Diglossic aphasia and the adaptation of the Bilingual ...

Diglossic aphasia and the adaptation of the Bilingual Aphasia Test to Palestinian Arabic and Modern Standard Arabic Reem Khamis Dakwara,∗, May Ahmarb, Rola Farahc, Karen Froudd a

Adelphi University, Garden City, NY, United States b

Columbia University, New York, United States c Technion Institute, Haifa, Israel d

Columbia University Teachers College, New York, United States

ARTICLEINFO Keywords: Diglossia Bilingual Aphasia Test Declarative and procedural memory Implicit/explicit dichotomy Differntial recovery Cultural and linguistic sensitivity Modern Standard Arabic Palestinian Arabic

ABSTRACT The Bilingual Aphasia Test (BAT) is a criterion-referenced test developed in multiple languages and languagepairs to identify differential recovery in bilingual aphasia. It was developed to allow equivalent and culturally non-biased examination of all languages spoken by bi- or multilingual individuals with aphasia, to enable valid comparison of the residual abilities in each of the languages spoken. This comparison is critical for clinical practice to inform clinical decisionmaking, especially with respect to language of intervention and progress evaluation for bilingual adults with aphasia. This paper describes the adaptation of the BAT to expand its use to examine recovery patterns in adults with aphasia within an Arabic diglossic situation – specifically, Modern Standard Arabic (MSA) and Palestinian Arabic (PA). In diglossia, two language varieties co-exist, one acquired naturally and used for daily communication, and the other learned formally in school, used for formal communication settings, reading and writing and typically considered to be of a higher value by its community members. The theoretical, clinical and social impacts of this unique adaptation of the BAT to a diglossic situation are discussed.

1. Introduction At present, there are no systematic studies of aphasia investigating recovery patterns in diglossic situations. Diglossia refers to a specific sociolinguistic situation in which two closely inter-related varieties of the same language are used for different social and communicative functions. One variety is acquired naturally and used for daily communication (often referred to as a Low Language Variety) and the other is a standard language variety that is explicitly taught and used for written texts and other formal contexts (referred to as a High Language Variety; Ferguson, 1959). Such diglossic situations necessitate differential engagement of implicit memory systems in the acquisition of the native dialect acquired naturally, in contrast with greater engagement of explicit memory mechanisms and metalinguistic knowledge that are used to acquire the standard language variety. One example of a diglossic situation is found in Arabic, with various spoken dialects that are specific to each speech community, and Modern Standard Arabic (MSA) acting as the formal (or “High”) language variety. MSA has long been considered as a unifying variety across the Arabicspeaking communities, but Ibrahim (2009) showed that there may be regional variation in MSA too; e.g., Cairene MSA in Egypt, or Damascene MSA in Syria. Paradis (1994, 2004) has proposed that in aphasia, implicit linguistic competence is affected, with explicit linguistic knowledge

∗

Corresponding author. Department of Communication Sciences and Disorders, Adelphi University, 154 Cambridge Avenue, Hy Weinberg Center, Room 117, Garden City,

NY 11560. USA. E-mail address: [email protected] (R. Khamis Dakwar). https://doi.org/10.1016/j.jneuroling.2018.04.013

R. Khamis Dakwar et al. Received 28 September 2017; Received in revised form 30 January 2018; Accepted 26 April 2018 Available online 09 May 2018 0911-6044/ © 2018 Elsevier Ltd. All rights reserved.

Journal of Neurolinguistics 47 (2018) 131–144

potentially affected to a much lesser extent – a view of acquired language pathology that relates directly to the distinction between declarative and procedural memory systems (e.g. Cohen & Squire, 1980; Squire, 1992). This approach to aphasia has been widely applied to the study and treatment of bilingual aphasia. For instance, observations of selective language loss in bilingual aphasia could be erroneously interpreted as cases of differential recovery, when a more accurate characterization could relate to the inhibition of implicit linguistic knowledge (correlated with L1) alongside a reliance on explicit metalinguistic knowledge (correlated with L2, if the L2 was formally learned) as a compensatory strategy. Under this framework it is clear that there is a pressing need to assess all languages spoken premorbidly by an adult with aphasia, using assessment instruments that are linguistically and culturally equivalent in each of the patient's languages (Paradis, 2004). Although many researchers and practitioners have addressed bilingual aphasia from this and other theoretical perspectives, little work has been done to relate Paradis’ model to diglossic aphasia. Examination of aphasic recovery patterns in diglossia necessitates the development of an assessment protocol that is culturally unbiased and enables equivalent comparison between residual language skills in both language varieties. The Bilingual Aphasia Test (BAT) was designed to enable identification of differential impairment in classic bilingual situations by assessing each of the languages of a bilingual or multilingual individual with aphasia in an equivalent way, across all linguistic domains, in both oral and written modes (Paradis, 2011). This widely used assessment tool has been adapted to several world languages and language pairs. However, the use of such a tool for assessing all language varieties in a diglossic situation has not been initiated. Such investigation is of potential benefit, not only for clinical practice with multilingual communities, but also for enhancing understanding of understudied and under-examined language varieties (typically the spoken dialects). Such knowledge would contribute to theoretical discussions related to language recovery in bilingual and diglossic speech communities, and to understanding of the different learning mechanisms associated with distinct language varieties. In this paper, we describe the adaptation of the BAT to two language varieties in Arabic diglossia: the Palestinian Arabic (PA) spoken dialect and Modern Standard Arabic (MSA). This adaptation will enable the assessment of residual linguistic abilities in each of the two language varieties used by Palestinian adults with aphasia. This is the first aphasia test designed specifically to address a diglossic situation, and it has the potential to enable a systematic investigation of recovery patterns in diglossic aphasia. 2. Arabic diglossia, the declarative/procedural model, and recovery patterns in aphasia Arabic is characterized as having multiple levels of variation (Holes, 2004) ranging from syntactic, phonological and lexical variation across different spoken dialects of Arabic to the sociolinguistic variation related to the presence of diglossia. In general, diglossic variation refers to the sociolinguistically governed alternation between two varieties of a language used for different pragmatic functions, one of which is acquired naturally while the other is learned formally. Holes (2004) argues that this description of Arabic diglossia is rather oversimplified, because of constant stylistic shifting between the language varieties, a situation referred to as “kaleidoscopic variation” (Holes, 2004: 49). Moreover, other foreign languages are also part of the linguistic norm in some parts of the Arab world, such as the use of French in Lebanon and Algeria or the use of Hebrew by Palestinians in Israel, as a remnant of the colonization of these areas. The different languages and varieties spoken in Arabic speech communities necessarily interact with the speech-language pathology services provided for those with communication disorders. Nevertheless, the newly emerging literature in this field is more focused on the examination of “classic” second language(s) within a speaker's repertoire and tends not to address the different varieties used by speakers in Arabic or other diglossic situation (e.g. African American English speakers in the U.S. or Cypriot Greek speakers in Cyprus; for further discussion, see KhamisDakwar, 2018, in press). In Arabic, for example, a recent study by Mamdoh and Gomaa (2015) examined dysfluencies in the productions of middleschool Egyptian Arabic speaking children who stutter in their native language – the spoken dialect (Egyptian Arabic in this case) – and in English, which was learned as a second language by all participants at age 4. However, their investigation omitted any examination of dysfluency patterns in MSA, which would be expected to form part of the language repertoire for middle school children in Egypt. Hence, they did not examine the influence of diglossia on the communication disorder. Similarly, studies of aphasia recovery in Arabic speaking individuals have compared recovery patterns across two “classic” languages such as English and Arabic (e.g., Faroqi-Shah & Waked, 2010; Vajramani, Akrawi, McCarthy & Gray, 2008), Hebrew and Arabic (e.g., Khamis, VenkertOlenik, & Gil, 1996) or have examined cross-linguistic treatment generalization between Arabic and English for Arab-English bilingual speakers (e.g. Koumanidi Knoph, 2013). By omitting evaluation of the different varieties of Arabic, these pioneering investigations, despite addressing a gap in the bilingual aphasia research, did not include examinations of all languages and language varieties that make up the linguistic repertoire of Arabic speakers. Adopting an ecologically valid crosslinguistic approach to assessment has the potential to provide clinicians and/or researchers with more comprehensive clinical information that may inform best clinical practices as well as theoretical knowledge. Indeed, Jones, Gitterman, and Obler (2012) provided a case study on the manifestation of agrammatism in a diglossic native speaker of African-American English and Standard American English, and in doing so they highlighted the need to further examine bilingual recovery in diglossia and to develop culturally appropriate assessments that are equivalent across diglossic language varieties, to provide adequate services for individuals with aphasia and/or other communication disorders in diglossic speech communities. Drawing upon the dichotomy of explicit-implicit learning mechanisms that are, by hypothesis, involved in acquiring the two language varieties in diglossic situations, studies of language recovery in diglossic aphasia may inform current theoretical approaches in this domain. For example, the declarative/procedural model of memory draws a distinction between two neurofunctionally and anatomically distinct memory systems that subserve implicit and explicit linguistic knowledge (Squire, 1992). Implicit linguistic knowledge, typically construed as the rule-governed aspects of linguistic knowledge, is “acquired incidentally, stored implicitly, used automatically” (Paradis, 2008:343), and its instantiation in the brain is known to involve circuits extending to parts of the right cerebellum, left neostriatum, and the frontal basal ganglia as well as portions of the

132

R. Khamis Dakwar et al.


perisylvian cortex (Paradis, 2008).1 Explicit linguistic knowledge, such as knowledge of words and their meanings, is sustained by declarative memory, and must be consciously learned; it involves the hippocampal system, mesial temporal lobes, parahippocampal gyri, and anterior cingular cortex (Paradis, 2008). This kind of knowledge is impacted by pathologies affecting memory retrieval, such as Alzheimer's disease (Paradis, 2008). These competencies instantiated in distinct cerebral systems lie behind observations of some paradoxical recovery patterns in aphasia, and also account for the double disassociation between aphasia and amnesia (Paradis, 1994). While both implicit and explicit systems are needed in language processing and representation, it may be the case (Paradis, 2004) that there are distinctions – especially in early second language learning – between a greater reliance on explicit, metalinguistic knowledge for an L2 compared to implicit, procedural knowledge that characterizes an L1. Hence, in bilingual aphasia, an implicit/explicit memory model may help to explain observations of enhanced performance in a patient's second language when knowledge of their first language is compromised – depending on how (or when) the second language was learned. Bilingual aphasia in this way contrasts with certain symptoms of Alzheimer's disease, where greater impairments of explicit memory are observed, while implicit knowledge like knowledge of the L1 is preserved (Paradis, 2008). These differences in language recovery in bilingual aphasia, i.e. better recovery of an explicitly-learned second language (which corresponds to the high language variety in diglossia), may appear to be a differential recovery, when in fact the implicit competence of both languages may be equally impaired but patients may rely on their L2's declarative knowledge to compensate for their lack of implicit, automatic, procedural language competence (Paradis, 2004). Furthermore, variability between morphosyntactic and semantic aphasic deficits within the same language may be related to the interaction between brain damage affecting implicit syntactic knowledge and explicit lexical semantic knowledge. Indeed, Khamis-Dakwar and Froud (2012) predicted the presence of differential recovery in Arabic diglossia due to the interplay between explicit and implicit learning mechanisms involved in acquiring the two language varieties of Arabic. This theoretical framework can be directly evaluated through the examination of aphasic recovery in diglossia; but this is only possible using an assessment that enables direct comparison between the different diglossic language varieties. This highlights the need to expand the use of the BAT to include diglossic language varieties, and not only to classic cases of bilingualism. 3. The bilingual aphasia test (BAT) The BAT has three parts. The first part is a questionnaire that can be completed either by the patient or by family members, to provide background information on the patient's history of multilingualism. There are fifty questions about language use, exposure, and linguistic beliefs and preferences of the adult with aphasia. The second part of the BAT systematically assesses the patient's languages, in all domains and across different modalities, through 472 test items in each language. The third part assesses translation abilities and grammaticality judgments across each language pair, providing the opportunity to investigate the interactions between the languages spoken by the patient. The BAT has previously been adapted to several Arabic dialects, including Jordanian (Paradis, Hallis, Taha, & Amayreh, n.d.) and Tunisian (Paradis, Abidi, & Maymouni, n.d.), but neither of these adaptations compared the dialect with MSA,2 despite the existence of diglossia in the targeted speech communities. By contrast, the adaptation reported in this paper rendered the BAT into two varieties of Arabic (comparable to different languages in more “classical” bilingual situations). The test was normed as per BAT guidelines (Paradis & Libben, 1987). For the purposes of developing this assessment, it was assumed, broadly speaking, that the native dialect (Palestinian Arabic, PA) is supported by implicit linguistic knowledge, while MSA is supported by declarative memory mechanisms; and that aphasia interferes with the normal functioning of the procedural memory systems that underlie implicit linguistic competence (correlated with knowledge of the native dialect). It is therefore predicted that, in adult speakers of Palestinian Arabic who have aphasia, knowledge of PA would be affected, while MSA is likely to remain available as a compensatory mechanism. Based on these assumptions, we predict that future administration of the BAT in the two language varieties would permit identification of differential recovery patterns in Arabic diglossia. 4. Aim of the work Given that best practices in the assessment of bilingual individuals with aphasia should incorporate examination of all languages using a valid and reliable instrument that is linguistically and culturally equivalent in each of the patient's linguistic systems, along with the predicted presence of persistent impairments determined by diglossic language representation and processing, we aim to describe the adaptation of the BAT to the Palestinian and Modern Standard Arabic varieties. We hope that this approach will catalyze further adaptations of the BAT to other Arabic-speaking communities exhibiting diglossia, and to non-Arabic diglossic communities that exemplify similar sociolinguistic situations. Diglossia is generally an understudied phenomenon, and this has affected the quality of the educational and clinical services provided to individuals from diglossic communities. This is the first explicit attempt to adapt the BAT for two Arabic language varieties. This paper describes the adaptation and norming of BAT to diglossic varieties as a first step towards future investigations of the predicted presence of diverse patterns of recovery in Arabic diglossia, based on differences in contexts of acquisition and use of the two language varieties. By developing the BAT for the two language varieties, it is possible to examine differential recovery in multilingual aphasia while controlling for structural differences between the two language varieties.

1 Paradis' perspective resonates with Ullman's declarative procedural model (Ullman, 2004) but these two perspectives differ in the specifics of the distinctions between declarative and procedural memory, and each incorporates divergent neuroanatomical bases for declarative memory (Ullman, 2006). 2 The available Arabic BAT language pairs are Arabic-English, Arabic-Armenian, Arabic-French, Arabic-Somali, and Arabic-Swahili. There are no explicit specifications of the Arabic language variety used in these adaptations, and a review of the available Arabic BATs reveals that MSA was used in some of the adaptations (e.g. Arabic-English) while regional dialects were used in others (e.g. Arabic-French).

133



We will first introduce diglossia with a focus on the structural differences between Palestinian Arabic and Modern Standard Arabic, the two language varieties used within the Palestinian speech community. This will be followed by a detailed description of the adaptation and norming of the BAT to these two language varieties, and a discussion of the social and clinical implications of the adapted BAT for Arabic diglossia. 5. Arabic diglossia and contrastive features of Palestinian Arabic and modern standard Arabic Diglossia was introduced to modern linguistic studies by Ferguson (1959), who observed a number of speech communities that exhibit this phenomenon: Colloquial and Classical Arabic, Dhimotiki and Katharevousa Greek, Creole and Standard French in Haiti, and Standard German and Swiss-German in parts of Switzerland. Other speech communities were later also identified as exhibiting diglossia, notably African American communities where African-American English (AAE) constitutes the native language and Standard American English (SAE) is used for formal communication, reading and writing (e.g., Ogbu, 1999; Birch, 2001; but note that the majority of discussions of AAE do not employ the term diglossia explicitly). Ferguson (1959: 336) defined diglossia as: “a relatively stable language situation in which, in addition to the primary dialects of the language (which may include a standard or regional standards), there is a very divergent, highly codified (often grammatically more complex) superposed variety, the vehicle of a large and respected body of literature, either of an earlier period or in another speech community, which is learned largely by formal education and is used for most written and formal spoken purposes but is not used by any sector of the community for ordinary conversation.” Ferguson referred to the two language varieties as “H” (the high, or superposed, variety) and “L” (the low, or naturally acquired, variety). Diglossia is not a case of bilingualism. In an epilogue written several decades after his original characterization of diglossia, Ferguson (1996) reflected on the original definition of the term, and underscored the problems inherent in extending it to bilingual situations. He reiterated that diglossia exists, by definition, in monolingual societies in which two or more varieties of the same language are used. “Diglossia” as a term was not intended to describe any functional distribution of two languages associated with different speech communities. Ferguson (1996) also highlighted the different learning mechanisms associated with each of the language varieties – that is, the H variety is learned explicitly through formal education while the L variety is acquired implicitly as a native language. Ferguson commented that his definition of diglossia was intended to capture “the kind of situation in which the ordinary formal language of the community is one that no one speaks without special effort and no one uses in ordinary conversations: it is acquisitionally and functionally superposed to the primary variety of the language” (Ferguson, 1996: 52). 5.1. Contrastive features between MSA and Palestinian spoken dialects In Arabic, structural differences between the spoken dialect and MSA are evident in all language domains. For the sake of this paper, we will provide a general linguistic description of selected differences between Palestinian Arabic and MSA, in order to illustrate the selected reversible features used for the development of part C of the PA-MSA BAT. 5.1.1. Phonology Ibrahim (1983) maintains that the phonological systems of all spoken dialects have some overlap and some differences when compared to the phonological system of MSA. In this section, we will outline the documented differences between Palestinian and Standard systems in the different domains, while acknowledging that there are more shared than non-shared features, as in all other domains. There are a number of vocalic differences between MSA and dialects. In Palestinian Arabic, unstressed long vowels in final positions are shortened (Raz, 1996), and the diphthongs/aʊ/and/aI/are monophthongized to/oː/and/eː/respectively. These vocalic differences underlie syllabic structural differences. For instance, MSA includes light syllables (CV, CVː, CVC) in addition to CVːC and CVCC structures in word-final positions, whereas Palestinian dialects allow heavier syllables including 2-consonant clusters in wordinitial positions or across morphemic boundaries – but not in final positions (Saiegh-Haddad & Henkin-Roitfarb, 2014). Stress allocation also varies between MSA and the different Arabic dialects (Saiegh-Haddad & HenkinRoitfarb, 2014). Moreover, the phonological inventory of MSA includes a large consonantal inventory composed of 28 conventional ‘consonants’ and a relatively small vowel inventory that includes the short and long vowels (/i/,/iː/,/a/,/aː/,/u/,/uː/) and two diphthongs (/aɪ/and/aʊ/). Biadsy, Hirschberg, and Habash (2009) outlined a number of phonological differences between Arabic dialects and MSA, including differences in the realization of various MSA consonants in different Palestinian spoken dialects. For instance, the phoneme which in MSA is a voiceless uvular stop/q/is usually cognate with the glottal stop phoneme/ʔ/in urban Palestinian Arabic. However, differences between MSA and the dialect vary across different dialects. For example, while the MSA uvular stop/q/is absent in the dialect of Acre and realized as/ʔ/, this difference is not evidenced in rural and nomadic Palestinian dialects such as the dialects of Baːqa or the Naqab (Majadly, 2012). In other cases, such as the Palestinian dialect spoken in Kufur-Kana, different phoneme, the voiced velar/ɡ/is the local cognate of MSA/q/. 5.1.2. Lexical semantics According to Ferguson (1959), lexical differences between language varieties are a defining feature of diglossia, with some lexical items being shared between both language varieties but others instantiating variation, such that the use of one lexical item may be indicative of the selected variety. There is limited information on the percentage of lexical items that are shared between MSA and Palestinian spoken Arabic, and how many items comprise lexical cognates. In general, for Ferguson, the lexicons of the two language varieties in diglossia may share lexemes, or include cognates with partial or no overlap. Computational studies attempting to parse Arabic dialect corpora based on pre-existing MSA resources provide some insight into the lexical distance between MSA and different Arabic dialects, but comparable lexical resources in the dialects are still absent (for a review see Farghaly & Shaalan, 2009). Maamouri et al. (2006) relate the absence of MSA-dialect dictionaries to the fact that the two diglossic

134



language varieties are perceived to constitute one language by speakers and hence dictionaries are usually produced for non-Arabic translations, not for dialect-MSA equivalents. 5.1.3. Morphosyntax Morphosyntactic differences between MSA and the spoken dialects of Arabic are visible at all levels of word and phrase structure. The word order VSO is unmarked in MSA, while it is marked in PA, with SVO being the norm. In PA, relatively free alternation between the two word orders is permissible, but in MSA they are associated with distinct agreement restrictions (Shlonsky, 1997). Case marking is not exhibited in spoken dialects of Arabic, but is seen in MSA productions. Few studies have examined linguistic competence in diglossia, so it is difficult to determine which structures should be targeted in assessment of aphasia, while controlling for age of acquisition effects. Khamis-Dakwar, Gordon and Froud (2012) examined the development of children's diglossic morphosyntactic knowledge using a forced choice grammaticality task. They looked at six morphosyntactic features that are known to be instantiated differently in MSA and Palestinian Arabic. The study revealed that diglossic knowledge related to plural marking, dual marking, wh-questions, and construct structures was acquired early compared to relative pronouns, passives, yes/no question, and adjective definiteness. These findings therefore guided the selection of morphosyntactic structures used in part C of the PA-MSA BAT. 6. Adaptation of the BAT to PA and MSA In this section, we will discuss the adaptation of the BAT to Palestinian Arabic and Modern Standard Arabic based on the guidelines published by Paradis and Libben (1987), which provide an outline of the different sections of the BAT, how to select items for the assessment, and methods for addressing crosslinguistic considerations. The Palestinian dialect adaptation is based on the Galilee sub-dialect used by Palestinians living in the North of Israel and in coastal urban areas (such as Haifa and Acre). However, we still refer to this adaptation as a general Palestinian BAT, rather than relating the adaptation to a specific sub-dialect. This is because while we acknowledge that Palestinian Arabic has different variations within it, the administration of the test should not be affected since the administrator should be familiar with the patient's dialect as per the guidelines of BAT. Moreover, we made sure that all stems used are exhibited similarly in all Palestinian dialects (so long as the differences are in pronunciation, and not in the morphosyntactic features selected for testing, this presents no problem). In translating the name of the test we decided to borrow the term “aphasia” for Arabic use and not “locked of tongue” (maħbuːs allisaːn), the phrase usually used to describe aphasia impairment in both Arabic language varieties. The use of this physical term rather than linguistic or cognitive terms to describe aphasia has implications for descriptions of recovery and treatment, and contributes to the construction of experiences during aphasia recovery, especially as it impacts caregivers’ perceptions and the roles of family members (Kardosh & Damico, 2009). Hence, we decided to maintain the term “aphasia” as an attempt to raise awareness and change the perspective about how aphasia and its rehabilitation are viewed in Arabic-speaking communities. The use of English loan words for medical terms is acceptable in Arabic speaking communities, and we believe the use of the word aphasia rather than the Arabic term also reflects current changes in perception of communication disorders within Arabic-speaking communities. 6.1. Part A: history of bilingualism This section of the BAT was directly translated into Modern Standard Arabic and Palestinian Arabic. However, since part A of the BAT was developed for classic bilingual cases and not for cases of diglossia, the questionnaire as it stands does not address issues affecting exposure and use of different varieties in diglossia, such as the knowledge of spoken Arabic and MSA in Arabic-speaking communities. Additionally, there are also misconceptions about Arabic among its speakers, who might assume when asked about Arabic that there is no difference between the formal and dialectal varieties. Hence, to inquire about how and when the different varieties of Arabic were acquired – a question that carries valuable clinical import, due to the implicit-explicit dichotomy, and a factor that may differ between individuals, especially for Arabic-speaking individuals in the diaspora – a sub-item was added to item 4: “what about MSA?” We decided to add a sub-item rather than a question, to adhere to BAT adaptation guidelines (that prohibit the addition of new questions, so as to ensure that all items are in the same order in all versions of the BAT). 6.2. Part B: instructions All part B instructions were translated to MSA. For the Palestinian BAT, administrators were instructed to use their spoken dialect in presenting these instructions. This decision was made due to the fact that it is generally not acceptable to write in spoken Arabic, although some poets and other writers make use of this practice. Additionally, dialectal phonology is not necessarily always reflected Table 1 PA and MSA words used for word translation task. Item Number

PA

MSA

English translation

23 24 25 26 27 28 29 30 31

‫ شنطة‬antˤaʃ ‫ قشاط‬ʔʃaːtˤ ‫كلسات‬kalsaːt ‫مصاري‬masˤari ‫ سطل‬satˤəl ‫ طاقية‬tˤaʔijje ‫شاكوش‬akuːʃ ʃ ّ‫شمسية‬amsijje ʃ ‫ قرقعة‬qurqaʕa

‫حقيبة‬aqiːba ħ ‫ حزام‬ħizaːm ‫جوارب‬awaːrib ʤ ‫مال‬maːl ‫ دلو‬dalw ‫ قبعة‬qubaʕa ‫ مطرقة‬mitˤraqa ‫ مظلة‬miðˤala ‫ سلحفاه‬sulħufaː

bag belt socks money bucket hat hammer umbrella turtle

135

R. Khamis Dakwar et al. 32

‫ تخت‬taxit

‫ سرير‬sariːr

Journal of Neurolinguistics 47 (2018) 131–144 bed

in the Arabic/MSA alphabet, not to mention the many different subvarieties of dialect that might be involved. Below we describe essential changes made within Part B and the reasoning for these changes. Any other items that are not described below were translated word by word to the targeted variety and did not require any additional crosslinguistic and/or crosscultural considerations. All item numbers relate to the standard BAT configuration as laid out by Paradis and Libben (1987). 6.2.1. Part B, items 10–17 (personal literacy history) Since Arabic dialects do not have written systems in diglossic situations, questions 10–17 are not applicable for the Palestinian BAT Language Background Part B. As such, these items were left empty (with stars next to each item) and participants are not asked for Palestinian dialect responses in the current revision of the BAT. This permits maintenance of the item numbering sequence as per BAT adaptation guidelines. 6.2.2. Part B, items 23–33 (word translation) Due to the different levels of lexical distance between MSA and its translational items in the Palestinian spoken dialect, direct translation of these ten items would have resulted in a mixture of MSA-PA non-cognates, partial cognates, and full cognates. To make sure that there is no overlap at all between words in MSA and their PA counterparts, all these items were changed so as to include only MSA-PA non-cognates, as shown in Table 1 below: 6.2.3. Part B, items 33–42 (simple and semi-complex commands) Similarly, for this task, lexical distance between MSA and PA was controlled and the commands had to be changed. For the last 5 commands, the cognates from the pointing sub-test were used whereas the first five commands were translated word by word and included the following: - two MSA-PA cognates that do not share phonological features: /sakkir/ (PA) - /ʔaƔliq/ (MSA) for ‘close-IMP’ and tˤalliʕ/ (PA) /ʔaxriʤ/ (MSA) (‘raise-IMP’); - two similar lexemes: /ʔiftaħ/ for ‘open-IMP’ and /ʔirfaʕ/ for ‘lift-IMP’; - and six items that were related by phonological features such as: /ʔi:d/ (PA) - /jad/ (MSA) for ‘hand’; /tumm/ (PA) and /fam/ (MSA) for ‘mouth.’ To ensure that the administrator presented each command in the relevant language variety, a clear direction was added for each item. The complex commands (items 42–47) were translated word by word into each language variety. 6.2.4. Part B, items 48–65 (verbal auditory discrimination) Considerations of dialectal variability and cultural specificity were considered for this adaptation of the BAT. Several items used in previous Arabic BAT adaptations were utilized for the verbal auditory discrimination subtest, but not all, due to dialectal considerations that affected the status of minimal pairs. For example, in the Jordanian BAT, the following minimal pairs are used: /θoːm/ (garlic), /koːm/ (bunch), /noːm/ (sleeping), and /ʕoːm/ (floating). In the Palestinian dialect, the word for ‘garlic’ is pronounced /tuːmi/ and hence it was changed to /toːm/ for ‘twins.’ Moreover, the term /ʕoːm/ is less frequently used for ‘floating’ and hence it was switched to /joːm/ ‘day.’ Pictures were altered accordingly as shown in Fig. 1. In addition to the linguistic considerations, cultural considerations were taken into account when determining the minimal pairs used in this section. For example, in the Jordanian BAT the following minimal pairs were presented: /lafː/ (turned), /kafː/ (palm), /zafː/ (attending wedding procession). Two of these items were changed for the Palestinian BAT, because the pictures used to exemplify turning were based on a boat steering wheel turning and the picture of the wedding procession was not culturally appropriate. Hence, these two words were changed to /sˤafː/ (classroom) and /rafː/ (shelf), as shown in Fig. 2 below. 6.2.5. Part B, items 66–152 (syntactic comprehension) For the assessment of syntactic comprehension, VSO was determined to represent canonical sentential word order in PA, and SVO

136



Fig. 1. Pictures used for verbal auditory discrimination subtest (item 58) in Palestinian and Jordanian Arabic.

Fig. 2. Pictures used for verbal auditory discrimination subtest (item 64) in Palestinian and Jordanian Arabic.

in MSA. Since the passive construction is distinct between MSA and PA (specifically, the “by” phrase is not available in Arabic dialects), the use of relative pronouns was substituted for passive as a non-standard (marked) construction in our adaptation (sentences 81–88). For example, sentence 81:

Sentence 87:

137



The second non-standard structure selected was topicalization of causal sentences (instead of topicalization in anaphora/reference constructions, as is available in English and Jordanian). The non-standard negative was maintained. For the reversible sentence constructions (items 137–152) which include possessive realized through possessive s-marker ’s and through construct nouns (as in “the manager of this building”), there should be two optional structures to express the same meaning (e.g., both analytic and synthetic possessives: the baby's mother/the mother of the baby). Since construct nouns in Arabic are not contrastive with another type, unlike English, we used masculine and feminine idˤaːfa (construct state) adding a masculine or feminine demonstrative - for example, alternating between Sentence 148:

And Sentence 140: This permitted the representation of reversible noun-phrase constructions in addition to the possessive, and is similar to the approach used in the French version of the BAT (and several others involving languages that do not instantiate the ’s/noun phrase alternation). Moreover, since spoken Arabic draws a distinction between animate and inanimate nouns, we changed the target nouns in part 6. The Jordanian BAT used the nouns car (sayara) and truck (shahina), but in MSA both of these would be inanimate feminine. We therefore chose car (sayara) (feminine) and bus (bas) (masculine).

6.2.6. Part B, items 153–172 (semantics) The items selected to evaluate semantic categories, synonyms and antonyms (section I) were all translated word by word to Palestinian dialect and MSA for the current BAT adaptation. In the antonyms section II (items 168–172), three derived forms of the antonym of the main word are presented for the patient to choose from. In dialectal Arabic, adverbs are usually produced as a preposition followed by a noun form (such as /bis-surʕa/ in-the-hurry, translated as quickly). For these items, a past tense verb form was substituted for the adverb form, similar to the adaptation of BAT to German. For example, in Item 170 in English, the patient is asked to identify the antonym for sad and is provided with the following choices: the adjective happy, the noun happiness, and the comparative adjective happier. For the adapted Palestinian BAT item 170, the patient is asked to identify the antonym of /zaʕlaːn/ ‘sad’, and presented with the noun /bastˤ/ ‘happiness,’ the adjective /mabsuːt/ ‘happy’, and the past tense /nbasˤat/ ‘he was happy’.

6.2.7. Part B, items 193–252 (repetition and judgment) Items for assessing lexical decision and the repetition of words and pseudowords were based on the adaptation guidelines and adjusted so that they conformed to the phonotactic rules of the targeted Arabic variety.

6.2.8. Part B, items 260–262 (series) In the section requiring enumeration of a series (260–262), the choice was made to change item #260 which requests the production of the months of the year. This was done in order to minimize possible effects of influence from Hebrew. Instead, we required the enumeration of the days of the week. Items 261 and 262, requesting counting from 1 to 25 and naming the letters of the alphabet, were not changed.

6.2.9. Part B, items 263–268 (verbal fluency) In the verbal fluency subtest, the instruction is typically given to list as many words starting with a certain letter/sound. For the third required sound, we substituted the sound /k/ with an /m/ because of the limited number of words that start with /k/ in Arabic (for a review see Newman, 2005, pp. 185–206). Although this means that the task will rely to a certain extent on pattern productivity (similar to derivational morphology), in addition to phonological productivity, the /m/ does provide more possibilities for noun productions than the very infrequent initial phoneme /k/.

6.2.10. Part B, items 367–427 (reading aloud, reading comprehension, copying, writing to dictation and spontaneous writing) Since PA and MSA do not share an equal sociolinguistic status and only MSA is recognized as a written system within the Arabic diglossic situation, all reading and dictation tasks were eliminated in the assessment of PA. This was done in order to avoid the inclusion of ecologically invalid tasks. In order to maintain the task numbering sequence and its consistency with other BAT adaptations, the items for these tasks were left empty (with stars next to each item). 6.3. Part C: language-pair specific items 6.3.1. Part C, items 428–457 (word recognition) Similar to the adaptation of part B, for the word recognition task in part C, MSA-PA non-cognates were selected. The assessment manual suggests oral presentation of words when a language has no written form, or when the participant is illiterate (Paradis & Libben, 1987). Hence, we 138



provided the written MSA words for the translational task as a list of PA words that were transcribed in the MSA alphabet, despite the fact that PA is not customarily used for writing in Arabic diglossia. 6.3.2. Part C, items 458–480 (translation and grammaticality) The following reversible contrastive features were used for creating the translation and grammaticality items for this task. 1. Translation of one contrastive feature: The dual verb conjugation evident was the reversible contrast used for the first sentence in this section. This construction is available in MSA productions but absent in Palestinian Arabic. For the second sentence, we used consecutive verb conjugation, which is evidenced in dialects but is not possible in MSA. For instance, in PA two consecutive verbs can appear with no interfering particles in between (e.g., biddi ashrab, literally “I-want I-drink”); while MSA requires the use of a particle (e.g., uriid Ɂan ashtarii, “I want to buy”). 2. Translation of two contrastive features: Sentences 3 and 4 in this section (items 430 and 431) require translation of two contrastive features. One sentence was formed based on the MSA and PA differences in tense marking. The spoken dialect uses past tense in conditional sentences whereas MSA uses present tense. This distinction was combined with differences in the future marker used in MSA/sa/versus PA /raħ/. The other sentence utilized the reversible feature related to the marker/n/in construct structures, which is present in PA but absent in MSA (e.g., /muhandisiː ʃ-ʃarika/ (MSA) versus /mhandsiːn i-ʃ-ʃirke/ (engineerpl.mas the-company) (PA), translated to English in both cases as “the company's engineers”). This was combined in the sentence with differences in MSA and PA pronominals (/ʔanna/, MSA and /ʔinno/, PA).3 3. Translation of three contrastive features: The fifth and sixth sentences in this section (items 468 and 469) contain three reversible features. For example, item 469 uses double negation (permitted in PA but not in MSA) in addition to a masculine plural adjective for the feminine plural noun (permitted in PA but not MSA), and the use of feminine plural marking on the verb (required in MSA but absent in PA). This example of a 3-contrast item is illustrated below. MSA

A comprehensive list of all the reversible contrastive features that were used for the development of the MSA-PA BAT Part C is provided in the Appendix to this paper. 7. Collecting normative data on the BAT for Arabic diglossia in the Palestinian community Sixty healthy Palestinian Arabic-speaking individuals (30 men and 30 women) participated in the normative data collection process for the MSA-PA BAT adaptation. Participants were recruited by the third author from the general population and outpatient clinics, and they had no known cognitive impairment and no stroke or neurological deficits. Participants were divided into three age groups: 20 participants (10 male) between the ages 50 and 59 years; 20 participants (10 male) between the ages of 60 and 69 years; and 20 participants (10 male) over age 70 years (age and years of schooling for the whole standardization group are summarized in Table 1 below). All participants were right handed native speakers of Palestinian Arabic from the northern area of Israel and all provided informed consent to participate in the normative data collection study. All study procedures were overseen by the Table 2 Normative data collected for Part B: numbers of errors produced. Part B: Errors produced by participants broken down by subtest and age group Item range

3

Subtest name'

N of items

N of errors – 50–59 years old


N of errors – 70 + years old

N of errors – Normal range (Paradis & Libben, 1987)

In PA, the relative particle/Ɂanna/, equivalent to the English relative pronoun/that/, has to morphologically include an attached pronoun –o, becoming/Ɂinno/, regardless of

whether or not the subject of the sentence is present. This is prohibited in MSA. 139

R. Khamis Dakwar et al. Spontaneous speech 18–22

Journal of Neurolinguistics 47 (2018) 131–144 *

5

0

0

0

Pointing Simple and semi-complex commands Complex commands

10 10

0 0

0 0

0 0

0

5

0

0

1

2**

48–65 66–152 153–157 158–162 163–167 168–172 173–182 183–192 193–252 253–259 260–262 263–268

Verbal auditory discrimination Syntactic comprehension Semantic categories Synonyms Antonyms I Antonyms II Grammaticality judgment Semantic acceptability Repetition and judgment Sentence repetition Series

1 3 0 0 0 0 0 0 0 0 0

2 4 0 0 0 0 0 0 0 0 0

1 5 0 0 0 1 0 1 0 0 0

3

Verbal fluency

18 87 5 5 5 5 10 10 30 7 3 6

269–288 289–313 314–323 324–333 334–343 344–346

Naming Sentence construction Semantic opposites Derivational morphology Morphological opposites Description

20 25 10 10 10 3

0 0 0 0 0

0 0 0 1 0

0 0 0 1 1

0 *** 1 2 2 *

347–361 362–366 367–376 377–386 387–392 393–397 398–402 403–407 408–417

Mental arithmetic Listening comprehension Reading (words) Reading (sentences) Reading (text) Copying Dictation (words) Dictation (sentences) Reading comprehension (words) Reading comprehension (sentences)

15 5 10 10 6 5 5 5 10

0 0 0 0 0 0 0 0 1

1 0 0 0 0 0 0 0 0

1 1 0 0 0 0 0 0 1

0–2 1 0 1 1 0 0 1 1

10

1

1

1

1

23–32 33–42 43–47

418–427

0–1

0–3 1 1 1 1 1 1 0–1 1 0 ***

Institutional Review Board at Teachers College, Columbia University. Testing took place in a quiet room, in the speech-language pathology clinic where participants had been recruited. Prior to starting the assessment, participants were informed about the test structure and were reminded to answer the questions in the language version being tested (MSA/PA). To ensure understanding of the instructions regarding MSA/PA use, participants were asked to judge whether a sentence produced by the examiner would be considered MSA (e.g. the girl sat on the bed /ʒalasat ʔaṭifla fawq ʔalsari:r/) or PA (e.g., we went out yesterday /ruћna ʔImbareћ miʃwar/) and to provide one word in MSA and another word in PA. The sequence of tasks and the objects and materials used were identical in the two versions tested. The order of PA or MSA presentation was counterbalanced across participants. All participants except for 5 completed the three parts of the BAT on the same day, with a few minutes break between parts. Testing always started with part “A”, followed by part “B”; half of the participants in each age group started with the MSA version and the other half started with the PA version. Part “C” concluded the assessment in all cases. Testing lasted about 2.5 h including breaks. For part A, all participants indicated the use of PA in their daily life as the spoken language of choice while reporting exposure to MSA on a daily basis through the media, TV shows, news and religious activities. Eighteen out of 20 participants in the 50-59 year-old age group reported exposure to and use of the Hebrew language and 4 out of the 18 reported exposure and use of English as well. All twenty participants in the 60-69 year-old group reported exposure and use of Hebrew, and 3 out of the 20 reported exposure and use of English as well. Finally, all twenty participants in the 70+ year-old age group reported exposure and use of Hebrew. Participant errors on each task in Part B are reported in Table 2, broken down by age group. Results also revealed that there was no significant difference in age between males (M = 64.3, SD = 8.8) and females (M = 64.83, SD = 7.8; t (58) = −0.247, p = 0.8), or in years of schooling (males: M = 10.6, SD = 2.8; females: M = 11.0, SD = 4; t (58) = −0.697, p = 0.48). For Part C, all participants completed the translation and grammaticality judgment tasks as described above. Numbers of errors for each task are provided in Table 3 below, broken down by age group. Table 3 Normative data collected for Part C: numbers of errors produced. Part C: Errors produced by participants broken down by subtest and age group Item range

Subtest name

N of items



N of errors – 70 + years old

428–432

Word recognition-MSA→PA

5

0

0

0

140

R. Khamis Dakwar et al. 433–437 Word recognition-PA→MSA 438–447 Translation of words MSA→PA

5

0

0

Journal of Neurolinguistics 47 (2018) 131–144 0

10

0

0

0

448–457 458–469

Translation of words PA→MSA Translation of sentences MSA→PA

10

0

0

0

12

1

1

1

470–481 482–497

Translation of sentences PA→MSA Grammaticality judgment MSA→PA

12

1

1

2

498–513

Grammaticality judgment PA→MSA

16 16

0 1

1 1

1 1

8. Potential recovery patterns for MSA and PA Recently, a few case studies of aphasia in Palestinian individuals have been published (i.e. Adam, 2014; 2016), but no reports of differential recovery in diglossia have been reported so far. We assume that this paucity is linked to the fact that diglossia is an understudied phenomenon in clinical research and that most current investigations of aphasia in Arabic are reporting on more commonly investigated questions (such as verb and noun processing, tense and inflection processing, or the Tree Pruning Hypothesis (TPH)), within the same variety used for communication (i.e., the spoken dialect). Such practices may in part be due to a limited awareness of potential diglossic recovery patterns in aphasia, as well as severely limited accessibility to appropriate assessments for diglossic aphasia. In turn, this reflects an Anglo-centric perspective on assessment and treatment studies in general. For example, in July 2017, the first author presented a webinar on principles and practices for aphasia assessment in Arabic diglossia using the Bilingual Aphasia Test; the audience was composed of Arab speech pathologists in Israel/Palestine working with adults with neurogenic disorders. During that session, one experienced practitioner who had worked for more than 5 years at a specialized hospital described a case of a patient who, post-stroke, spoke mainly in MSA. At the time, the speech pathologist could not understand this presentation; but following the webinar it was much clearer that apparent differential recovery in relation to diglossia could be interpreted differently. Based on such anecdotal evidence, we hope that availability of the developed BATs for PA and MSA will catalyze further case reports of apparent differential recovery in aphasia, and that increased awareness and knowledge of differential recovery patterns and their relationship to different learning mechanisms will be useful for clinical practice. Given that the standard variety is formally learned at school, to the extent that it contains morphosyntactic elements that differ from those of the naturally-acquired variety, and that those elements are learned, and not acquired, the model predicts that a bidialectal patient with aphasia will have metalinguistic knowledge available in the standard variety that can serve as a compensatory mechanism. Such knowledge is not available in the naturally-acquired variety and hence cannot compensate for the lack of access to implicit, procedural linguistic knowledge. Some tasks in the BAT assess mainly explicit or implicit kinds of knowledge, as shown in Table 4 below. Hence, some predictions are possible that can help to identify whether cases of apparent differential recovery in the standard variety could be related to this procedural/declarative distinction. The analysis of specific recovery patterns would be based on Paradis’ suggestion that knowledge of L1 phonology, morphology and syntax is supported by implicit procedural memory, while lexical knowledge is more explicit and declarative in nature (Paradis & Libben, 1987). As such, three possible recovery patterns are expected when administering the MSA and PA BAT: (1) the parallel recovery of Standard and spoken Arabic with relatively better performance in MSA for explicit tasks that are learned meta-linguistically (see Table 4); (2) apparent differential recovery of MSA, reflected in better performance in MSA for all tasks explicit and implicit; and (3) the differential recovery of PA, likely seen in individuals who had lower ability in MSA premorbidly. 9. Social and clinical implications This paper describes the adaptation of Bilingual Aphasia Test to the diglossic language pair, Modern Standard Arabic and the Palestinian spoken dialect. This addresses a gap in the field of bilingual assessment in aphasia, since the use of BAT to assess aphasia in bilingual/multilingual individuals does not yet address linguistic varieties within the sociolinguistic situation of diglossia. There is increased awareness in the field of speech-language pathology of the need to assess all languages spoken by a bilingual/multilingual individual with a communication disorder. Recent studies have documented the interrelationships between many different languages and language varieties in the manifested symptoms of aphasia – for example, the work of Jones et al. (2012) documenting the manifestation of African American English (AAE) features in the productions of an adult AAVE speaker with agrammatism. Hence, the expansion of BAT to assess language impairment across languages as well as varieties of languages within diglossia has social and clinical implications. Diglossia is exhibited in several communities across the world, and overlooking this aspect of the linguistic profile of patients from such speech communities means that their communicative needs are not being identified or met. Initiating systematic assessment of language impairment across all varieties in aphasia may enhance the quality of the clinical services provided to individuals with aphasia and their families from diglossic populations. This is true in particular in light of latest work on the effectiveness of utilizing Table 4 Characterization of key knowledge domains addressed by specific subtests in Part B. Subtest (item numbers) Part B

Main linguistic knowledge addressed

Spontaneous speech (18–22, 514–539)

All domain

Pointing (23–32)

Explicit knowledge (lexical)

Orders (33–47)

Implicit knowledge (semantic -lexical)

Verbal auditory discrimination (48–65)

Explicit knowledge (phonology -lexical)

Comprehension of syntactic structures (66–152)

Implicit knowledge (syntactic) 141

R. Khamis Dakwar et al. Semantic categories (153–157)

Journal of Neurolinguistics 47 (2018) 131–144 Explicit knowledge (semantic -lexical)

Synonyms (158–162)

Explicit knowledge (semantic -lexical)

Antonyms (163–172)


Grammaticality judgment (173–182)

Implicit knowledge (syntactic)

Semantic judgment (183–192)

Explicit knowledge (semantic)

Repetition of words (193–251)


Lexical decision (194–252)


Sentence repetition (253–259)


Series (260–262) Verbal Fluency (263–268)


Naming (269–288)


Sentence construction (289–313) Semantic opposites (314–323)


Derivational morphology (324–333)

Implicit knowledge (morphological)

Morphological opposites (334–343)

Implicit knowledge (morphological)

Description of a story (344–346, 504–565)

All domain, plus working memory

Mental arithmetic (347–361)

All domain, plus working memory and mathematical processing

Text listening comprehension (362–366)


Reading words aloud (367–376)


Reading sentences aloud (377–392)


Text reading comprehension (387–392)


Copying of words (393–397)


cognitive-linguistic strategies in treatment (see Kohnert & Peterson, 2012). It should be noted, of course, that the BAT is a criterionreferenced test that is most commonly used to evaluate linguistic and cognitive skills in two languages in a comparable way. Clinical decisions should not be made solely on the results from any single assessment approach, including the BAT, but will always require the skilled integration of information from assessments of functional communication and life participation. These assessments will be most effective when they are administered in ways that are appropriate for, and sensitive to the needs of, individuals from culturally and linguistically diverse (CLD) populations. The BAT approach to assessment, by including sections on translation and cross-linguistic judgment, may serve a vital role for clinical work with CLD populations, by supporting informed decision-making in the selection of such strategies in therapy. The BAT therefore supports the enhancement of functional communication in multilingual aphasia. In addition to the clinical contribution of BAT development in diglossia, such an undertaking has the potential to contribute to the theoretical discussion of recovery patterns in aphasia, especially within the declarative/procedural model. This is because based on that account, we can predict the presence of differential recovery patterns in diglossic situations. Hence, expansion of the study of language recovery is called for, especially in other diglossic situations (such as Swiss-German and German) and within Arabicspeaking communities with different levels of exposure and use of MSA.

10. Conclusions There is a lack of available language tests designed to assess all language varieties spoken by members of diglossic speech communities. Development of the BAT to assess all language varieties in diglossia would enable clinicians to equivalently compare the abilities in the two language varieties that are represented within a diglossic situation. This initiative would also enable researchers to examine differential recovery predicted by the dichotomy of explicit-implicit learning mechanisms in diglossia, and hence contribute to the theoretical literature. The two tests are accessible to all via the McGill University website (https://www. mcgill.ca/linguistics/research/bat#modern: the MSA BAT is available under the Modern Standard Arabic tab, and the PA BAT under Palestinian Arabic). The major implications of the initiative to extend the BAT to diglossic situations, however, remain clinical in nature, since functional communication and cultural participation in diglossic speech communities is inherently reliant on facility in both language varieties.

Declarations of interest None. Acknowledgments We would like to thank Dr. Michel Paradis for his continuous mentoring and support of our project. We would also like to thank Dr. Heather Green for her contributions to the project at its early stages, and Dr. Uri Horesh and Dr. Mira Goral for reviewing early drafts. Lastly, a special thanks to the participants in this study for their willingness to support the adaptation of the BAT to Modern Standard Arabic and Palestinian Arabic. Appendix A. Supplementary data

142



Supplementary data related to this article can be found at http://dx.doi.org/10.1016/j.jneuroling.2018.04.013. Appendix For the MSA-Palestinian Arabic BAT, the following reversible contrastive features were considered for the development of Part C. Note that reversible contrastive features are required to be obligatory in both languages (not optional in one) and must be equivalent, namely the use of feature A in Language B must be as unacceptable as the corresponding Feature B in Language A.

Drop the masculine plural noun marker “-n” in first part of idaafa construction (plural masculine noun + noun) → no drop of marker Noun + Noun 2. Masculine plural adjective for feminine plural noun → Feminine plural adjective for feminine plural noun 3. Stative non-habitual action mood: Active participle → Present tense verb from 4. No nominative case for subject nominal (accusative instead) → nominative case for subject nominal 5. double negation Neg + V + sh → Neg + V+ Ø 6. Masculine plural verb marking for feminine plural noun → Feminine plural verb marking for feminine plural noun 7. Dual noun + plural adjective → Dual noun + dual adjective 8. Yes/No question: Sentence (rising intonation) → Hal + Sentence 9. V + that + neutral pronoun → V + that+ Ø 10. Relative pronoun “illi” for both masculine and feminine antecedents →/llaði:/for masculine and/llati/for feminine 1.

11. 12. 13. 14. 15. 16. 17.

Present tense marker for masculine singular/plural/ba + stem/→ ʔu + stem Dual noun + Plural verb → Dual noun + Dual V nfa3al → Passive fu3ila If (iza)+ Present tense → If + past tense conjugation present tense masculine plural: V+ u: + Ø → V + u:n V (plural fem)+ Noun (Plural fem.) → V (masculine fem)+ Noun (Plural fem) V + V (Ø present marker → V + an + V)

18. raħ future marker → sa- future marker 19. Ø Plural feminine verb suffix (plural fem with plural masc suffix instead) → -na plural feminine suffix 20. Conjugation future tense feminine singular/plural: raħ + V → awfa + V

English Translations of Part C items (relevant contrasts are indicated after the first item in each pair)

458. I saw the two boys eating an apple (12) 460. I want to buy coffee in an hour (17) 462. If you study you will succeed (14, 18) 464. I heard that the company engineers are big people (1, 9) 466. Is your brother living in two big buildings? (7, 8, 3) 468. I don't allow small girls to play video games (5, 2, 6)

470. I saw two dogs eating a bone 472. We want to hear the sound of the car 474. If you move here, you will like the area 476. We knew that the school teachers are important 478. Does your mother like the two neighboring schools? 480. There are no colored teachers teaching here

Grammaticality judgment (relevant contrasts indicated in parentheses)

Bad MSA (except 486 & 492 = correct)

482. This is the ball the boy threw (10) 484. I am preparing coffee (11) 486. The men like cabbage (15) 488. The window was broken yesterday (13) 490. The neighboring women drank the juice (16) 492. She will buy a new dress (20) 494. The teachers are home (4)

Bad PA (except 500 & 504 = correct)

500. I am seeing a movie (11) 502. The workers listen to the ruling (15) 504. The sheep was slaughtered yesterday (13) 506. The lawyer women studied for the test (16) 508. They will celebrate the holiday (20) 510. I saw the carpenters in the carpenter's shop (4) 512. The mothers give the medicine to the kids (dative) (19) 143


496. The aunties are travelling to Hebron (19) 498. This is the Table 1 bought (10)


500. I am seeing a movie (11) 502. The workers listen to the ruling (15)

References Adam, H. (2014). Verb and noun production in aphasia: Evidence from Palestinian Arabic. International Journal of Brain and Cognitive Sciences 2014, 3(1), 1–5. Adam, H. (2016). Tense and gender production in Arabic-speaking aphasics. Journal of Foreign Language Teaching and Applied Linguistics, 3(1), 1–6. Biadsy, F., Hirschberg, J., & Habash, N. (2009). Spoken Arabic dialect identification using phonotactic modeling. Proceedings of the EACL 2009 workshop on computational approaches to semitic languages (pp. 53–61). Association for Computational Linguistics. http://www.aclweb.org/anthology/W09-0807.pdf. Birch, B. (2001). Grammar standards: It's all in your attitude. Language Arts, 78(6), 535–542. Cohen, N. J., & Squire, L. R. (1980). Preserved learning and retention of pattern-analyzing skill in amnesia: Dissociation of knowing how and knowing that. Science, 210(4466), 207–210. Farghaly, A., & Shaalan, K. (2009). Arabic natural language processing: Challenges and solutions. ACM Transactions on Asian Language Information Processing (TALIP), 8(14), 1–19. Faroqi-Shah, Y., & Waked, A. N. (2010). Grammatical category dissociation in multilingual aphasia. Cognitive Neuropsychology, 27(2), 181–203. http://dx.doi.org/10. 1080/02643294.2010.509340. Ferguson, C. (1959). Diglossia. Word, 15, 325–340. Ferguson, C. (1996). Epilogue: Diglossia revisited. In A. Elgibali (Ed.). Understanding Arabic: Essays in contemporary Arabic linguistics in honor of el-said badawi (pp. 49– 67). Cairo: Cairo Press. Holes, C. (2004). Modern Arabic: Structures, functions, and varieties. Washington, D.C.: Georgetown University Press. Ibrahim, M. (1983). Linguistic distance and literacy in Arabic. Journal of Pragmatics, 7(5), 507–515. Ibrahim, Z. (2009). Beyond lexical variation in Modern Standard Arabic. Cambridge Scholars Publishing. Jones, J., Gitterman, M., & Obler, L. K. (2012). A case study of a bidialectal (African-American vernacular English/Standard American English) speaker with agrammatism. In M. Gitterman, M. Goral, & L. Obler (Eds.). Multilingual aphasia (pp. 257–277). Bristol, UK: Multilingual Matters. Kardosh, B. M., & Damico, J. S. (2009). The contribution of language in shaping clinical culture: Palestinian aphasics and families living in Israel. Asia Pacific Journal of Speech, Language and Hearing, 12, 243–252. Khamis-Dakwar, R., & Froud, K. (2012). Aphasia, language, and culture: Arabs in the United States. In M. Gitterman, M. Goral, & L. Obler (Eds.). Multilingual aphasia (pp. 275– 291). Bristol, UK: Multilingual Matters. Khamis-Dakwar, R., Froud, K., & Gordon, P. (2012). Acquiring diglossia: Mutual influences of formal and colloquial Arabic on children's grammaticality judgments. Journal of Child Language, 39, 1–29. Khamis-Dakwar, R. (2018). Diglossia. To appear in Damico. In J.S., & M. J. Ball (Eds.). The SAGE encyclopedia of human communication sciences and disorders (in press). Khamis, R., Venkert- Olenik, D., & Gil, M. (1996). Bilingualism in aphasia: The effect of L2 treatment on language performance of L1. Journal of the Israeli Speech, Hearing & Language Association, 19, 73–83. Kohnert, K., & Peterson, M. (2012). Generalization in bilingual aphasia treatment. In M. Gitterman, M. Goral, & L. Obler (Eds.). Multilingual aphasia (pp. 89–105). Bristol, UK: Multilingual Matters. Koumanidi Knoph, M. (2013). Language intervention in Arabic-English bilingual aphasia: A case study. Aphasiology, 27(12), 1440–1458. Maamouri, M., Bies, A., Buckwalter, T., Diab, M., Habash, N., Rambow, O., et al. (2006). Developing and using a pilot dialectal Arabic treebank. Proceedings of LREC, genoa, Italy. Majadly, H. (2012). The Arabic dialect of Baqa al-Gharbiyya: Aspects of phonology and morphology. Qasimi College Collection, 6, 37–80. http://www.qsm.ac.il/ arblanguage/docs/majalla/6/5–eng–haifa.pdf. Mamdoh, H., & Gomaa, M. A. (2015). Assessment of severity of stuttering in native versus foreign language in secondary (late) bilingual children. Indian Journal of Otolaryngology and Head & Neck Surgery, 67(2), 132–134. Newman, D. L. (2005). Contrastive analysis of the segments of French and Arabic. In A. Elgibali (Ed.). Investigating Arabic: Current parameters in analysis and learning. Leiden, Netherlands: Brill. Ogbu, J. U. (1999). Beyond language: Ebonics, proper English, and identity in a Black-American speech community. American Educational Research Journal, 36(2), 147–184. Paradis, M. (1994). Neurolinguistic aspects of implicit and explicit memory: Implications for bilingualism. In N. Ellis (Ed.). Implicit and explicit learning of second language (pp. 393–419). (London). Paradis, M. (2004). A neurolinguistic theory of bilingualism. Vol. 18. John Benjamins Publishing. Paradis, M. (2008). Language and communication disorders in multilinguals. In B. Stemmer, & H. A. Whitaker (Eds.). Handbook of the neuroscience of language. Oxford: Elsevier Press. Paradis, M. (2011). Principles underlying the bilingual aphasia test (BAT) and its uses. Clinical Linguistics and Phonetics, 25(6–7), 427–443. Paradis M., Abidi R. and Maymouni Z., (n.d.). Bilingual aphasia test (Tunisian Arabic version), http://www.mcgill.ca/linguistics/research/bat. Paradis M., Hallis Y., Taha K. and Amayreh M., (n.d.). Bilingual aphasia test (Jordanian Arabic version), http://www.mcgill.ca/linguistics/research/bat. Paradis, M., & Libben, G. (1987). The assessment of bilingual aphasia. Hillsdale, NJ: Erlbaum. Raz, S. (1996). Prominence and vowel duration in some spoken Arabic dialects. Israel Oriental Studies, 16, 193–199. Saiegh-Haddad, E., & Henkin-Roitfarb, R. (2014). The structure of Arabic language and orthography. In E. Saeigh-Haddad, & M. Joshi (Eds.). Handbook of Arabic Literacy: Insights and perspectives (pp. 3–28). New York, NY: Springer. Shlonsky, U. (1997). Clause structure and word order in Hebrew and Arabic. New York: Oxford University Press. Squire, L. R. (1992). Declarative and nondeclarative memory: Multiple brain systems supporting learning and memory. Journal of Cognitive Neuroscience, 4(3), 232–243. Ullman, M. T. (2004). Contributions of memory circuits to language: The declarative/procedural model. Cognition, 92(1), 231–270. Ullman, M. T. (2006). The declarative/pocedural model and the shallow-structure hypothesis. Applied PsychoLinguistics, 27(1), 97–105 ([Commentary on target article by H. Clahsen and C. Felser.]). Vajramani, G., Akrawi, H., McCarthy, R., & Gray, W. (2008). Bilingual aphasia due to spontaneous acute subdural haematoma from a ruptured intracranial infectious aneurysm. Clinical Neurology and Neurosurgery, 110, 823–827.

144