The Narrative Abilities of an English-Spanish Bilingual with Prader-Willi Syndrome

The aim of this study was to analyze the narrative abilities of a 33-year-old English-Spanish bilingual with Prader-Willi syndrome (PWS). The few previous linguistic studies examining monolinguals with PWS have focused primarily on these individuals’ narrative capacity, revealing a performance deficit in this area (Lewis et al. 2002; Garayzábal-Heinze et al. 2012). The present study is novel in that it examines a bilingual speaker and also tests his narrative abilities in both languages. Two wordless picture books from Mayer’s (1967, 1969) Frog story series were used as the elicitation method. The PWS bilingual produced, over two experimental sessions, four narratives (two in each language), which were compared to four analogous narratives produced by a 25-year-old typically developing bilingual with a comparable linguistic background and proficiency level in Spanish and English. Following Gonçalves and collaborators’ (Gonçalves et al. 2001a, 2001b, 2001c) narrative evaluation protocol, the narratives were analyzed according to three dimensions: narrative structure and coherence, narrative process and complexity, and narrative content and multiplicity. Overall, the results revealed that the PWS bilingual (1) presented a poor narration ability in both languages, with narrative content and multiplicity being the least impaired; (2) showed better narrative abilities during the second experimental session (i.e., narrative abilities improved with experience/practice); and (3) did not show typically developing behavior but a comparable performance to that of monolingual speakers with PWS. These findings suggest that bilingualism should not be discouraged in PWS populations and that special attention should be given to the development of their narrative abilities in their school curriculum.


Introduction
Individuals with developmental and intellectual disabilities (henceforth DDs and IDs, respectively) are often discouraged from learning two languages simultaneously or from having access to a second language (henceforth L2) under the assumption that this will have a negative effect on the linguistic development of their first language (henceforth L1) (Paradis et al. 2011).This practice not only denies these individuals the possibility of becoming bilingual, but also-in all probability-the option of taking advantage of the general benefits associated with bilingualism (which are discussed in Section 1.1).Now, however, there is a growing body of research focusing on bilingualism in individuals with special needs that is not only questioning this practice but pointing out a completely opposite scenario (i.e., the encouragement of bilingualism and multilingualism).It is undeniable that there are valuable data provided by previous studies that have focused on bilingualism in individuals with DDs; nevertheless, this line of research has been narrowed to Specific Language Impairment (SLI), Autism Spectrum Disorders (ASD), and Down syndrome (DS) (see Kay-Raining Bird et al. 2016 for a complete overview).Thus, in order to have a more comprehensive view of bilingualism in non-typically developing (henceforth non-TD) individuals, it is imperative to expand the focus to other syndromes and conditions, as is the case with the emerging studies on bilingualism in Williams syndrome (Perovic and Lochet 2015).Accordingly, this study seeks to contribute to this developing field by providing new data from a bilingual speaker with PWS, an understudied DD.In this paper, I analyze the narrative abilities of an English-Spanish bilingual with PWS in both languages in comparison with a typically developing (henceforth TD) English-Spanish bilingual of similar age and sociolinguistic background.

Bilingualism in Typically Developing Individuals
Even though nowadays it seems that being bilingual is more of an asset than a problem, this has not always been the case.From the early 1920s, when the first studies focused on bilingualism and cognition arose, to the early 1960s, when Peal and Lambert's (1962) revolutionary research work was published, researchers expounded bilingualism's negative effects on intellectual abilities.Studies from this 40-year-period systematically reported the inferior linguistic abilities of bilinguals when compared to monolinguals.As Diaz (1983) and Hakuta and Diaz (1985) summarized, this bilingual disadvantage was mainly reflected in deficient articulation (Carrow 1957), poor vocabulary (Barke and Williams 1938), poor writing skills, and a greater number of grammatical errors (Harris 1948).However, while in verbal tasks researchers repeatedly showed either a bilingual disadvantage or a neutral effect, contradictory findings were found for non-verbal experiments (see Peal and Lambert 1962;Diaz 1983 for a complete review).With the publication of Peal and Lambert's (1962) seminal work, the reliability of previous findings started to be questioned, since, for the first time, a balanced bilingual group matching in terms of class, sex, and age to those of a monolingual group was included in a study (Canadian English-French bilinguals compared to Canadian monolingual French speakers).This special attention to participant selection was crucial as the careless criteria used to determine bilingualism in previous studies could have influenced the negative results obtained for bilinguals.For example, Brunner (1929) determined levels of bilingualism in terms of parents' origin, which today is known to not be a very precise method of determining the degree of bilingualism.Peal and Lambert's results showed that when participants were carefully selected, bilinguals showed an advantage over monolinguals in both verbal and non-verbal tasks.Nevertheless, these findings were not free from controversy and some researchers pointed out that the positive results for bilingualism obtained could have been the result of an involuntary bilingual bias.The authors, in an attempt to include only balanced bilinguals, could have favored the positive effects of bilingualism because they only included bilinguals who obtained a certain result in the intelligence test administered (see Hakuta and Diaz 1985, pp. 322-23, for a complete review of the possible bilingual bias in this study).Be that as it may, since the early 1960s a substantial body of research has revealed that bilingualism seems to lead to different cognitive benefits.
Following Adescope et al.'s (2010) comprehensive review of previous work focused on the cognitive effects of bilingualism, bilinguals appear to have: better attentional control in non-verbal tasks (Bialystok 2001;Emmorey et al. 2008); greater metalinguistic awareness (Galambos and Goldin-Meadow 1990;Yelland et al. 1993) and metacognitive awareness 1 (Ransdell et al. 2006); better optimization of the working memory capacity (Bialystok et al. 2004;Morales et al. 2013); better management of problem-solving skills (Bialystok 1999(Bialystok , 2006)); superior abstract or symbolic analysis capacity (Cummins 1976); greater creative thinking (Ricciardelli 1992;Baker 2001); and more diverse communication strategies (Thomas 1992).Recently, different studies have revealed an interaction between bilingualism and certain narrative abilities in both TD and non-TD individuals.Andreou (2015) analyzed the narrative performance of TD bilingual children at the microstructural (morphosyntactic abilities) and macrostructural (coherence and global organization abilities) levels.Her results showed, on the one hand, that bilinguals outperformed monolinguals at a macrostructural level and, on the other hand, that bilingual children developed better global narrative 1 While metalinguistic awareness refers to the ability to think about the language, metacognitive awareness refers to the capacity to reflect on the learning process itself.
Languages 2018, 3, 23 3 of 28 abilities earlier than their monolingual counterparts (9 vs. 11 years old approximately).Better narrative abilities at macrostructural level were also found in bilinguals with SLI (Tsimpli et al. 2016) and with high-functioning autism (Baldimtsi et al. 2016) when compared to their monolingual peers.In addition, studies from the last decade have revealed that bilingualism seems to delay the age of onset of dementia and cognitive aging (Bialystok et al. 2007;Bialystok and Craik 2010).Finally, one obvious, but often undervalued, benefit of bilingualism in scientific research is the intrinsic capacity to communicate and interact in two languages (or more in the case of multilingualism) and all the social benefits associated with it.Being bilingual gives a person the opportunity to have access to another language, to its culture, and, most importantly, to the speakers of that language.In the end, humans are social beings wanting to socialize and be part of a community.

Bilingualism in Individuals with Developmental and Intellectual Disabilities
Despite the large number of benefits to becoming bilingual, populations with DDs have been discouraged from being raised bilingual or from studying an L2 for fear of increasing their linguistic learning challenges (Cleave et al. 2014).Nevertheless, cases of bilingualism have been attested in individuals with DS (Kay-Raining Bird et al. 2016), ASD (Seung et al. 2006), or SLI (Sanz-Torrent et al. 2008), to name a few, and it is precisely these three conditions on which previous research investigating DDs and bilingualism has almost exclusively focused.Since the aim of this study is to analyze the narrative abilities of a bilingual with PWS, the revision of previous findings is primarily focused on DS.The reason for this is that within the heterogeneity of DDs, PWS and DS are, in principle, closer than PWS and SLI or PWS and ASD.As Cimolin et al. (2010, p. 1) state, "Prader-Willi and Down syndrome are two different chromosomal disorders characterized by some common clinical features, such as obesity, muscular hypotonia, ligament laxity, and mental retardation".SLI differs from DS and PWS in that it is a language disorder "diagnosed when a child's language development is deficient for no obvious reason" (Bishop 2006, p. 217), or, in other words, when it "cannot be explained in terms of mental or physical handicap, hearing loss, emotional disorder, or environmental deprivation" (Bishop 1992, p. 3).Thus, according to this description, individuals with SLI should display, by definition, typical non-verbal abilities and intelligence.Previous research, however, has shown that these individuals are not exempt from experiencing difficulty with some particular non-linguistic tasks, as seems to be the case with visuo-spatial short-term memory tasks (Hick et al. 2005) or mathematical tasks (Cowan et al. 2005;Mainela-Arnold et al. 2011) to name a few.ASD, for its part, is related to impairment of verbal and non-verbal communication, social interactions, and repetitive behaviors and interests (Tager-Flusberg 2004;Faras et al. 2010).Traditionally, ASD has been directly associated with intellectual disabilities, but recent research has revealed that, despite the fact that around 50% of children with an autistic disorder present intellectual disabilities, some individuals may show average or even superior intelligence (see Charman et al. 2011 for an overview).
Previous research that focused on bilinguals with DS is very scarce, and, at its beginning, was based on case study research, which, interestingly enough, revealed homogeneous outcomes: the possibility of becoming bilingual, or even multilingual, and facing the same language difficulties in the different languages.For example, Vallar and Papagno (1993) studied a 23-year-old L1 Italian woman with DS who was also fluent in English (and French to a lesser extent).The authors focused on vocabulary acquisition and their results revealed that the participant's phonological short-term memory appeared to not be impaired, and, in fact, seemed to facilitate the process of vocabulary acquisition.For their part, Woll and Grove (1996) studied the case of two bimodal, bilingual identical twins (English-British Sign Language) with DS whose parents were deaf.These twins were administered different verbal and non-verbal tests and their spontaneous production was analyzed in both languages.The results evidenced that the twins were fluent when signing and that the shortcomings detected were consistent across the two language modalities.
According to Kohnert and Medina (2009) and the recent exhaustive bibliography review by Kay-Raining Bird et al. (2016), apart from case studies, only five group studies analyzing bilingualism in DS have been carried out 2 .It must be stressed that four out of these five studies were conducted by the same group of researchers and that the participants partially overlapped, with English-French being almost the only bilingual language combination under analysis (see Kay-Raining Bird et al. 2016, p. 7).Trudeau et al. (2011) carried out a cross-sectional and a longitudinal study with 18 sequential English-French bilinguals with DS.They investigated both expressive and receptive vocabulary in English (L1) and French (L2) and found that the bilinguals' French vocabulary was smaller and had greater variability than their English vocabulary but that it became more uniform as their overall vocabulary size increased.The longitudinal study showed that, while in English a vocabulary growth was detected over time for 13 participants, this was not the case for the L2, since only nine individuals showed progress in French.Regardless, what these results suggest is that under no circumstances did the fact of being exposed to an L2 appear to have a negative effect on the vocabulary acquisition of the L1.In order to evaluate the cognitive effects of being exposed to two languages, Edgin et al. (2011) administered a battery of cognitive tests to both English-Spanish bilinguals with DS and English monolinguals with DS (control group).Results showed parallel performance of both groups for all the measures under analysis.Likewise, Kay-Raining Bird et al. (2005) investigated the language abilities (mainly vocabulary comprehension) of eight bilinguals (English-one other language) with DS in comparison to 14 English monolinguals with DS and they found that both bilinguals and monolinguals with DS shared a similar profile.For their part, Feltmate and Bird (2008) analyzed the morphosyntactic and vocabulary abilities of eight English-French bilinguals with DS and four English monolinguals with DS.Their findings showed no detrimental effects of bilingualism (i.e., bilinguals with DS did not demonstrate linguistic deficits when compared to monolinguals with DS).Two of the aforementioned studies, Kay-Raining Bird et al. (2005) and Feltmate and Bird (2008), also included a TD bilingual group as a comparison baseline.Both studies agreed that while bilinguals with DS did not seem to perform much different from their monolingual counterparts, they did appear to have linguistic difficulties when compared to TD bilinguals.Nevertheless, Cleave et al. (2014) found that, in specific areas, bilinguals with DS did not differ greatly from TD bilinguals, as is the case of the ability to learn new words.
Overall, previous case studies and group studies consistently reveal that bilinguals with DS show different linguistic and cognitive deficiencies when compared to TD bilinguals but not when compared to their monolingual peers.Thus, these results, as well as those from other studies based on bilingualism in individuals with DDs (see Kay-Raining Bird et al. 2016), show that there seems to be no scientific foundation to deny individuals with these special needs the possibility of managing more than one language and, consequently, the opportunity to benefit from the positive cognitive outcomes that appear to derive from being bilingual.

Clinical Characteristics of the Prader-Willi Syndrome and their Impact on Language/Speech
The Prader-Willi Syndrome (PWS) is a rare disease first described by Prader et al. (1956), and is not related to sex (Burd et al. 1990) or other factors such as race, social class, or geographical origin (Alexander and Hanson 1988;Butler and Thompson 2000).Genetic research has shown that PWS is rarely inherited and the recurrence risk within the same family is normally low (Buiting and Horsthemke 2006).It must be emphasized that although PWS has a relative low incidence (it affects one in every 10,000-15,000 humans) (Lewis et al. 2002;Dimitropoulos et al. 2013), it is one of the most frequent syndromes among those recognized (Alexander and Hanson 1988).In a simplistic way, the reason why this malformation occurs is related to the fact that, at conception, the crossing over that 2 Both Kohnert and Medina's (2009) and Kay-Raining Bird et al.'s (2016) reviews comprise research papers published mainly in English, and French to a lesser extent.It is possible that other studies have been published in languages other than English or French but have gone unnoticed.
Languages 2018, 3, 23 5 of 28 happens between the 23 paternal and the 23 maternal chromosomes, when they line up to form 46 chromosomes, is imperfect, which causes health and cognitive "abnormalities" in the individual.For example, DS, one of the genetic disorders with higher incident rates, is the result of three copies of chromosome 21.Prader-Willi Syndrome, for its part, is the result of problems in chromosome 15, which may have three main manifestations (see Cassidy et al. 2012 for a comprehensive genetic explanation): (1) Chromosomal deletion (the absence of part of the paternal chromosome 15); (2) Uniparental disomy (two maternal copies of chromosome 15); and (3) Imprinting defect (defined by (Cassidy et al. 2012, p.10), as a "defect in the genomic region that controls the imprinting process").Chromosomal deletion affects about 75% of the PWS population, followed by Uniparental disomy, which causes about 20% of cases and, finally, by Imprinting defect, which only has about 5% incidence (Cassidy et al. 2012;Dimitropoulos et al. 2013).
Even though there is variation in terms of the severity and outcomes, individuals with PWS present specific behavioral, physical, and cognitive characteristics.Their most salient trait is food anxiety, which normally leads to morbid obesity if their insatiable desire to eat is not controlled.They also tend to be stubborn, manipulative, and compulsive, and to experience temper tantrums as well as other behavioral issues.As for physical characteristics, they tend to be short and overweight and to have poor muscle tone (hypotonia) and idiosyncratic facial features.Likewise, intellectual disabilities (ranging from mild to moderate) and delay of language development are rather common in this population (see Greenswag and Alexander 1988;Cassidy et al. 2012 for a comprehensive review of the main characteristics of PWS).Previous studies show that PWS individuals have a mean Intellectual Coefficient (IQ) of 62.3 3 (Curfs et al. 1991) and that their non-verbal IQ is normally higher than their verbal one (Dimitropoulos et al. 2013).Therefore, these results demonstrate that linguistic abilities are not one of their strengths.It should also be emphasized that both receptive and expressive language are impaired, the latter being poorer than the former (Kleppe et al. 1990;Lewis 2006).Similarly, if the Mean Length of Utterance (MLU 4 ) is taken as a measure of language productivity, this linguistic skill is also commonly impaired, since PWS individuals' speech samples show shorter MLUs than those of the same age TD peers (Lewis et al. 2002;Lewis 2006).
Succinctly, as Lewis et al. (2002, pp. 287-88) and Dimitropoulos et al. (2013, p. 194) put it, PWS individuals show relative strengths in the following capacities: reading decoding, reading comprehension, receptive vocabulary, visual-spatial skills, picture completion, block design, object assembly, and long-term memory.On the other hand, they show deficiencies in short-time memory, as well as with auditory verbal processing and linear or temporal order processing skills.Additionally, PWS individuals show a poor expressive vocabulary, poor pragmatic abilities, and a considerable language delay, since their first words are produced between 18 months and six years (Lewis 2006), which contrasts with the pattern observed in TD children, who at 18 months start to produce their first signs of grammaticalization and generally by four years their core grammatical system is completed (Serra et al. 2000, p. 51).
Some clinical features of PWS, together with reduced cognitive and phonological abilities, result in some speech idiosyncrasies (Lewis et al. 2002;Lewis 2006).For example, having a narrow overjet and palatal arch, as well as an undersized jaw and some dentition problems, affects their articulatory skills.In accordance with Lewis (2006, p. 274), the principal speech-sound problems of PWS are "sound distortions and omissions, vowel errors, simplification of consonant blends, and difficulty sequencing syllables".In addition, poor muscle tone and abnormal laryngeal growth have an impact on their voice level.Akefeldt et al. (1997) studied the oral motor function, pitch level, and resonance of 11 individuals with PWS and found certain abnormal vocal features.Similar outcomes were found in Munson-Davis (1988), Kleppe et al. (1990), andDefloor et al. (2001).Likewise, nasality has also been reported as a common characteristic of this syndrome (Branson 1981;Edmonston 1982).With regard to speech fluency, previous research has revealed that PWS individuals consistently show dysfluency.Defloor et al. (2000) studied the speech fluency of 15 PWS individuals through four different speech modalities (spontaneous, repetition, monologue, and automatic series) and their results revealed "within-word dysfluencies such as part-word repetitions, whole-word repetitions, blocks, prolongations, and broken words" (Defloor et al. 2000, p. 95).These authors defended that these dysfluencies did not obey to stuttering and/or cluttering, but to cognitive and language deficits, a position also defended by Branson (1981); Kleppe et al. (1990);and Lewis (2006).
In sum, previous research on PWS has mainly focused on its clinical, cognitive, and behavioral characteristics while its linguistic abilities have received little attention.Overall, linguistic research has focused on two main topics: speech characteristics and narrative abilities.Having presented an overview of the main findings related to speech characteristics, in the next section I will concentrate on the narrative abilities of the PWS population.I will first provide an overview of the narrative framework that forms this study and I will subsequently scrutinize the narrative abilities of monolingual speakers with PWS.

Narrative Abilities in Prader-Willi Syndrome
One of the most frequent activities in a person's daily life, to a greater or lesser extent, consists in narrating personal events (Fivush and Nelson 2004;Gonçalves et al. 2011), a capacity that emerges at a very young age as a way to organize human experience (Bruner 1990) and that does not differ significantly from the way fictitious stories are narrated (Anderson and Conway 1993).Given that narrating is a very common and familiar activity for all of us, we might be inclined to believe that it is a relatively simple and straightforward task.Nonetheless, the reality is that narrating properly, either personal or invented stories, is a highly demanding task that entails the integration of linguistic, cognitive, and social domains (Tsimpli et al. 2016).Thus, given the complexity of this activity, even for TD individuals, it should not be surprising that PWS individuals show consistent deficiencies in this respect, as will be detailed later in this section.Children and adult's narrative productions have been extensively used by linguistic researchers to analyze the writer/speakers' narrative abilities and/or their linguistic competence or productivity.It is worth noting that while there are different reliable analysis proposals to evaluate the quality of written or oral narratives produced by TD individuals, the options specifically conceived to analyze the narrative production of individuals with special needs are more limited.In this study, I follow Gonçalves and collaborators' narrative evaluation scheme (Gonçalves et al. 2001a(Gonçalves et al. , 2001b(Gonçalves et al. , 2001c)), which consists in the analysis of three dimensions (narrative structure and coherence, narrative process and complexity, and narrative content and multiplicity), and their corresponding subdimensions, which are presented in Table 1 following the authors' description.
This methodology was originally conceived to evaluate the narrative abilities of individuals with psychological disorders and its reliability was satisfactorily tested in patients with agoraphobia (Gonçalves et al. 2002).Subsequently, it was extended to the assessment of the narrative abilities in individuals with neurodevelopmental disorders such as Williams syndrome, Smith-Magenis syndrome, and PWS (Gonçalves et al. 2011;Garayzábal-Heinze et al. 2012).

Orientation
Reference to the circumstances that surround an action.

Structural sequence
Events are presented in a logical sequence.

Evaluative commitment
The narrator shows emotional involvement.

Integration
The narrative is presented in an integrated way (it makes sense).

Narrative Process and Complexity
1. Objectifying Sensorial information is included.

Emotional subjectifying
Reference to emotional states is included.

Cognitive subjectifying
The narrator's thoughts are included.

Metaphorizing
Metaphors are included.

Narrative Content and Multiplicity
1. Themes Variety of themes/topics is included.

Events
Variety of events/actions is included.

Scenarios
Explicit reference to the atmosphere that surrounds an action.

Characters
Variety of characters is included.
It is important to highlight that although a great deal of individual variability exists between the PWS population, some common idiosyncratic narrative features have been identified in previous studies.Lewis et al. (2002) and Garayzábal-Heinze et al. (2012), using different techniques and narrative analysis, studied the narrative abilities of monolinguals with PWS.By means of a narrative retelling task, Lewis et al. (2002) investigated whether 24 participants (eight preschools, eight school-age children, and eight adolescents/adults) included the grammar components and content items presented in the original story in their own narrative production.In addition, and in order to measure the participants' comprehension capacity, participants were asked three factual and three inferential questions.Results revealed important deficits in both narrative abilities and comprehension questions in the three groups.However, these deficiencies were more evident in younger participants, which indicates that they do not disappear with age but rather improve in adulthood, which, in turn, can be interpreted as an experience/practice gain.Garayzábal-Heinze et al. (2012), for their part, compared the narrative abilities of two individuals with PWS with those observed in two individuals with Williams syndrome and two with Smith-Magenis syndrome (all microdeletion syndromes characterized by cognitive impairment and language deficits).Using the well-known picture book Frog, Where Are You? (Mayer 1969) to elicit speech samples, participants were asked to explain the story to the examiner, who recorded the sessions.Narratives were transcribed and analyzed according to Gonçalves et al.'s (2001aGonçalves et al.'s ( , 2001bGonçalves et al.'s ( , 2001c) ) narrative evaluation proposal presented in Table 1.Results showed that all participants, regardless of the syndrome, exhibited poor narrative abilities to different extents.PWS individuals had the lowest performance in both the narrative structure and coherence and the narrative process and complexity dimensions, while the opposite trend was observed in the narrative content and multiplicity dimension: PWS and Williams syndrome participants obtained the highest scores.These results, together with those of Lewis et al. (2002), are evidence that narration is an arduous task for these individuals and point to the need to include this content in their school curriculum.
Since research on the linguistic abilities of individuals with PWS has been sparse and has only addressed monolingual speakers, in this study I investigate the narrative abilities of an English-Spanish bilingual with PWS in comparison to a TD English-Spanish bilingual and to the two Spanish monolingual speakers with PWS in Garayzábal-Heinze et al.'s (2012) study.The results will shed light on how the narrative abilities of a bilingual with PWS compare in the two languages and will represent a first step towards the establishment of a new line of research that may help speech therapists and other intervention professionals to make more informed decisions with regard to PWS and bilingualism.At the same time, the outcomes of this study will contribute to the line of research initiated by Kay-Raining Bird and collaborators focused on bilingualism in populations with IDs by providing new data from a different DD with ID and from a previously understudied language combination.This type of study is crucial since, as Garayzábal-Heinze et al. (2012, p. 49) argue, the analysis of the narrative production in non-TD individuals with IDs is of particular relevance because it is normally accompanied by an affected language system and this, in the authors' words, "could elucidate about the genetic contributions to language and narrative organization in atypical neurodevelopmental genetically based disorders".

Research Questions
In this study, I intend to answer the following research questions: Research question 3: How are the narrative abilities of an English-Spanish bilingual with PWS compared to a TD English-Spanish bilingual?

Participants
One bilingual with PWS (Yves) and one TD bilingual (Olivia) 5 participated in the study.Both participants were asked to complete a language background questionnaire.Yves is a 33-year-old male English-Spanish bilingual who was diagnosed with Prader-Willi syndrome (Chromosomal deletion) when he was 11 years old.Olivia is a 25-year-old female TD English-Spanish bilingual.Both participants were raised in Ontario (Canada) and are simultaneous English-Spanish bilinguals.In both cases, English is the dominant language and Spanish is the heritage language.They both belong to Spanish-speaking families (both parents) from Spain who maintain Spanish as the family communication language but speak predominantly English in social and professional circles.Olivia lives independently and works in an Anglophone context; Yves lives in an Anglophone group home and visits his family every other weekend.Both participants gave their consent to participate in this study, which was approved by the ethics committee of the University of Ottawa (file reference: #06-16-19).

Instruments
Besides completing a language background questionnaire, both participants were asked to complete different tasks to evaluate their proficiency level in English and Spanish and their narrative abilities in both languages.

Proficiency Tests
In order to assess the participants' proficiency level in both English and Spanish, they were administered the Oxford Quick Placement Test (UCLES 2001) and a short version of the Wisconsin Spanish Placement Test (2009) (two of the most common standardized tests used in the second language acquisition field to evaluate knowledge of English and Spanish as foreign languages), as well 5In order to preserve the participants' anonymity, the two names are fictional.
Languages 2018, 3, 23 9 of 28 as separate cloze tests in both English and Spanish (used extensively in the Language Acquisition Research Laboratory at the University of Ottawa).The tests' results, reported in Table 2, revealed that both participants had a high competence in both languages.In English, both participants showed 96% accuracy in the cloze test and were considered advanced speakers of English according to the results from The Oxford Quick Placement Test (Yves showed a C1 level conforming to the CEFR 6 and Olivia a C2 level).In the case of Spanish, results also revealed an advanced level of proficiency, since both participants showed more than 70% of accuracy in the cloze test and in the Wisconsin Spanish Placement Test.

Narrative Elicitation Tasks
Participants were presented with two wordless picture books, A Boy, a Dog and a Frog (Mayer 1967) and Frog, Where Are You? (Mayer 1969), and were asked to explain the story to the investigator.This elicitation method, popularly known as the Frog story task, has been widely used in linguistic research in both TD and non-TD populations and for different languages (see Strömqvist and Verhoeven 2004 or Slobin 2005 for a bibliography).In both picture books, there are two main characters, a boy and his dog, that look for or try to catch a frog.Both picture books are related but are completely autonomous, since the context and the adventures the characters experience throughout the two picture books are different.In A Boy, a Dog and a Frog (25 images) (Mayer 1967), a boy and his dog are on a fishing trip and see a frog they would like to catch.The frog escapes and they both try to chase it.They finally give up and go back home, only to have the frog follow them there.The picture book Frog, Where Are You? (25 images) (Mayer 1969) shows the frog escaping from the boy's home during the night, followed by the boy and his dog searching for it in the forest.Before finding the frog, their adventure includes encounters with a variety of animals.
Taking into account the different characters involved and the different situations and contexts depicted by the images, many researchers defend and agree in considering the Frog story series as a legitimate tool to elicit narrative texts (oral and/or written).With the appropriate data analysis tools, these narratives can provide appropriate speech and written samples from which to analyze the participants' narrative abilities.

Procedure
Data collection took place in four sessions.In the first session, the participants were given the research project details and were both asked to orally narrate A Boy, a Dog and a Frog (Mayer 1967) in English and Frog, Where Are You? (Mayer 1969) in Spanish.The narratives were all audio recorded.During the second session, the narration languages were reversed so that A Boy, a Dog and a Frog (Mayer 1967) was asked to be narrated in Spanish and Frog, Where Are You? (Mayer 1969) in English.In both sessions the task procedure was the same: participants were asked to narrate one story first and, once finished, the participant and the investigator talked about trivial things about the story they had just narrated.Once the participants felt comfortable and ready to continue, they were asked to narrate the second story.During the first experimental session participants did not have access to the pictures before starting the narrative task, but they did know the title of the story they were about to narrate.During the second experimental session, however, they knew both stories since they had already narrated them in the previous session (though in a different language).Between the first and the second sessions, there was an average of a two-and-a-half-month gap.In sessions three and four, participants completed the language background questionnaire, the two cloze tests, and the two proficiency tests previously mentioned.

Data Treatment
The eight oral narratives (both participants produced two narratives in Spanish and two narratives in English) were transcribed in CHAT, the transcription system for the CHILDES Project (MacWhinney 2000).Once transcribed, following Gonçalves et al.'s (2001aGonçalves et al.'s ( , 2001bGonçalves et al.'s ( , 2001c) methodology, the narratives were evaluated according to the dimensions and subdimensions presented in Table 1 (see Garayzábal-Heinze et al. 2012, pp. 63-65;Gonçalves et al. 2011, pp. 292-94 for a detailed explanation of the three dimensions and their analysis).
The narratives produced were evaluated by four linguists specialized in the language acquisition field, and each subdimension, following the original scheme, was evaluated on a 5-value Likert scale (1 = very low, 2 = low, 3 = moderate, 4 = high, 5 = very high).The dependent variable was the mean obtained by the four evaluators for each subdimension (the standard deviation was between 0 and 1.26 for all the subdimensions evaluated).

Results
Due to the fact that this is a case study involving two participants, a descriptive data analysis was performed.This type of data analysis is the most extended practice within the field due to the recurrent limited number of participants (Garayzábal-Heinze et al. 2012;Feltmate and Bird 2008).I first present an overview of the characteristics of the narratives and then an analysis of the participants' narrative abilities.

Characteristics of the Narratives
Table 3 shows the number of words, the mean length of utterances by words (MLUw), and the Type-token ratio for each of the participants' narratives.These three measures reflect, respectively, the length of the transcribed recording, the participants' linguistic productivity, and their vocabulary variation.Overall, Olivia showed better linguistic productivity and richer vocabulary than Yves.However, it is interesting to note that, while in Yves' case the narratives' length and MLUw increase through experience/practice (here understood as a convergence of two factors: the knowledge of the characters and the plot of both stories from session 1 to session 2 and the narrative experience itself acquired between sessions) in both languages, the opposite behavior was detected in Olivia.In her case, the narratives are shorter in both languages in the second session when compared to the first.Another opposite trend observed between the two participants is that while Yves' vocabulary richness decreases after experience/practice, the opposite was observed in Olivia, who showed more vocabulary variation in the second session than in the first.As pointed out by an anonymous reviewer, the most feasible explanation for this opposite trend between participants is that while Olivia is more precise when explaining events (higher vocabulary richness and lower extension in session 2), Yves uses more repeated words (higher extension and lower vocabulary richness in session 2).This "repetition strategy" suggests a poor expressive vocabulary, a limitation that previous studies focused on the expressive language of monolingual speakers with PWS have found to a greater or lesser degree (Edmonston 1982;Kleppe et al. 1990; Van Borsel et al. 2007).

Holistic Narrative Performance
Overall, the results reflect a poorer narrative performance by Yves when compared to Olivia (see Figure 1).While Olivia obtained a global score higher than 50 (maximum score 60) in the four narratives, Yves scored between 31.5 and 35.75 in three narratives (with the exception of the story Frog, Where Are You? (Mayer 1969) narrated in English, where he obtained a score of 46.25 out of 60).

Global Narrative Performance by Dimensions
Figures 2 and 3 show, respectively, Olivia's and Yves' global performance for the three analyzed dimensions in each of the four narratives.The findings indicate a high and homogeneous command of the three analyzed dimensions in the case of Olivia, as all her scores, regardless of the narrative and the language, were above 15 (maximum score was 20).Yves, on the other hand, showed specific shortcomings.In fact, as Figure 3 reveals, the most problematic areas for the PWS participant were the ones related to the narrative process and complexity and the narrative structure and coherence dimensions since, with the exception of the story Frog, Where Are You? (Mayer 1969) in English, Yves' scores did not exceed 11 points, which is indicative of a poor command of these two dimensions.However, in the case of the narrative content and multiplicity dimension, a slightly better performance was observed as his scores were higher, ranging from 12.75 to 17.25.In what follows, each dimension is analyzed in detail for both Yves' and Olivia's performance.

Global Narrative Performance by Dimensions
Figures 2 and 3 show, respectively, Olivia's and Yves' global performance for the three analyzed dimensions in each of the four narratives.The findings indicate a high and homogeneous command of the three analyzed dimensions in the case of Olivia, as all her scores, regardless of the narrative and the language, were above 15 (maximum score was 20).Yves, on the other hand, showed specific shortcomings.In fact, as Figure 3 reveals, the most problematic areas for the PWS participant were the ones related to the narrative process and complexity and the narrative structure and coherence dimensions since, with the exception of the story Frog, Where Are You? (Mayer 1969) in English, Yves' scores did not exceed 11 points, which is indicative of a poor command of these two dimensions.However, in the case of the narrative content and multiplicity dimension, a slightly better performance was observed as his scores were higher, ranging from 12.75 to 17.25.In what follows, each dimension is analyzed in detail for both Yves' and Olivia's performance.
shortcomings.In fact, as Figure 3 reveals, the most problematic areas for the PWS participant were the ones related to the narrative process and complexity and the narrative structure and coherence dimensions since, with the exception of the story Frog, Where Are You? (Mayer 1969) in English, Yves' scores did not exceed 11 points, which is indicative of a poor command of these two dimensions.However, in the case of the narrative content and multiplicity dimension, a slightly better performance was observed as his scores were higher, ranging from 12.75 to 17.25.In what follows, each dimension is analyzed in detail for both Yves' and Olivia's performance.A boy, a dog and a frog (English) 1 (Mayer 1967) A boy, a dog and a frog (Spanish) 2 (Mayer 1967) Frog, where are you?(Spanish) 1 (Mayer 1969) Frog, where are you?(English) 2 (Mayer 1969)

Structure and coherence
Process and complexity Content and multiplicity

Narrative Structure and Coherence Dimension: Analytic Analysis
Overall, Olivia performed within the high range (between 4 and 5 points) in all the narrative structure and coherence subdimensions (see Figure 4).Yves, on the other hand, performed within the average-low range (between 2 and 3 points) for all the subdimensions (except in the case of the story Frog, Where Are You? (Mayer 1969) in English, in which he always obtained higher scores as shown in Figure 5).I will delve into this issue with more detail in the discussion section.Globally, it is noteworthy that Yves' lower scores were obtained for the evaluative commitment subdimension, which is indicative of an inability, in this specific case, to get emotionally involved in the narratives.A boy, a dog and a frog (English) 1 (Mayer 1967) A boy, a dog and a frog (Spanish) 2 (Mayer 1967) Frog, where are you?(Spanish) 1 (Mayer 1969) Frog, where are you?(English) 2 (Mayer 1969)

Structure and coherence
Process and complexity Content and multiplicity A boy, a dog and a frog (English) 1 (Mayer 1967) A boy, a dog and a frog (Spanish) 2 (Mayer 1967) Frog, where are you?(Spanish) 1 (Mayer 1969) Frog, where are you?(English) 2 (Mayer 1969)

Orientation
Structural Sequence Evaluative Commitment Integration

Narrative Structure and Coherence Dimension: Analytic Analysis
Overall, Olivia performed within the high range (between 4 and 5 points) in all the narrative structure and coherence subdimensions (see Figure 4).Yves, on the other hand, performed within the average-low range (between 2 and 3 points) for all the subdimensions (except in the case of the story Frog, Where Are You? (Mayer 1969) in English, in which he always obtained higher scores as shown in Figure 5).I will delve into this issue with more detail in the discussion section.Globally, it is noteworthy that Yves' lower scores were obtained for the evaluative commitment subdimension, which is indicative of an inability, in this specific case, to get emotionally involved in the narratives.

Narrative Structure and Coherence Dimension: Analytic Analysis
Overall, Olivia performed within the high range (between 4 and 5 points) in all the narrative structure and coherence subdimensions (see Figure 4).Yves, on the other hand, performed within the average-low range (between 2 and 3 points) for all the subdimensions (except in the case of the story Frog, Where Are You? (Mayer 1969) in English, in which he always obtained higher scores as shown in Figure 5).I will delve into this issue with more detail in the discussion section.Globally, it is noteworthy that Yves' lower scores were obtained for the evaluative commitment subdimension, which is indicative of an inability, in this specific case, to get emotionally involved in the narratives.A boy, a dog and a frog (English) 1 (Mayer 1967) A boy, a dog and a frog (Spanish) 2 (Mayer 1967) Frog, where are you?(Spanish) 1 (Mayer 1969) Frog, where are you?(English) 2 (Mayer 1969) A boy, a dog and a frog (English) 1 (Mayer 1967) A boy, a dog and a frog (Spanish) 2 (Mayer 1967) Frog, where are you?(Spanish) 1 (Mayer 1969) Frog, where are you?(English) 2 (Mayer 1969)

Orientation
Structural Sequence Evaluative Commitment Integration

Narrative Process and Complexity Dimension: Analytic Analysis
When focusing on the narrative process and complexity dimension, again, we observe that Olivia performed in the high range for all the subdimensions under analysis (see Figure 6), while Yves, with the exception of the narrative Frog, Where Are You? (Mayer 1969) in English, performed within the average-low range (see Figure 7).In general, Yves' lower scores were obtained when making reference to the cognitive experiences of the different characters (cognitive subjectifying) and when metaphorizing, which is linked to the ability to make inferences.

Narrative Process and Complexity Dimension: Analytic Analysis
When focusing on the narrative process and complexity dimension, again, we observe that Olivia performed in the high range for all the subdimensions under analysis (see Figure 6), while Yves, with the exception of the narrative Frog, Where Are You? (Mayer 1969) in English, performed within the average-low range (see Figure 7).In general, Yves' lower scores were obtained when making reference to the cognitive experiences of the different characters (cognitive subjectifying) and when metaphorizing, which is linked to the ability to make inferences.

Narrative Content and Multiplicity Dimension: Analytic Analysis
In the narrative content and multiplicity dimension, Olivia performed again within the high range for almost all the subdimensions, since the range of her mean scores were between 3.75 and (see Figure 8).In Yves' case, it is striking that, unlike in the case of the previous dimensions, his scores were, in general, above the 3-point line (see Figure 9), which means at least an average command of the four subdimensions.A more in-depth analysis shows that the specification of characters and A boy, a dog and a frog (Spanish) 2 (Mayer 1967) Frog, where are you?(Spanish) 1 (Mayer 1969) Frog, where are you?(English) 2 (Mayer 1969)

Narrative Process and Complexity Dimension: Analytic Analysis
When focusing on the narrative process and complexity dimension, again, we observe that Olivia performed in the high range for all the subdimensions under analysis (see Figure 6), while Yves, with the exception of the narrative Frog, Where Are You? (Mayer 1969) in English, performed within the average-low range (see Figure 7).In general, Yves' lower scores were obtained when making reference to the cognitive experiences of the different characters (cognitive subjectifying) and when metaphorizing, which is linked to the ability to make inferences.In the narrative content and multiplicity dimension, Olivia performed again within the high range for almost all the subdimensions, since the range of her mean scores were between 3.75 and (see Figure 8).In Yves' case, it is striking that, unlike in the case of the previous dimensions, his scores were, in general, above the 3-point line (see Figure 9), which means at least an average command of A boy, a dog and a frog (Spanish) 2 (Mayer 1967) Frog, where are you?(Spanish) 1 (Mayer 1969) Frog, where are you?(English) 2 (Mayer 1969)

Narrative Content and Multiplicity Dimension: Analytic Analysis
In the narrative content and multiplicity dimension, Olivia performed again within the high range for almost all the subdimensions, since the range of her mean scores were between 3.75 and 5 (see Figure 8).In Yves' case, it is striking that, unlike in the case of the previous dimensions, his scores were, in general, above the 3-point line (see Figure 9), which means at least an average command of the four subdimensions.A more in-depth analysis shows that the specification of characters and events are two of Yves' strengths, since, for these two subdimensions, he performed within the average-high range in all the narratives.

Discussion
The main objective of this study was to analyze the narrative abilities of an English-Spanish bilingual adult with PWS in both languages in comparison to the narratives produced by a TD English-Spanish bilingual adult with a similar sociolinguistic background.Overall, the results showed a rich narrative capacity in both languages for the TD bilingual (Olivia) and a poor performance for the individual with PWS (Yves).Olivia consistently performed within the high range in the four narratives produced (scores higher than 50 out of 60) while Yves' scores reflected an average command (scores between 30-35 points in all the narratives, with the exception of the narrative Frog, Where Are You? (Mayer 1969) in English, in which he obtained 46.25 points).These results showed that Yves' narrative abilities, unlike Olivia's, were lower than what is considered typical development.Thus, this outcome is in line with previous findings from studies focused on the narrative abilities of monolinguals with PWS, which consistently demonstrated low narrative quality (Lewis et al. 2002;Garayzábal-Heinze et al. 2012).
When differences between narratives were analyzed, it needs to be noted that while Olivia showed homogeneous narrative abilities in the four narratives produced, a better narrative quality became clear for Yves the second time he narrated a story, regardless of the language variable.As A boy, a dog and a frog (Spanish) 2 (Mayer 1967) Frog, where are you?(Spanish) 1 (Mayer 1969) Frog, where are you?(English) 2 (Mayer 1969)

Themes
Events Scenarios Characters

Discussion
The main objective of this study was to analyze the narrative abilities of an English-Spanish bilingual adult with PWS in both languages in comparison to the narratives produced by a TD English-Spanish bilingual adult with a similar sociolinguistic background.Overall, the results showed a rich narrative capacity in both languages for the TD bilingual (Olivia) and a poor performance for the individual with PWS (Yves).Olivia consistently performed within the high range in the four narratives produced (scores higher than 50 out of 60) while Yves' scores reflected an average command (scores between 30-35 points in all the narratives, with the exception of the narrative Frog, Where Are You? (Mayer 1969) in English, in which he obtained 46.25 points).These results showed that Yves' narrative abilities, unlike Olivia's, were lower than what is considered typical development.Thus, this outcome is in line with previous findings from studies focused on the narrative abilities of monolinguals with PWS, which consistently demonstrated low narrative quality (Lewis et al. 2002;Garayzábal-Heinze et al. 2012).
When differences between narratives were analyzed, it needs to be noted that while Olivia showed homogeneous narrative abilities in the four narratives produced, a better narrative quality A boy, a dog and a frog (Spanish) 2 (Mayer 1967) Frog, where are you?(Spanish) 1 (Mayer 1969) Frog, where are you?(English) 2 (Mayer 1969)

Themes
Events Scenarios Characters

Discussion
The main objective of this study was to analyze the narrative abilities of an English-Spanish bilingual adult with PWS in both languages in comparison to the narratives produced by a TD English-Spanish bilingual adult with a similar sociolinguistic background.Overall, the results showed a rich narrative capacity in both languages for the TD bilingual (Olivia) and a poor performance for the individual with PWS (Yves).Olivia consistently performed within the high range in the four narratives produced (scores higher than 50 out of 60) while Yves' scores reflected an average command (scores between 30-35 points in all the narratives, with the exception of the narrative Frog, Where Are You? (Mayer 1969) in English, in which he obtained 46.25 points).These results showed that Yves' narrative abilities, unlike Olivia's, were lower than what is considered typical development.Thus, this outcome is in line with previous findings from studies focused on the narrative abilities of monolinguals with PWS, which consistently demonstrated low narrative quality (Lewis et al. 2002;Garayzábal-Heinze et al. 2012).
When differences between narratives were analyzed, it needs to be noted that while Olivia showed homogeneous narrative abilities in the four narratives produced, a better narrative quality became clear for Yves the second time he narrated a story, regardless of the language variable.As shown in Figure 1, the scores for the two stories presented (A Boy, a Dog and a Frog (Mayer 1967) and Frog, Where Are You? (Mayer 1969)) were higher during the second experimental session.This means that Yves' narrative abilities for A Boy, a Dog and a Frog (Mayer 1967) were slightly superior in Spanish (session 2) than in English (session 1), with a difference of 4.25 points.Similarly, a better performance for the narrative Frog, Where Are You? (Mayer 1969) was observed in English (session 2) than in Spanish (session 1), with a difference of 14.25 points between the narratives in the two languages.This second-time effect indicates that narrative abilities in the PWS individual seem to improve with experience/practice, a finding consistent, to a certain extent, with Lewis et al.'s (2002) results, where older individuals (more experienced) showed better narrative abilities than the youngest ones (less experienced).Also, these results suggest that Yves has a good long-term memory, as has been defended in previous literature on PWS individuals (Curfs et al. 1991;Dimitropoulos et al. 2013), and that this strength, together with the experience/practice gained between the experimental sessions, will potentially make him more confident the second time he performs the task.Consequently, this will lead him to produce, in general terms, richer narratives during the second experimental session within his intrinsic limitations.If we compare Yves' performance between languages and narratives, it is noteworthy that the difference between the English and Spanish narratives is higher when narrating the story Frog, Where Are You? (Mayer 1969).This 10-point difference between the two narratives in the two languages (and between the two sessions) seems to reaffirm the idea that narrative abilities principally improve due to the experience/practice gained, since the story Frog, Where Are You? (Mayer 1969) in English was the last one to be produced.Likewise, the data presented in Table 3 showed that the narratives produced in session two were not only richer in terms of narrative abilities, but also in terms of length, language development (MLUw), and vocabulary variation, although not in the same direction for the two participants in this study.While in session two Yves produced longer narratives and showed a greater MLUw but a less rich vocabulary, Olivia produced shorter narratives richer in terms of vocabulary variation.In light of these preliminary results, both Yves and Olivia showed comparable narrative abilities in Spanish and English.Therefore, bilingualism does not appear to have a negative effect on the development of this PWS bilingual's narrative abilities, a finding that supports Kay-Raining Bird et al.'s (2016) claim that bilingualism among non-TD individuals should not be discouraged.
I will now analyze the results in relation to the three dimensions under study: narrative structure and coherence, narrative process and complexity, and narrative content and multiplicity.
Before delving into each of these three dimensions individually, a global perspective on each one of them is presented.Figures 2 and 3 reflect, respectively, Olivia and Yves' global narrative performance for each dimension and for each narrative.The data showed a consistent narrative control by Olivia in the three analyzed dimensions and different levels of performance in Yves' case.Leaving the story Frog, Where Are You? (Mayer 1969) in English aside (as he probably showed a better command due to the experience/practice acquired), Yves' narrative abilities, in both languages, could be described according to the following scale: narrative content and multiplicity > narrative structure and coherence > narrative process and complexity.In other words, this PWS bilingual showed better narrative abilities when expressing content than when expressing coherence or complexity.Likewise, coherence abilities were slightly superior to the ones attested to express process and complexity.Due to the fact that this is a case study, results should be considered as a first approximation to the analysis of narrative abilities in bilingual speakers with PWS and should, therefore, be interpreted cautiously.However, interestingly enough, Garayzábal-Heinze et al. ( 2012) also obtained the gradation scale presented above when analyzing the narrative's abilities of two Spanish monolingual adolescents with PWS (even though they do not discuss this in their paper).These similar outcomes seem to suggest a tendency that should be confirmed or disproved in future studies including larger samples of monolinguals and bilinguals with PWS.This must by no means be understood as an underestimation of the validity of case study research, whose "power" is often identified as wanting in comparison to large group-based research.Under no circumstances should both approximations be considered opposed but complementary and even necessary in order to have a complete and thorough picture of the phenomenon under analysis.While it is true that large samples and their corresponding statistical analyses provide objective validation to the study, case study research represents, in Flyvbjerg's (2006, p. 228) words, the "force of the example," which is not as valued as it should be taking into account that "more discoveries have arisen from intense observation than from statistics applied to large groups" (Beveridge 1951, quoted here from Flyvbjerg 2006, p. 226).
The following are the results obtained for the three dimensions and their corresponding subdimensions.In the narrative content and multiplicity dimension, Yves performed in the average-high range in all four subdimensions (themes, events, scenarios, and characters) in the four narratives (see Figure 9).This means that Yves' abilities to introduce and change topics (themes), to present different actions (events), to settle the scenario where the narrative takes place (scenarios), and to introduce and involve different characters to the story (characters) were not as impaired as other narrative abilities evaluated in this study.However, if Yves' results are compared to Olivia's performance, some shortcomings are patent.Yves' relative strengths-establishing a sequence of events and incorporating and involving different characters-are illustrated in Example 1 (see Appendixs A and B for complete narrative examples of both participants for the picture book Frog, Where Are You? (Mayer 1969) in Spanish).

Yves' Narrative 7
English Translation basis that an event is "a narrative sequence that allows the subject to answer the fundamental narrative question 'what happened?'"(Gonçalves et al. 2001c, p. 6) and that the number of events included in a narrative production determines its quality, among other indicators, it could be said that in the narrative of example 1, Yves showed a relatively rich sequence of events within his narrative limitations, which is translatable to the other narratives he produced.From line 1 to line 22 he presented an event per line, with the exception of line 11, where he reproduced what the child in the story was saying and did not introduce a new action.Similarly, in almost every line, he introduced a new character that was rapidly incorporated into the narrated action.All of the animal characters that were presented played a role-albeit brief-and on no occasion were unfoundedly introduced.These outcomes partially coincide with Garayzábal-Heinze et al.'s (2012) results for monolingual speakers with PWS.On the one hand, both Garayzábal-Heinze et al.'s (2012) and this study reflect an acceptable performance in the themes and events subdimensions.On the other hand, the studies show contradictory results for characters.While Garayzábal-Heinze and collaborators found low scores in monolingual speakers, the present study showed high performance for the bilingual speaker.It is premature to attribute this discordant outcome to a possible bilingualism effect given the limited data handled, but it is an issue that should be studied in depth in future research in order to elucidate whether being bilingual could play a role in this respect.Garayzábal-Heinze et al. (2012, p. 58) attributed the low scores observed in monolingual speakers with PWS when introducing variety of characters to "their behaviour problems, namely aggressiveness, stubbornness, emotional lability, and difficulties in adopting the other point of view."When interpreting Yves' results, we could hypothesize that bilinguals with PWS may show a more social and empathic profile than their monolingual counterparts because of their constant alternation between two languages and two cultures.However, it is worth noting that Yves is passionate about animals and has a high level of "expertise" in different animal species and their characteristics.Consequently, given the fact that the characters in the Frog story series are mainly animals, this could have played a role when it comes to Yves' high scores, since, as previous studies have noted, PWS individuals tend to have obsessions (Cassidy 1997;State et al. 1999).Thus, the alternative hypothesis that Yves' special interest in animals could have "helped" him to enrich his narrative because these characters particularly caught his attention cannot be ruled out at this point.
In future research, participants should be presented with stories including either human or non-human characters in order to neutralize a possible character bias.Focusing on the narrative structure and coherence dimension, Yves, unlike Olivia, performed in the average-low range in all four subdimensions under analysis (see Figure 5) with the exception of the story Frog, Where Are You? (Mayer 1969) in English.This means that handling orientation (contextualization of the story to make it understandable), structural sequence (organization and sequentialization of the story), evaluative commitment (emotional participation of the narrator), and integration (narrating the story in an integrated way) when narrating was hard for Yves.In fact, it seemed to be especially hard for him to get emotionally involved.Example 2 allows us to exemplify Yves' narrative shortcomings in relation to the narrative structure and coherence dimension (see Appendixs C and D for complete narrative examples of both participants for the text A Boy, a Dog and a Frog (Mayer 1967) in English).
When Yves and Olivia's fragments are compared, the reader will quickly notice the different degree of narrative elaboration and, consequently, the narrative abilities of the two participants.Both fragments reflect the same story passage and it is evident that Olivia's narrative is almost twice as long as Yves', a relevant issue that revealed the position each one took with respect to what information was presented and how.Focusing on orientation, it is worth noting that Olivia clearly situates the reader at the beginning of the story by creating the picture of a boy and his dog being ready to enjoy a fishing day out (lines 1 and 2 of Example 2).Opening by presenting and situating the characters establishes the starting point of the story and situates the recipient of the narration, who, with the information given, will be able to recreate such a scenario in their mind and will be able to model it according to the successive information.On the other hand, Yves did not include an opening description to contextualize the story but started by entering directly into one of the first events included in the story.This makes it very difficult for an interlocutor (who cannot see the picture book and is listening to the story for the first time) to make a proper mental representation of the scenario, since they do not possess the minimal necessary information to recreate it.Likewise, whereas in Olivia's narrative the listener constantly receives explicit and descriptive information that helps them to construct an accurate visual image of the different situations described (e.g., on top of a lily pad in line 3, uprooted root in line 5 or they fall head first in line 8 of Example 2), in Yves' case these clues are practically non-existent, since he concentrated mainly on narrating what happened, without going into greater detail.From a structural sequence point of view, it is important to highlight that Yves' narration is limited to a sequence of events not linked at all by discourse markers (e.g., from line 3 to 14 in Example 1, and from line 3 to 6 in Example 2), or, if linked, by using basic copulative (y 'and') (e.g., lines 15 and 22 in example 1), adversative (pero 'but') (e.g., line 2 in example 1), or consecutive conjunctions (así que 'so') (e.g., line 2 Example 2).For the sake of fairness, and although it is true that Olivia's narratives did not include a great deal of discourse markers either, it must be pointed out that her level of description and the effective way she connected the information made the recipient feel that they were listening to a detailed and linked-up story (an impression that was not present while listening to Yves' narratives, as he did not integrate the information in a very successful way).
Yves' non-TD-narration-style had a direct impact on his narrative structure and coherence quality evaluation, which translated into low scores for orientation, structural sequence, and integration variables (evaluative commitment will be discussed further).As argued in the previous paragraph, the structure and coherence of Yves' narratives were impaired and required the recipient's predisposition to give the narrative a complete meaning.According to Garayzábal-Heinze et al. (2012), when analyzing the narrative structure and coherence dimension of monolinguals with PWS, they found that PWS individuals were able to identify only the beginning of the story but not the development or ending.The results of this study, however, do not seem to support this claim.According to the data obtained here, the development of the story would be less of a problem for this PWS bilingual than providing an introduction and a closing to the story, since his real challenge seems to lie in the latter outcomes.Yves did not offer evident or strong closure to his narratives (see lines 21 and 22 of Example 1), nor a proper introduction (see lines 1 and 2 of example 2).This seems to favor this PWS individual being described as meticulous and analytical rather than having a holistic or global perspective when narrating.Another interesting issue is Yves' low emotional involvement or commitment in his narratives.Whereas Olivia tended to get emotionally involved in her stories by including evaluation adverbs (see lines 11 and 17 of example 2) and even dared to explain the story A Boy, a Dog and a Frog (Mayer 1967) in Spanish in the first person as if she were the child in the story, Yves adopted a neutral position towards the characters and limited himself to explaining what happened without getting emotionally involved with either the characters or the facts.
When evaluating Yves and Olivia's narrative process and complexity dimension, four subdimensions were analyzed: objectifying (inclusion of sensorial elements), emotional subjectifying (inclusion of emotional states), cognitive subjectifying (cognitive experiences of the characters) and metaphorizing (inclusion of metaphors).As stated before, this was the most impaired dimension in Yves' narratives and, furthering the idea from the previous paragraph, he seemed not to allow his emotions to "interfere" and play an active role in his narrative process.Thus, his narratives were comparatively less complex and diverse than Olivia's.The next step should be to explore in more depth this dimension in future research in order to elucidate whether, within their intrinsic limitations, bilinguals with PWS outperform monolingual speakers in story structure complexity, as Tsimpli et al. (2016) and Baldimtsi et al. (2016) found for bilinguals with SLI and high-functioning autism spectrum disorder, respectively.Overall, Yves scored in the low range in all subdimensions (see Figure 7), with the exception of the emotional subjectifying variable.The only exception to this was in Frog, Where Are You? (Mayer 1969) in English, where he obtained better results in all four dimensions.As a rule, he did not include metaphors and did not provide the reader with a detailed description of the sensorial elements or the characters' thoughts.Nevertheless, he did include some emotional states of the characters (lines 13 and 15 in Example 2), although to a lesser extent and with less detail than Olivia did (see lines 20-23 in Example 2).One possible explanation for the low performance in this dimension could be the idiosyncratic social profile of individuals with PWS, which will condition them to focus primarily on non-social elements (Garayzábal-Heinze et al. 2012).Assuming that this PWS bilingual is more social and empathetic than the PWS monolinguals included in Garayzábal-Heinze et al.'s (2012) study, this more socially oriented attitude would allow him to "socialize" the narratives by including more characters but not delving into more introspective details.Obviously, this possible explanation, as well as the interpretation of other findings and the findings reported in this study, need to be taken cautiously and should be confirmed with more extensive data.

Conclusions and Implications for Further Research
To the best of my knowledge, this study is the first to discuss the narrative abilities of an English-Spanish bilingual with PWS in comparison to a TD English-Spanish bilingual.Each participant produced four narratives (two in English and two in Spanish) over two experimental sessions and these narratives were analyzed according to three dimensions: narrative structure and coherence, narrative process and complexity, and narrative content and multiplicity (Gonçalves et al. 2001a(Gonçalves et al. , 2001b(Gonçalves et al. , 2001c)).The results obtained revealed that the bilingual with PWS did not show TD narrative behavior, which is in line with previous studies focusing on the narrative abilities of monolinguals with this syndrome (Lewis et al. 2002;Garayzábal-Heinze et al. 2012).
Another interesting result was that both participants showed comparable narrative abilities in both languages.Thus, a possible interpretation of these results is that bilingualism does not seem to have a negative impact on the development of the narrative abilities of the PWS individual's dominant language, since he is capable of narrating equally in both English and Spanish with the limitations expected from his condition.Therefore, the results from this study do not support the extended practice of denying individuals with PWS access to a second language.Actually, our findings support Kay-Raining Bird et al.'s (2016) defense of promoting bilingualism among non-TD individuals.The limited research carried out within this field has not only proved the ability of these individuals to become bilingual, but has also shown that the second language does not have a negative effect on the development of the first or dominant language.Given that nowadays there exists a tacit agreement among language acquisition researchers to think of bilingualism as an asset rather than a problem, the natural question to be asked is why non-TD individuals, overall, still continue to be denied access to a second language when in fact this access should be facilitated.In future research, it is my intention to test whether, as is the case with TD individuals, bilingual with PWS outperform their monolingual counterparts in different linguistic and non-linguistic abilities.
Another important result was that both participants performed better during the second session of data collection, as is evidenced by their richer narratives.However, this improved performance did not go in the same direction for both participants.While Olivia showed better lexical richness in session two than in session one, Yves exhibited more extended narratives in session two, a greater MLU, and a better control of the narrative abilities in the three analyzed dimensions.This leads to another very interesting finding, which is that while Olivia showed a consistent control of the narrative abilities in the three analyzed dimensions, Yves showed a different level of skillfulness according to the following scale: narrative content and multiplicity > narrative structure and coherence > narrative process and complexity.These results seem to suggest that narrative abilities, which have been revealed as a problematic area for PWS, seem to improve with experience/practice.Therefore, the curricula focused on this population should be carefully oriented towards the development of these abilities in order to make them more effective and, as a consequence, be able to produce "richer" narratives in any or all of the languages they can speak.Research based on the study of bilingualism in non-TD individuals contributes novel data to the field but, more importantly, provides intervention professionals with valuable information that will help them to make informed decisions on how to deal with bilingualism in non-TD bilinguals, as is the case with the PWS bilingual in this study.

Figure 1 .
Figure 1.Holistic scores for Olivia and Yves' narrative performance.

Figure 1 .
Figure 1.Holistic scores for Olivia and Yves' narrative performance.

Figure 4 .
Figure 4. Olivia's scores for the narrative structure and coherence dimension.

Figure 4 .
Figure 4. Olivia's scores for the narrative structure and coherence dimension.

Figure 4 .
Figure 4. Olivia's scores for the narrative structure and coherence dimension.

Figure 4 .
Figure 4. Olivia's scores for the narrative structure and coherence dimension.

Figure 5 .
Figure 5. Yves' scores for the narrative structure and coherence dimension.

Figure 5 .
Figure 5. Yves' scores for the narrative structure and coherence dimension.

Figure 6 .
Figure 6.Olivia's scores for the narrative process and complexity dimension.

Figure 7 .
Figure 7. Yves' scores for the narrative process and complexity dimension.

Figure 6 .Figure 5 .
Figure 6.Olivia's scores for the narrative process and complexity dimension.

Figure 6 .
Figure 6.Olivia's scores for the narrative process and complexity dimension.

Figure 7 .
Figure 7. Yves' scores for the narrative process and complexity dimension.

Figure 8 .
Figure 8. Olivia's scores for the narrative content and multiplicity dimension.

Figure 9 .
Figure 9. Yves' scores for the narrative content and multiplicity dimension.

Figure 8 .Figure 8 .
Figure 8. Olivia's scores for the narrative content and multiplicity dimension.

Figure 9 .
Figure 9. Yves' scores for the narrative content and multiplicity dimension.

Figure 9 .
Figure 9. Yves' scores for the narrative content and multiplicity dimension.

Funding:
This research was funded by the Ontario Graduate Scholarship program from both the Government of Ontario and the University of Ottawa, by the Excellence Scholarship from the University of Ottawa and by the Doctoral Research Scholarship program from the Fonds de Recherche du Québec-Société et Culture (FRQSC) (Scholarship number: 206102).

•
Research question 1: Does an English-Spanish bilingual with PWS show poor narrative abilities in terms of narrative structure and coherence, narrative process and complexity, and narrative content and multiplicity as previous research has revealed in the case of monolinguals with PWS?Does the bilingual speaker show relative strength in the narrative content and multiplicity dimension compared to the other two dimensions, as Garayzábal-Heinze et al. (2012) found for monolingual Spanish speakers with PWS?
• Research question 2: Are the narrative abilities of an English-Spanish bilingual with PWS comparable in both languages?•

Table 2 .
Cloze and proficiency tests' results for Yves and Olivia.

Table 3 .
Length, mean length of utterances by words (MLUw), and Type-token ratio values for each participant and narrative.