in both accuracy (F (3, 102) = 43.275, p < .001)
and speed (F (3, 102) = 34.926, p < .001). Post-hoc comparisons with Bonferroni correction revealed
that both groups of dyslexics were slower and less accurate than controls (p < .001). Interestingly,
MD were also slower than BD (p < .01), even though accuracy was similar (p = 1.000), whereas no differ-
ences were found between the two control groups in either accuracy (p = .735) or speed (p = 1.000).
As for nonword reading, group was significant in both accuracy (F (3, 102) = 53.974, p < .001)
and speed (F (3, 102) = 29.834, p < .001). Post-hoc comparisons with Bonferroni correction
showed that both groups of dyslexics performed similarly (p = 1.000) and worse than the two
control groups (p < .001) for what concerned accuracy. MD were slower than the other three
groups of children (p < .001), whereas BD performed worse than BC (p < .01) and MC (p < .05),
but better than MD (p < .01).
Finally, Group was significant also in text reading, considering both accuracy (F (3, 102) = 7.088, p
< .001) and speed (F (3, 102) = 28.354, p < .001). Post-hoc comparisons with Bonferroni correction
revealed that MD and BD were equally (p = .329) slower than controls (p < .001), but MD performed
similarly to BD (p = .088), MC (p = .515) and BC (p = 1.000), whereas BD performed worse than MC (p
< .001) and than BC (p < .01), indicating that BD displayed in this case the poorest performance.
Working Memory. Significant group differences were found in the two working memory tasks, with
the Forward Digit Span (F (3, 102) = 10.413, p < .001) and the Backward Digit Span (F (3, 102) = 7.536 p
< .001). Post-hoc comparisons with Bonferroni correction revealed that in the FDS MD performed
worse than MC (p < .001) and BC (p < .05), and similarly to BD (p = 1.000), while BD performed worse
than the two control groups (p < .001). In the BDS, MD underperformed both control groups (p
< .01) and performed similarly to BD (p = 1.000), while BD performed worse than controls (p < .05).

Table 2. Mean scores (standard deviation) of the bilingual language exposure questionnaire.
BD (n.22) BC (n. 30)
AFE to Italian (years) 2.67 (2.34) 2.32 (1.80)
QE to Italian (percentage) 0.66 (0.14) 0.64 (0.13)
TLE to Italian 7.47 (2.06) 8.07 (2.15)
CLE to Italian 2.17 (0.75) 2.39 (0.77)
162 M. VENDER ET AL.

Bilingual Language Exposure Questionnaire. The results concerning the bilingual children’s AFE to
Italian, their current QE to the L2, the TLE, and the CLE are reported in Table 2. A series of indepen-
dent-sample t-tests were carried out showing that the two groups did not differ for AFE (t(48) = .598,
p = .552), QE (t(48) = .577, p = .567), TLE (t(48) = .991, p = .327) and CLE (t(48) = 1.009, p = .320).

3.2. Pluralization task

As illustrated above, our task comprised a first part assessing the subjects’ ability to apply the plural
inflection rules to existing words, and a second part testing their capacity to extend these rules to
invented words. Mean scores and standard deviations for the Wug Test are discussed separately, con-
sidering determiners on the one side and nouns on the other side.
Determiners. Mean scores for accuracy and standard deviations in the production of plural
articles with both words and nonwords are reported in Table 3.
Considering the accuracy of the four groups in the production of plural determiners, we observe
that performance was at ceiling with words, whereas some minor differences are reported with non-
words, where dyslexics, both monolinguals and bilinguals, performed worse than the two control
groups. When we look at the single conditions, we observe that M a > i was the most challenging,
with dyslexics, especially monolinguals, committing more errors than controls. Errors took always
the form of substitution for MD, MC and BC (100% of the cases) and for BD, who showed just one
article omission out of 30 errors (corresponding to the 0.03% of the cases). The typical error was
the production of a plural feminine determiner instead of a plural masculine determiner and vice
versa (e.g. the feminine le as the plural form of the masculine il, where the correct form would be
the masculine plural i).
The statistical analysis confirmed the presence of a significant difference among the four groups
with determiners in the second part of the task regarding nonwords (as in the first part, concerning
words, all subjects showed a 100% accurate performance).
We performed a series of three-way mixed analyses of variance (ANOVAs) with Condition (1, 2, 3, 4
and 5) as within-subject variable and Bilingualism (Bilingual, Monolingual) and Dyslexia (Dyslexic,
Control) as between-subject variables.
Concerning the production of the determiners of nonwords, we found that type of Condition was
significant (F (4, 99) = 4.009, p < .01, partial η2 = .139), whereas none of the interaction was significant:
the Condition × Bilingualism interaction (F (4, 99) = .760, p = .554, partial η2 = .030), the Condition ×
Dyslexia interaction F (4, 99) = 1.537, p = .197, partial η2 = .058), and the Condition × Bilingualism ×
Dyslexia interaction (F (4, 99) = .656, p = .624, partial η2 = .026). To understand this result in detail,
we ran a series of two-way ANOVAs with Bilingualism (Bilingual, Monolingual) and Dyslexia (Dyslexic,
Control) as fixed factors and each condition as a dependent variable. We found that there was a sig-
nificant effect of Dyslexia in F a > e (F (1, 102) = 4.719, p < .05, partial η 2 = .044), M o > i (F (1, 102) =

Table 3. Mean (standard deviation) accuracy for the Wug Test – determiners.
MD BD MC BC
Words (total) 1.00 (0.00) 1.00 (0.00) 1.00 (0.00) 1.00 (0.00)
Feminine a > e 1.00 (0.00) 1.00 (0.00) 1.00 (0.00) 1.00 (0.00)
Masculine o > i 1.00 (0.00) 1.00 (0.00) 1.00 (0.00) 1.00 (0.00)
Masculine a > i 1.00 (0.00) 1.00 (0.00) 1.00 (0.00) 1.00 (0.00)
Masculine e > i 1.00 (0.00) 1.00 (0.00) 1.00 (0.00) 1.00 (0.00)
Feminine e > i 1.00 (0.00) 1.00 (0.00) 1.00 (0.00) 1.00 (0.00)
Nonwords (total) 0.93 (0.10) 0.95 (0.10) 0.99 (0.02) 0.97 (0.05)
Feminine a > e 0.94 (0.16) 0.98 (0.07) 1.00 (0.00) 1.00 (0.00)
Masculine o > i 0.94 (0.16) 0.94 (0.13) 1.00 (0.00) 0.98 (0.08)
Masculine a > i 0.85 (0.22) 0.91 (0.18) 1.00 (0.00) 0.94 (0.15)
Masculine e > i 0.93 (0.17) 0.94 (0.22) 0.99 (0.06) 0.97 (0.13)
Feminine e > i 0.99 (0.07) 0.98 (0.07) 0.99 (0.06) 0.99 (0.06)
 INTERNATIONAL JOURNAL OF BILINGUAL EDUCATION AND BILINGUALISM 163

5.021, p < .05, partial η 2 = .047) and M a > i (F (1, 102) = 9.323, p < .01, partial η 2 = .084), with no effect
of language or significant interactions. Conversely, no differences amongst the four groups were
found in the last two conditions with the vowel ending in −e, M e > i (p = .152) and F e > i (p = .789).
Nouns. Mean scores for accuracy and standard deviations in the production of plural nouns with both
words and nonwords are reported in Table 4.
Concerning the plural forms of nouns, we can notice that performance is again at ceiling with
words, confirming that participants were able to produce the target plural form in Italian, as
shown in Table 4. Specifically, a complete accuracy is reported in all conditions, except for M a > i
were all groups showed a slightly lower performance. Conversely, all groups manifested more diffi-
culties with nonwords, with MD showing the highest error rate and performing even more poorly
than BD, who were instead as accurate as unimpaired children. Analyzing the results of each
single condition, we can observe that all groups displayed a similar performance in F a > e,
whereas MD were less accurate than the other groups in the other conditions. BD displayed
overall a surprisingly good performance, being more accurate than MD in each condition, and man-
ifesting more difficulties than the two control groups only in M a > i, and even outperforming MC in
the two conditions with the vowel ending in −e. Finally, both groups of bilingual children outper-
formed the two monolingual groups in M e > i and F e > i.
We performed a series of three-way mixed analyses of variance (ANOVAs) with Condition (1, 2, 3, 4
and 5) as within-subject variable and Bilingualism (Bilingual, Monolingual) and Dyslexia (Dyslexic,
Control) as between-subject variables.
Concerning words, we found that type of Condition was significant (F (4, 99) = 28.391, p < .001,
partial η2 = .218), whereas none of the interaction was significant: the Condition × Bilingualism inter-
action (F (4, 99) = .300, p = .878, partial η2 = .003), the Condition × Dyslexia interaction F (4, 99) =
2.402, p = .124, partial η2 = .023), and the Condition × Bilingualism × Dyslexia interaction (F (4, 99)
= .860, p = .356, partial η2 = .008). This indicates that the condition M a > i, which is the only condition
in the real-word task whereby children commit some errors, is more difficult than the other four con-
ditions, where all groups of children show a perfect accuracy. The lack of interactions demonstrates
that this holds for all participants, without distinctions among groups.
As for nonwords, not only the type of Condition was significant (F (4, 99) = 65.651, p < .001, partial
η2 = .726), but significance was found in the Condition × Bilingualism interaction (F (4, 99) = 8.392, p
< .001, partial η2 = .253), in the Condition × Dyslexia interaction F (4, 99) = 16.060, p < .001, partial η2
= .394), and in the Condition × Bilingualism × Dyslexia interaction (F (4, 99) = 11.903, p < .001, partial
η2 = .325).
This suggests that the different types of conditions were processed differently by bilingual chil-
dren with respect to monolingual children and by dyslexic children with respect to unimpaired chil-
dren. To appreciate these results in more detail, separate analyses were conducted for each sentence
type, performing a two-way ANOVA with Bilingualism and Dyslexia as fixed factors and with each of

Table 4. Mean (standard deviation) accuracy for the Wug Test – nouns.
MD BD MC BC
Words (total) 0.98 (0.05) 0.98 (0.04) 0.97 (0.05) 0.96 (0.07)
Feminine a > e 1.00 (0.00) 1.00 (0.00) 1.00 (0.00) 1.00 (0.00)
Masculine o > i 1.00 (0.00) 1.00 (0.00) 1.00 (0.00) 1.00 (0.00)
Masculine a > i 0.89 (0.25) 0.90 (0.18) 0.86 (0.26) 0.78 (0.34)
Masculine e > i 1.00 (0.00) 1.00 (0.00) 1.00 (0.00) 1.00 (0.00)
Feminine e > i 1.00 (0.00) 1.00 (0.00) 1.00 (0.00) 1.00 (0.00)
Nonwords (total) 0.55 (0.15) 0.72 (0.17) 0.71 (0.16) 0.75 (0.19)
Feminine a > e 0.82 (0.26) 0.85 (0.27) 0.94 (0.12) 0.87 (0.21)
Masculine o > i 0.85 (0.22) 0.94 (0.13) 1.00 (0.00) 0.94 (0.15)
Masculine a > i 0.33 (0.31) 0.56 (0.36) 0.59 (0.40) 0.62 (0.38)
Masculine e > i 0.65 (0.35) 0.88 (0.26) 0.80 (0.28) 0.82 (0.30)
Feminine e > i 0.12 (0.27) 0.36 (0.46) 0.24 (0.36) 0.51 (0.40)
164 M. VENDER ET AL.

the five conditions as dependent variable. No differences were reported in F a > e: no effect of Dys-
lexia (F (1, 102) = 2.884, p = .093, partial η 2 = .027), no effect of Bilingualism (F (1, 102) = .324, p = .570,
partial η 2 = .003) and no Bilingualism × Dyslexia interaction (F (1, 102) = 1.605, p = .208, partial η 2
= .015). In M o > i we found a significant effect of Dyslexia (F (1, 102) = 7.597, p < .01, partial η 2
= .069), no effect of Bilingualism (F (1, 102) = .418, p = .520, partial η 2 = .004) and a significant Bilingu-
alism × Dyslexia interaction (F (1, 102) = 6.719, p < .05, partial η 2 = .062), indicating that MD performed
worse than BD.
Considering M a > i, we found only a significant effect of Dyslexia (F (1, 102) = 4.804, p < .05, partial
η 2 = .045), no effect of Bilingualism (F (1, 102) = 3.264, p = .074, partial η 2 = .031) and no interaction (F
(1, 102) = 1.820, p = .180, partial η 2 = .018), confirming that dyslexics, both monolinguals and bilin-
guals, performed worse than unimpaired children.
Quite different results were shown in M e > i, where we found only a significant effect of Bilingu-
alism (F (1, 102) = 4.434, p < .05, partial η 2 = .042), but no effect of Dyslexia (F (1, 102) = .591, p = .444,
partial η 2 = .006) and no interaction (F (1, 102) = 2.981, p = .087, partial η 2 = .028). This indicates that
bilinguals, both dyslexics and controls, performed better than monolinguals. Similarly, in F e > i,
which was the most difficult condition for all groups, there was a significant effect of Bilingualism
(F (1, 102) = 11.594, p < .001, partial η 2 = .102), no effect of Dyslexia (F (1, 102) = 3.259, p = .074,
partial η 2 = .031) and no interaction (F (1, 102) = .036, p = .850, partial η 2 = .000), confirming again
that bilinguals outperformed monolinguals, irrespective of the presence of dyslexia.
Successively, we analyzed the typology of errors committed by the four groups in the inflection of
invented nouns. Errors were classified in three different categories: (i) ‘invariant’, when the plural form
of the nonwords was identical to the singular (e.g. il bolfa > i bolfa, instead of the target form i bolfi),
(ii) ‘gender error’, when the gender of the noun was changed from masculine to feminine or vice
versa, with the production a plural inflection of the noun in agreement with the determiner (e.g. il
bolfa > le bolfe) and (iii) ‘unpredictable’, when the plural form of the noun could not be captured
by any existing morphological rule (e.g. il bolfa > i bolfe). The distribution of the errors made by
the four groups in these three categories is shown in Table 5.
Data reveal that the production of an invariant form is the most common error for all four groups
and in all conditions. Specifically, this kind of error represents the 88% of the errors committed by BD,
the 91% by MD, the 93% by BC and the 98% by MC.

Table 5. Classification of the errors in the inflection of invented nouns in each condition.
MD BD MC BC
F a > e target 0.82 (0.26) 0.85 (0.27) 0.94 (0.12) 0.87 (0.21)
Invariant 0.12(0.20) 0.07 (0.17) 0.05 (0.12) 0.12 (0.20)
Wrong gender 0.03 (0.09) 0.01 (0.07) 0.00 (0.00) 0.00 (0.00)
Unpredictable 0.03 (0.09) 0.06 (0.22) 0.01 (0.06) 0.01 (0.06)
M o > i target 0.85 (0.22) 0.94 (0.13) 1.00 (0.00) 0.94 (0.15)
Invariant 0.06 (0.13) 0.03 (0.09) 0.00 (0.00) 0.03 (0.13)
Gender 0.05 (0.16) 0.01 (0.07) 0.00 (0.00) 0.01 (0.06)
Unpredictable 0.04 (0.11) 0.02 (0.07) 0.00 (0.00) 0.02 (0.06)
M a > i target 0.33 (0.31) 0.56 (0.36) 0.59 (0.40) 0.62 (0.38)
Invariant 0.64 (0.33) 0.38 (0.35) 0.40 (0.40) 0.35 (0.38)
Wrong gender 0.03 (0.09) 0.04 (0.11) 0.00 (0.00) 0.03 (0.13)
Unpredictable 0.00 (0.00) 0.02 (0.07) 0.01 (0.08) 0.00 (0.00)
M e > i target 0.65 (0.35) 0.88 (0.26) 0.80 (0.28) 0.82 (0.30)
Invariant 0.33 (0.37) 0.12 (0.26) 0.20 (0.28) 0.17 (0.30)
Wrong gender 0.00 (0.00) 0.00 (0.00) 0.00 (0.00) 0.01 (0.06)
Unpredictable 0.02 (0.07) 0.00 (0.00) 0.00 (0.00) 0.00 (0.00)
F e > i target 0.12 (0.27) 0.36 (0.46) 0.24 (0.36) 0.51 (0.40)
Invariant 0.86 (0.29) 0.61 (0.46) 0.75 (0.36) 0.48 (0.39)
Wrong gender 0.00 (0.00) 0.03 (0.14) 0.02 (0.06) 0.00 (0.00)
Unpredictable 0.02 (0.07) 0.00 (0.00) 0.00 (0.00) 0.01 (0.06)
 INTERNATIONAL JOURNAL OF BILINGUAL EDUCATION AND BILINGUALISM 165

3.3. Correlations with reading skills, working memory and language exposure
Carrying out a Pearson correlation analysis, we found positive correlations between performance with
articles and nouns in the nonword part and some of the reading measures. The production of the
correct plural determiner was correlated to speed in word reading (r = .199, p < .05) and nonword
reading (r = .196, p < .05), but just marginally in text reading (r = .186, p = .056), as well as to accuracy
in word reading (r = .337, p < .001), nonword reading (r = .354, p < .001) and text reading (r = .229, p
< .05). The production of the correct plural noun correlated with speed in word reading (r = .207, p
< .05), nonword reading (r = .233, p < .05), but not in text reading (r = .144, p = .140), and with accuracy
in word reading (r = .239, p < .05), nonword reading (r = .212, p < .05), but not in text reading (r = .027,
p = .781).
The results of this correlational analysis show that there is a relationship between performance in
the morphological task and reading abilities, both with articles and nouns, thus confirming morpho-
logical skills are related to reading.
To check whether performance was related to children’s working memory, we carried out a
second correlational analysis. Results confirmed that the subjects who displayed a better perform-
ance in the production of plural determiners in nonwords had higher phonological memory abilities,
as measured by the FDS task (R = .213, p = < .05) and Central Executive abilities, as measured by the
BDS task (R = .219, p < .05). Similarly, performance in the production of the correct plural nouns in
nonwords was related to the FDS task (R = .219, p < .05) but not to the BDS task (R = .168, p = .085),
suggesting that the Phonological Loop plays a more prominent role in comparison to the Central
Executive in this kind of task.
Finally, we analyzed the presence of correlations between performance in the nonword task and
the data collected through the Bilingual Language Exposure Questionnaire. Results showed only a
moderate correlation between performance with invented nouns and AFE to Italian (R = .306, p
< .05). Conversely, accuracy in the task was correlated neither to the amount of exposure to Italian
(R = .253, p = .076), nor to the cumulative (R = 137, p = 342) or the TLE (R = .225, p = 116). Similarly, per-
formance with articles was correlated neither to AFE (R = .132, p = .360), amount of exposure (R =
−237, p = .098), TLE (R = −.042, p = .772) nor with CLE (R = −.132, p = .361). Moreover, performance
with articles and nouns in the nonword part was not correlated to nonverbal intelligence, as
measured by the Raven task (respectively: R = .142, p = .147 for articles and R = .110, p = .260 for
nouns) and to receptive vocabulary, as measured by the PPVT-R (respectively: R = −0.54, p = .582
for articles and R = −.097, p = .324 for nouns).

4. Discussion
In this study, we administered a protocol based on Berko’s Wug Test, with the aim of analyzing how
bilingualism and dyslexia interact in a task assessing the subjects’ morphological skills where bilin-
guals are generally reported to perform better than monolinguals and dyslexics are typically
shown to display a weaker performance in comparison to that of normally developing children. In
terms of overall performance, all groups tended to display a ceiling performance with words, consid-
ering both determiners and nouns, while displaying marked difficulties with nonwords.
Regarding determiners, we found a clear effect of dyslexia and no effect of bilingualism. Results
indicate that both groups of dyslexics underperformed in comparison to controls in the production
of plural articles of nonwords, consistently with the results reported by Vender et al. (2017). This
suggests that impaired children have more difficulties than controls in generating the plural inflec-
tion of articles with nonwords, whereas they do not display any difficulties with existing words.
This fact could reflect a lower ability to focus on the given form in comparison to controls, which
is mitigated in the case of real words by the familiarity with the stimuli and children’s implicit knowl-
edge of the noun gender and declension class. Looking at the single conditions, we found that dys-
lexics performed differently from controls in the two easiest conditions, F a > e and M o > i, and in
166 M. VENDER ET AL.

condition M a > i. In the latter condition, impaired subjects had a significantly higher tendency in
comparison to controls to shift the gender of the nonwords from the masculine to the feminine, pro-
ducing the feminine determiner le (instead of i) as the plural form of the masculine il. These results
extend those reported in Vender et al. (2017), where MD performed differently from controls only in
condition M a > i. Moreover, both studies revealed that dyslexics performed similarly to controls in
conditions M e > i and F e > i, which correspond to declension Class III. A possible explanation for
this difference could lie in the gender ambiguity that characterizes nouns belonging to Class III,
which can be either feminine or masculine. We thereafter propose that dyslexics, being aware of
this peculiarity, appear to pay more attention to the gender of the relevant stimuli (which can be
extracted by the gender-transparent determiner: ‘lamose’ -> feminine), in comparison to the other
conditions. Finally, no effect of bilingualism is found in the production of articles of nonwords, indi-
cating that bilinguals attained the same standard as monolinguals in this respect.
For what concerns the plural inflection of invented nouns, we found that bilinguals performed
better than monolinguals, irrespective of dyslexia, thus pointing to a general advantage of bilingual-
ism which also extends to dyslexia. Indeed, the BD displayed higher accuracy than the MD, perform-
ing similarly to the two control groups.
Analyzing each condition, no differences were reported in F a > e, where all groups manifested a
similarly accurate performance, whereas differences arose in M o > i, with dyslexics performing worse
than controls. In this case, the presence of a significant interaction between bilingualism and dyslexia
reveals that the effect of dyslexia, which has a negative impact on performance, is limited to mono-
lingual children, indicating thus a positive effect of bilingualism in dyslexia. In M a > i, instead, we
found only a significant effect of dyslexia, which suggests that dyslexics, irrespective of bilingualism,
performed more poorly than controls. This result confirms the previous one reported by Vender et al.
(2017) and is arguably due to the lower ‘type’ frequency of nouns belonging to Class IV in Italian and
to its opacity for what concerns gender; this is also supported by the fact that M a > i is the only con-
dition in which all groups of children did not reach a 100% accuracy with words. In condition M e > i
and F e > i, instead, results reveal a marked positive effect of bilingualism, without any effect of dys-
lexia. As in the previous study, MD performed more poorly than MC, but this difference is not stat-
istically significant, whereas both groups of bilinguals outperformed monolinguals, suggesting that
bilingualism enhances morphological skills, increasing the sensitivity to the relevant rules and the
ability to apply them in online processing.
Looking at the error patterns, the most frequent error made by all four groups was that of produ-
cing an invariant form, which retains the same form of the singular also for the plural. Two factors
might explain this error pattern. Firstly, production of non-target singular forms instead of target
plurals is a common pattern in first language acquisition, also from a cross-linguistic perspective
(see for instance Marinis 2003 on Greek). Secondly, invariable nouns are an open and productive
declension class in Italian, where we find most loanwords like bar, sport, test, etc. or clipped forms
like moto < moto-cicletta ‘motorbike’, foto < foto-grafia ‘photo’ (see D’Achille and Thornton 2003,
2008 on some other Italian nouns falling in this class). Therefore, this performance might result
from children’s erroneous interpretation of nonwords as loanwords falling in the class of invariables.
However, although the tendency to produce invariant plurals is the most common error for all groups
of subjects, this strategy seems to be adopted more frequently by monolinguals with respect to bilin-
guals, who appear to stick more closely to the morphological rules of Italian, applying them more
consistently than their monolingual peers, especially in the most difficult conditions. This suggests
that bilinguals, both dyslexics and controls, show more sophisticated morphological abilities,
which leads them to recognize the morphological patterns of the invented words more efficiently
than monolinguals.
The correlation analysis confirmed that there is a relationship between the performance in
nonword pluralization of both articles and nouns and reading measures, concerning both accuracy
and speed. We propose that the importance of morphological skills in reading are due to the crucial
role that they play in the accomplishment of the orthographic stage in Frith’s (1986) model of
 INTERNATIONAL JOURNAL OF BILINGUAL EDUCATION AND BILINGUALISM 167

reading acquisition. The ability to break down the word into morphemes, which is effectively mas-
tered at this stage, has a determinant role in enabling instant word recognition and in enhancing
reading proficiency. Moreover, the significant correlation between general performance and
working memory (WM) abilities, considering both the FDS and the BDS, suggests that the subject’s
working memory abilities are able to account, at least in part, for their linguistic behavior in the
pluralization task.3 Conversely, general performance in the task is not related to nonverbal intelli-
gence nor to vocabulary, indicating that these factors do not impact on performance. Finally, we
found no correlations between performance and the amount of present and past exposure to
Italian of bilingual children: this result seems to confirm the one reported by Bialystok, Peets,
and Moreno (2014), indicating that the advantage displayed by bilingual children in tasks
tapping their morphological skills, such as the Wug Test, shows up very early. Although there
must be a critical minimal amount of exposure needed to produce enhanced performance, the
advantages provided by bilingualism are arguably related to the fact of being exposed to two lin-
guistic systems, and not to the amount of exposure to the second language. As for the AFE, the
results of the correlational analysis suggest that children exposed later to Italian are even more pro-
ficient in the task than children exposed earlier. This is in line with other studies (Bohman et al.
2010; Paradis 2011) suggesting that children exposed later to their L2 in a preschool or school
program tend to acquire it faster and more efficiently in comparison to children acquiring it
earlier at home (Paradis and Kirova 2014).
Summarizing, the results of the present study lead to three important findings. The first important
result is that MD children display the worst performance, committing more errors than the two
control groups, also crucially more errors than BD children. This allows us to answer our first research
question, confirming that (monolingual) dyslexics are not able to apply inflectional rules to novel
nouns as efficiently as unimpaired children, thus indicating that morphological skills are impaired
in dyslexia.
The second important result concerns the advantage that bilingualism confers in this kind of task:
we found indeed that bilingual children, both dyslexics and controls, performed significantly better
than monolinguals in the pluralization of nonwords with the vowel ending in −e. This result provides
an answer to our second research question, confirming that there is an advantage for bilingual chil-
dren over monolinguals, even though this advantage is more evident in the most difficult conditions
with the ending in −e corresponding to conditions 4 and 5 of the task.
Finally, the perhaps most interesting result concerns the interaction between dyslexia and bilin-
gualism: data show that bilingual dyslexics outperformed monolingual dyslexics in all conditions,
except for condition F a > e, where all groups had a similar performance, and condition M a > i,
where all groups had a poorer performance and only the (negative) effect of dyslexia was observed.
As observed above, the latter result is possibly related to the very low frequency in the input of this
declension class in Italian, as well as to its unproductivity and misleading phonological shape with
respect to gender.
All in all, our results point to a general positive effect of bilingualism in the production of plural
noun inflections of nonwords which also extends, crucially, to dyslexia, suggesting that bilingual dys-
lexics are endowed with better morphological abilities than monolingual dyslexics, approaching and
even outweighing, as in the most difficult conditions, the performance of monolingual unimpaired
children.
Along the lines of the studies reviewed in the introduction, our results indicate that bilingualism
might have a positive effect on morphological and metalinguistic abilities, both in impaired and
unimpaired children. Specifically, bilingual success in this kind of task might be determined by an
advantage in selective attention, which helps them to focus on relevant information (in our task,
the rule to apply, e.g. producing the ending in –i to pluralize feminine nouns belonging to Class
III) and to inhibit misleading information (i.e. the tendency to produce the ending in –e for feminine
nouns, irrespective of the declension class to which they belong). These considerations open the way
to a more general discussion concerning the nature of the advantages of bilingualism and the
168 M. VENDER ET AL.

question regarding how and why they show up. Although the advantages in morphological and
metalinguistic awareness reported in the literature and discussed here are essentially linguistic, it
should be reminded that bilingualism has been shown to have consequences also on nonverbal
tasks and, more generally, on the cognitive functioning of bilinguals. Specifically, the fact that a bilin-
gual has to constantly deal with two languages in her mind does not affect only her language pro-
cessing abilities but, crucially, also her cognitive and EF. Indeed, both languages are always jointly
activated, thus creating a competition that the bilingual has to solve by inhibiting the interference
of the language that she is not using in a specific moment. Therefore, bilinguals have been proposed
to develop a stronger ability than monolinguals in inhibiting the access to the nonrelevant language
and in focusing selective attention to the relevant one (Bialystok et al. 2009). Since inhibition and
selective attention are part of the EF system (Miyake et al. 2000), it has been proposed that bilingu-
alism affects also the domain of EF, a claim that has been supported by a number of studies reporting
enhanced EF in bilingualism (Bialystok, Craik, and Luk 2008; Costa, Hernández, and Sebastián-Gallés
2008; amongst others). However, further studies are needed to explore more in detail this proposal,
which has been criticized by other researches evidencing on the contrary a lack of bilingual advan-
tage in EF tasks (Hilchey, Saint-Aubin, and Klein 2015; Paap, Johnson, and Sawi 2015).
As a final point, a possible limit of the present study lies in the choice to recruit bilingual children
with different mother-languages, which is motivated by the great difficulty to find a sufficiently
representative sample of bilingual subjects with a diagnosis of dyslexia speaking the same L1.
Namely, one could object that the L1 spoken by the subject could have exerted an influence on
the performance in our task. However, it is worth noting that the advantage shown in the
present study by bilinguals as a group demonstrates that bilingualism enhances the morphological
abilities of the children independently from the L1. On the other hand, we would certainly endorse,
for the future, a line of research that compares monolinguals with two distinct groups of bilinguals,
whose L1 is respectively similar and different from Italian with respect to their inflectional system, in
order to be able to ascertain the presence of a positive or negative transfer. This has been already
partially verified by a study in which the pluralization task that we used in this research was admi-
nistered to two groups of typically developing bilinguals speaking Albanian or Romanian as their L1:
findings revealed that both groups of bilinguals performed better than the monolinguals in con-
dition F e > i, suggesting therefore that the bilingual advantage was not related to the L1
spoken (Melloni, Vender and Delfitto, to appear).
Nevertheless, future researches could also address morphological competence more completely,
considering not only the pluralization of nouns, as in the present study, but also other types of inflec-
tional and derivational phenomena.
Finally, in order to deal with a further limitation of this research lying in the small size of the exper-
imental samples, future research could consider larger groups of participants, in order to decrease the
risk of both Type 1 and Type 2 errors (Bakker, van Dijk, and Wicherts 2012).

5. Conclusion
This study aimed at comparing monolingual and bilingual children, with and without a diagnosis of
dyslexia, in a task assessing their morphological skills. As discussed above, results point to an advan-
tage of bilingualism that crucially extends to dyslexia: in the most difficult conditions, indeed, bilin-
gual dyslexics performed even better than monolingual unimpaired children, indicating that their
ability to apply morphological rules in an abstract way across the board, based on the input provided
to them by their second language, and more particularly the ability to apply them to novel words, was
significantly better not only with respect to monolingual dyslexics, but also with respect to monolin-
gual typically developing children.
These results indicate that bilingualism does not have a negative impact on dyslexic children, at
least for what concerns their morphological competence. This constitutes an important indication for
all the educators, speech therapists and teachers who think that bilingualism may have a negative
 INTERNATIONAL JOURNAL OF BILINGUAL EDUCATION AND BILINGUALISM 169

influence on dyslexia and tend to provide families of dyslexic children with a negative advice with
respect to the use of their home language, in order to avoid alleged negative consequences of bilin-
gualism on dyslexic children. Our results suggest that this is generally not the case: on the contrary, it
seems that in some cases bilingualism can be a positive, protective factor on dyslexia, especially in
enhancing the subjects’ morphological abilities. The conclusion emerging from these preliminary
results is thus that bilingualism can be seen as an opportunity to catch, even in the case of commu-
nicative impairment and developmental language pathologies.

Notes
1. According to Bialystok, Peets, and Moreno (2014, 182), the Wug Test developed by Berko taps morphological
awareness since it ‘assesses children’s sensitivity to morphological structure in English’. However, whereas mor-
phological awareness tasks imply conscious access to the morphemic structure of words and the ability to reflect
on that structure and to manipulate it (Carlisle 1995), the original Wug Test and the noun pluralization task that
we designed in this study seem to tap morphological processing skills more than morphological awareness and
general metacognition in children. For this reason, we will refer throughout this paper to morphological proces-
sing or morphological skills, although it cannot be excluded that this task involves some sort of metalinguistic skill
or conscious access to the morphemic structure of words (especially in a language like Italian and in older children
exposed to literacy, as in our study).
2. It is worthwhile reminding that declension classes do not trigger morphosyntactic agreement within the noun
phrase (for instance, in the noun phrase quest-o pirat-a ‘this pirate’ there is masculine gender agreement
between the determiner and the noun, but they belong to two different declension classes).
3. It is worth specifying that this mild correlation was found across all subjects and thus independently from bilin-
gualism; no effect of bilingualism was indeed reported for the two WM tasks that we administered, consistently
with other studies reporting that bilinguals and monolinguals show a similar performance in WM tasks (Engel de
Abreu 2011, amongst others; see Grundy and Timmer (2017) for a comprehensive meta-analysis on bilingualism
and WM capacity). It follows that performance in the WM task cannot be held responsible for the advantage
shown by bilingual children in our morphological task. Moreover, it must be noticed that the lack of WM advan-
tages in this study, apparently at odds with the results of studies on executive functions (see for instance Jalali-
Moghadam & Kormi-Nouri (2015) for a study on bilingual poor readers), could be due to the lower complexity of
our tasks, which tapped directly the subjects’ verbal WM, without specifically addressing the other components of
the EF, i.e. inhibitory control and cognitive flexibility (Diamond 2013). Finally, the fact that we found an advantage
in the pluralization of nonwords but not in the forward and backward digit span is arguably due to the fact that
these tasks rest on different abilities, morphological competence on the one side and memory on the other side.

Disclosure statement
No potential conflict of interest was reported by the authors.

Funding
This project has received funding from the European Union’s Seventh Framework Programme for research, technological
development and demonstration under grant agreement no 613465.

Notes on contributors
Maria Vender, Ph.D, is a Postdoctoral Researcher at the University of Veronainvolved in the FP7 European Project
AThEME, Advancing The European Multilingual Experience. Her research interests concern the investigation of linguistic
and cognitive deficits in language and learning disabilities, focusing in particular on the relationship between bilingual-
ism and developmental dyslexia.
Shenai Hu is an Associate Professor at the Department of Foreign Language Education, Xiamen University. She earned
her Ph.D in Linguistics in the Autonomous University of Barcelona and the University of Milan-Bicocca. Her research inter-
ests are in the areas of child language, reading disorders and language disorders with a special focus on bilingual chil-
dren. She is particularly interested in how aspects of language form - syntax in particular - interact with other aspects of
language development.
170 M. VENDER ET AL.

Federica Mantione (Ph.D., University of Verona in Italy, 2016) is a Postdoctoral Researcher in the Department of Psychol-
ogy and Cognitive Science at the University of Trento, Italy. She primarily studies the underlying causes of morphosyn-
tactic difficulties in the language of children with learning disabilities (in particular, Developmental Dyslexia). She is also
interested in cognitive adaptation to language variation phenomena measured by electroencephalogram (EEG) in sen-
tence comprehension tasks.
Silvia Savazzi, Ph.D, is an Associate Professor at the University of Verona. She is the director of the Perception and Aware-
ness (PandA) Lab. Her major research interest relates the study of perception and awareness in healthy participants and
brain-damaged patients with visual field defects or neglect. In her research, she uses several techniques: behavior, elec-
trophysiology (EEG), neuroimaging (EROS - fNIRS) and direct brain stimulation (TMS and intraoperative electrocortical
mapping).
Denis Delfitto graduated in philosophy and linguistics at the Scuola Normale Superiore in Pisa. From 1990 to 2001 he was
Associate Professor at the Utrecht University and from 1996 to 2001 he coordinated the syntax/semantics research group
at the Utrecht Institute of Linguistics. Since November 2001 he is Full Professor in General Linguistics at the Department
of Humanities of the University of Verona. His research interests include the systems of interpretation in natural language,
language impairment and many topics in the philosophy of mind.
Chiara Melloni, Ph.D., is an Assistant Professor in General Linguistics at the University of Verona, Italy. She is a researcher
in the FP7 European Project AThEME, Advancing The European Multilingual Experience, and conducts theoretical and
experimental research on the formal and interpretive properties of the lexicon. Her research interests concern word for-
mation and inflection, approached from a theoretical, comparative and developmental perspective.

ORCID
Maria Vender http://orcid.org/0000-0003-1985-9184

References
Bakker, M., A. van Dijk, and J. M. Wicherts. 2012. “The Rules of the Game Called Psychological Science.” Perspectives on
Psychological Science 7 (6): 543–554.
Barac, R., and E. Bialystok. 2012. “Bilingual Effects on Cognitive and Linguistic Development: Role of Language, Cultural
Background, and Education.” Child Development 83: 413–422.
Bar-Shalom, E., S. Crain, and D. Shankweiler. 1993. “A Comparison of Comprehension and Production Abilities of Good
and Poor Readers.” Applied Psycholinguistics 14: 197–227.
Belacchi, C., T. G. Scalisi, E. Cannoni, and C. Cornoldi. 2008. Manuale CPM. Coloured Progressive Matrices. Standardizzazione
Italiana. Firenze: Giunti OS Organizzazioni Speciali.
Berko, J. 1958. “The Child’s Learning of English Morphology.” Word 14: 150–177.
Bialystok, E. 1986. “Factors in the Growth of Linguistic Awareness.” Child Development 57: 498–510.
Bialystok, E., and R. Barac. 2013. “Cognitive Effects.” In The Psycholinguistics of Bilingualism, edited by F. Grosjean, and P. Li,
192–213. Chichester, England: Wiley- Blackwell.
Bialystok, E., F. I. M. Craik, D. W. Green, and T. H. Gollan. 2009. “Bilingual Minds.” Psychological Science in the Public Interest
10: 89–129.
Bialystok, E., F. I. M. Craik, and G. Luk. 2008. “Cognitive Control and Lexical Access in Younger and Older Bilinguals.”
Journal of Experimental Psychology: Learning, Memory, and Cognition 34: 859–873.
Bialystok, E., K. F. Peets, and S. Moreno. 2014. “Producing Bilinguals Through Immersion Education: Development of
Metalinguistic Awareness.” Applied Psycholinguistics 35: 177–191.
Bohman, T., L. Bedore, A. Peña Mendez-Perez, and R. Gillam. 2010. “What you Hear and What You Say: Language
Performance in Spanish-English Bilinguals.” International Journal of Bilingual Education and Bilingualism 13: 325–344.
Byrne, B. 1981. “Deficient Syntactic Control in Poor Readers: Is a Weak Phonetic Memory Code Responsible?” Applied
Psycholinguistics 2: 201–212.
Carlisle, J. F. 1995. “Morphological Awareness and Early Reading Achievement.” In Morphological Aspects of Language
Processing, edited by L. B. Feldman, 189–209. Hillsdale, NJ: Erlbaum.
Casalis, S., P. Colé, and D. Sopo. 2004. “Morphological Awareness in Developmental Dyslexia.” Annals of Dyslexia 54: 114–
138.
Cornoldi, C., and G. Colpo. 1995. Nuove prove di lettura MT per la scuola media inferiore. Firenze: Edizioni OS.
Cornoldi, C., and G. Colpo. 1998. Prove di Lettura MT per la Scuola Elementare-2. Firenze: Edizioni OS.
Costa, A., M. Hernández, and N. Sebastián-Gallés. 2008. “Bilingualism Aids Conflict Resolution: Evidence from the ANT
Task.” Cognition 106: 59–86.
Cummins, J. 1978. “Bilingualism and the Development of Metalinguistic Awareness.” Journal of Cross-Cultural Psychology
9: 131–149.
 INTERNATIONAL JOURNAL OF BILINGUAL EDUCATION AND BILINGUALISM 171

D’Achille, P., and A. M. Thornton. 2003. “La flessione del nome dall’italiano antico all’italiano contemporaneo.” In Italia
linguistica anno Mille - Italia linguistica anno Duemila. Atti del XXXIV congresso internazionale di studi della Società di
Linguistica Italiana, edited by N. Maraschio, and T. Poggi Salani, 211–230. Roma: Bulzoni.
D’Achille, P., and A. M. Thornton. 2008. “I nomi femminili in -o.” In Prospettive nello studio del lessico italiano. Atti del IX
Congresso Internazionale della Società di Linguistica e Filologia Italiana, edited by E. Cresti, 473–481. Firenze: Firenze
University Press.
Diamond, A. 2013. “Executive Functions.” Annual Review of Psychology 64: 135–168.
Dressler, W. U., and A. M. Thornton. 1996. “Italian Nominal Inflection.” Wiener Linguistische Gazette 57–59: 1–26.
Dunn, L. M., and L. M. Dunn. 2000. “Peabody–Test di Vocabolario Recettivo.” In Adattamento italiano e standardizzazione a
cura di, edited by G. Stella, C. Pizzioli, and P. E. Tressoldi. Torino, Italy: Omega Edizioni.
Engel de Abreu, P. M. J. 2011. “Working Memory in Multilingual Children: Is There a Bilingual Effect?” Memory 19: 529–537.
Frith, U. 1986. “A Developmental Framework for Developmental Dyslexia.” Annals of Dyslexia 36: 67–81.
Grundy, J. G., and K. Timmer. 2017. “Bilingualism and Working Memory Capacity: A Comprehensive Meta-Analysis.”
Second Language Research 33: 325–340.
Hilchey, M. D., J. Saint-Aubin, and R. M. Klein. 2015. “Does Bilingual Exercise Enhance Cognitive Fitness in Traditional non-
Linguistic Executive Processing Tasks?” In The Cambridge Handbook of Bilingual Processing, edited by J. Schwieter, 586–
613. Cambridge: Cambridge University Press.
Jalali-Moghdam, N., and R. Kormi-Nouri. 2015. “The Role of Executive Functions in Bilingual Children with Reading
Difficulties.” Scandinavian Journal of Psychology 56: 297–305.
Jeffries, S., and J. E. Everatt. 2004. “Working Memory: Its Role in Dyslexia and Other Specific Learning Difficulties.” Dyslexia
10: 196–214.
Jiménez, J. E., E. Garcìa, A. Estévez, A. Dìaz, R. Guzman, I. Hernandez-Valle, M. Rosario, M. Rodrigo, and S. Hernandéz. 2004.
“An Evaluation of Syntactic-Semantic Processing in Developmental Dyslexia.” Electronic Journal of Research in
Educational Psychology 2: 127–142.
Joanisse, M. F., F. Manis, P. Keating, and M. S. Seidenberg. 2000. “Language Deficits in Dyslexic Children: Speech
Perception, Phonology and Morphology.” Journal of Experimental Child Psychology 77: 30–60.
Lyytinen, P., A. M. Poikkeus, M. L. Laakso, K. Eklund, and H. Lyytinen. 2001. “Language Development and Symbolic Play in
Children with and Without Familial Risk for Dyslexia.” Journal of Speech Language and Hearing Research 44: 873–885.
Mantione, F. 2016. On the Production of Functional Categories in Children with Dyslexia. A Study on Pronouns, Articles
and Prepositions. Doctoral dissertation, University of Verona.
Melloni, C., M. Vender, and D. Delfitto. to appear. Nonword Pluralization: Evidence for an Advantage of Bilingualism in
Albanian-Italian and Romanian-Italian Bilingual Children. Proceedings of the 14th Generative Approaches to
Second Language Acquisition Conference (GASLA14).
Marinis, T. 2003. The Acquisition of the DP in Modern Greek. Vol. 31. Amsterdam / Philadelphia: John Benjamins Publishing.
McLoughlin, D., C. Leather, and P. Stringer. 2002. The Adult Dyslexic: Interventions and Outcomes. London: Whurr.
Miyake, A., N. P. Friedman, M. J. Emerson, A. H. Witzki, A. Howerter, and T. D. Wager. 2000. “The Unity and Diversity of
Executive Functions and Their Contributions to Complex “Frontal Lobe” Tasks: A Latent Variable Analysis.”
Cognitive Psychology 41: 49–100.
Paap, K. P. in press. “The Bilingual Advantage Debate: Quantity and Quality of the Evidence.” In The Handbook of the
Neuroscience of Multilingualism, edited by J. W. Schwieter. Wiley-Blackwell.
Paap, K. P., H. A. Johnson, and O. Sawi. 2015. “Bilingual Advantages in Executive Functioning Either do not Exist or are
Restricted to Very Specific and Undetermined Circumstances.” Cortex 69: 265–278.
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.
Paradis, J., and A. Kirova. 2014. “English Second-Language Learners in Preschool: Profile Effects in their English Abilities
and the Role of Home Language Environment.” International Journal of Behavioral Development 38: 342–349.
Pickering, S. J., and S. E. Gathercole. 2001. Working Memory Test Battery for Children. London: Pearson Assessment.
Ramus, F., and G. Szenkovits. 2008. “What Phonological Deficit?” The Quarterly Journal of Experimental Psychology 61: 129–
141.
Raven, J., J. H. Court, and J. C. Raven. 1998. Raven Manual, Section 1 (General Overview) and Section 2 (Coloured Progressive
Matrices). Oxford, UK: Oxford Psychologist Press.
Rispens, J. 2004. “Syntactic and Phonological Processing in Developmental Dyslexia.” Doctoral dissertation., University of
Groningen, Groningen.
Robertson, E. K., and M. F. Joanisse. 2010. “Spoken Sentence Comprehension in Children with Dyslexia and Language
Impairment: The Roles of Syntax and Working Memory.” Applied Psycholinguistics 31: 141–165.
Sartori, G., R. Job, and P. E. Tressoldi. 2007. Batteria per la Valutazione della Dislessia e della Disortografia Evolutiva-2.
Firenze: Giunti OS Organizzazioni Speciali.
Scarborough, H. S. 1990. “Very Early Language Deficits in Dyslexic Children.” Child Development 61: 1728–1743.
Snowling, M. J. 1995. “Phonological Processing and Developmental Dyslexia.” Journal of Research in Reading 18: 132–138.
Snowling, M. J., A. Gallagher, and U. Frith. 2003. “Family Risk of Dyslexia is Continuous: Individual Differences in the
Precursors of Reading Skill.” Child Development 74: 358–373.
172 M. VENDER ET AL.

Unsworth, S., F. Argyri, L. Cornips, A. Hulk, A. Sorace, and I. Tsimpli. 2012. “The Role of age of Onset and Input in Early Child
Bilingualism in Greek and Dutch.” Applied Psycholinguistics 33: 1–41.
Vender, M. 2017. Disentangling Dyslexia: Phonological and Processing Impairment in Developmental Dyslexia. Frankfurt:
Peter Lang.
Vender, M., M. Garraffa, A. Sorace, and M. T. Guasti. 2016. “How Early L2 Children Perform on Italian Clinical Markers of SLI:
A Study of Clitic Production and Nonword Repetition.” Clinical Linguistics & Phonetics 30: 150–169.
Vender, M., F. Mantione, S. Savazzi, D. Delfitto, and C. Melloni. 2017. “Inflectional Morphology in Dyslexia: Italian Children’s
Performance in a Nonword Pluralization Task.” Annals of Dyslexia 67: 401–426.
Vellutino, F. R. 1979. Dyslexia: Theory and Research. Cambridge, MA: MIT Press.
Vygotsky, L. 1962. Thought and Language. Cambridge, MA: MIT Press.
Waltzman, D., and H. S. Cairns. 2000. “Grammatical Knowledge of Third Grade Good and Poor Readers.” Applied
Psycholinguistics 21: 263–284.
World Health Organization. 2004. International Statistical Classification of Diseases and Health Related Problems (ICD-10,
2nd ed.). Geneva, Switzerland.