Phonological aspects of al-Issa Arabic, a Bedouin dialect in the north of Jordan

Abstract

This study offers an account of selected key phonological aspects of al-Issa Arabic, spoken in three villages in al Mafraq: Ad-Dafyana, Mansiyat al-Gublan and Um As-srab. Forty suitable participants plus four language consultants were recruited. The dialect has a number of distinguishing features that make it stand out from the rest of Bedouin Jordanian Arabic dialects, e.g. the Bani Saxar dialect (Palva 1980), Bani Ḥassan dialect (Irshied 1984), Abbādi dialect (Sakarna 1999), and Wadi Ramm Arabic (Mashaqba 2015). These features include the unlimited affrication of ∗k, the realization of the palato-alveolar/j/ as [y] in all word positions (yaʔyaʔa phenomenon), and the unconditioned retention of Classical Arabic diphthongs ∗aw and ∗ay in all contexts. In terms of syllable structure, the core syllable types observed are: CV, CVC, CVV, CVVC, CCVVC, CVC₁C_1, and C₁C₂VVC. There is a strong ban against C₁C₂C₃ and C₁C₂ coda clusters in the dialect. Following Kiparsky's (2003) typological classification, the dialect is a VC-dialect where a vowel is inserted between C₁ and C₂ in C₁C₂C₃ clusters. To avoid gutturals in coda positions, the dialect exhibits of the gahawa syndrome where CVG.CV underlying forms are attested as CV.Ga.CV. Stress is predictable in the vast majority of forms and is controlled by a trochaic foot ('CV.CVC). Stress falls on the right-most heavy syllable in the last three syllables. Where a word lacks a heavy syllable, it falls on the antepenultimate, or a penultimate in two syllable words. However, stress assignment interacts with epenthesis and the gahawa syndrome: the inserted vowel attracts the main stress in the absence of an ultimate heavy syllable.

Arabic dialects, Jordanian Arabic, al-Issa Arabic, Bedouin dialects, phonology, stress.

1. Introduction

Arabic is the most widely spoken of the living Semitic languages, a group of languages belonging to the Afroasiatic family (Watson 2002). Arabic has two standard varieties: Classical Arabic (CA) and Modern Standard Arabic (MSA). CA, also referred to as fuṣḥā, is the standard language that was spoken in the pre-Islamic era by the Quraysh tribe. MSA, also referred to as contemporary faṣīḥ, is the form of language used in formal contexts, such as academic lectures, media, religious ceremonies, and conferences. Additionally, Arabic comprises a large number of dialects spoken in around 60 countries that differ from one another (and from MSA and CA) phonologically, morphologically, syntactically and lexically. These dialects may not be comprehensible to speakers from other regions (Huneety 2015; Watson 2002; Mashaqba et al., 2020a). Arabic dialects are classified geographically into: the Levant, Mesopotamia, the Arabian Peninsula, Egypt, and the Maghreb (Behnstedt and Woidich 2013). Another broad typological classification of Arabic dialects is that of Palva (2006): Western dialects Maġribī and Eastern dialects Mašriqī. The former group covers the dialects spoken in North Africa, whereas the latter dialect group is spoken in the Middle Eastern countries. Sociologically, Arabic varieties are divided into sedentary ħaḍarī and Bedouin badawī based on the history of settlement and language change (Palva 2006; Rosenhouse 2006).1

The Jordanian population is divided into nomadic and semi-nomadic tribes ‘Bedouins’, sedentary villagers fallaḥīn and town dwellers madaniyyīn (Dann 1984). Each of these groups has some distinguishing linguistic features, yet they are mutually recognized by speakers of Jordanian Arabic dialects (Suleiman 1985; Mashaqba et al., 2020b). Based on the realization of the imperfect prefix b- and CA realization of ∗q in the verb gāl ‘to say’, Cleveland (1963) offers a typology of JA dialects in four groups: yigūl, bəgūl, bəkūl and bəʔūl dialects. Similar to Najdi and al-Ḥsā dialects spoken in the north-west and east peninsula respectively, Group I yigūl dialects refer to Bedouin varieties in the southern and eastern desert of Jordan. In this group the reflex of ∗q is/g/, and verbs lack the imperfect indicative marker. Group II bəgūl dialects are spoken in sedentary trans-Jordanian dialects and in the West Bank and Jordan River; the reflex of ∗q in bəgūl dialects is also the voiced velar stop/g/. Although Group II dialects share most of Group I's features, they differ in the use of the imperfect bi- and the use of the negation morpheme –š. Group III bəkūl refers to central Palestinian dialects which realize ∗q as/k/. Group IV bəʔūl dialects represent urban varieties, where ∗q is realized as a glottal stop/ʔ/, ∗ṯ is realized as/t/, and ∗ḏ is realized as/d/. This group marks class membership as it is spoken by upper social classes (Suleiman 1985).

However, the linguistic situation in Jordan is more complicated, for many reasons. Jordan hosts many Palestinian skilled workers (Plascov 1981: 33), and many traders and academics moved from neighbouring countries in the 1920s and 1930s (Aruri 1972). More importantly, the annexation of the West Bank into Jordan in the aftermath of the 1948 Arab–Israeli War led to an increase in the proportion of the urban variety and a decrease in the nomadic patterns (Aruri 1972); for details of the linguistic situation in Jordan, please refer to Mashaqba (2015).

Many researchers have addressed the linguistic features of the Bedouin dialects, e.g. Palva (1976) on Al-ᶜAjārmah dialect, Palva (1980) on the Bani Ṣaxar dialect, Irshied (1984) and Irshied and Kenstowicz (1984) on the Bani Ḥassan dialect, Bani Yasin and Owens (1984) on the Bdūl dialect, Sakarna (1999) on the ᶜAbbādi dialect, Mashaqba (2015) on Mashaqba and Huneety (2017) on Ahl Al-Jabal Bedouin.2 All these studies come to the conclusion that each Jordanian Arabic (JA) dialect exhibits some linguistic features that make it stand out from other Arabic dialects. However, to date, the linguistic features of al-Issa Arabic have not been described, although many of them differ from the other Bedouin JA dialects. More importantly, the dialect is as at risk of losing some of its undocumented features, for many reasons. The most prominent reason is the displacement of thousands of Syrians to Jordan following the Syrian civil war which began in 2010. Since that time, hundreds of cases of marriage between Syrians and members of the al-Issa tribe have been registered, affecting the linguistic features of all Bedouin dialects (Achilli 2015: 8; Berti 2015: 47; p.c. with the head of al-Issa tribe).3 The present study, therefore, aims to fill a gap by examining selected phonological aspects of al-Issa Arabic.

The al-Issa tribe is one of the largest Bedouin tribes in the northeast of Jordan; it is descended from the Tayʔ tribe in the north of Arabia4. The tribe settled in Jabal al-Druze in Syria before moving into Jordan in the Ottoman period. Members of the al-Issa tribe reside in three main villages in the Mafraq region of northern Jordan: Ad-Dafyana and Mansiyat al-Gublan, 40 kms from the centre of the al-Mafraq governorate, and Um As-srab which is 12 km away. The al-Issa tribe is named after the founding father, Issa Bin Muhanna Bin Māniᶜ, and is divided into as-Waylim and al-ᶜAli. As-Waylim comprises the following sub-tribes: al-Māḏ̣i, al-Ḥwitta, al-Wṭafa, al-Ḥarīz and al-Ṭarwān, living in ad-Dafyana and Mansiyat al-Gublan; al-ᶜAli comprises al-Ršida, al-Gaṭṭāša, and al-Rmiḥa, and they live in Um As-srab (p.c. with the head of al-Issa tribe).

2. Methods

The researchers led two months of fieldwork in Um As-srab, Mansiyat al-Gublan and Ad-Dafyana villages, from mid-September to mid-November 2018. One week-fieldwork was also held in mid of December to double-check the linguistic features of al-Issa Arabic. During the fieldworks, ethical issues of the research were considered, including assuring participants that all data would remain confidential and anonymous, and that they had the right to withdraw from the study at any time. They were also assured that recorded materials would be destroyed as soon as they were no longer needed. After being informed about the research aims and objectives, participants signed a consent form to show their acceptance to taking part in the study.

The selection of subjects was subject to the following criteria. The total number of participants was 60 adults (30 males and 30 females), taken from the different al-Issa sub-tribes. All participants are native speakers of al-Issa Arabic and have spent all their life in one of the three villages. All are over 60 years old (average age = 63.5) to ensure that they speak the dialect of al-Issa. Participants’ levels of education range from primary school to secondary school. All participants are healthy, free from any speech impediments, and are good storytellers. Four additional language consultants were recruited to record and double-check the phonological features of the dialect.

Two data collection methods were followed. The first involved recording general conversations among twenty participants to provide a rich and sufficient corpus. In the other, twenty participants were asked to converse spontaneously for 10 min on various topics, such as Bedouin food, way of life, traditions, animals, means of communication, jobs, and memorable events. All topics were carefully selected by the researchers in order to ensure familiarity of the participants with such topics. This was followed by raising some direct questions and picture identification of animals, colours and objects to elicit additional linguistic features. Data was double-checked by the four language consultants who helped understand some lexical items and explain some phonological aspects. Data was recorded using a digital, multi-directional Olympus LS-11 voice recorder, at a distance of 15 cm from the speaker. Data were recorded indoors in a quiet room order to avoid outdoor noise.

3. The phonological aspects of al-Issa Arabic

The study aims to address the major phonological aspects of al-Issa Arabic. Below is an examination of the phoneme system, followed by syllable structure, the notable phonological processes, and then stress assignment rules.

3.1. The consonantal system

Applying the minimal pair test, we find that al-Issa Arabic has 26 main consonants plus four marginal ones. These phonemes belong to different places of articulation as follows: seven plosives (/b/,/ḅ/,/t/,/d/,/g/,/ʔ/, and/ṭ/), one affricate/č/, thirteen fricatives (/f/,/ṯ/,/s/,/š/,/ṣ/,/z/,/ḏ/,/ḏ̣/,/x/,/ġ/,/ᶜ/,/ḥ/, and/h/), three nasals (/m/,/ṃ/ and/n/), two laterals (/l/ and/ḷ/), two flaps (/r/ and/ṛ/), and two glides (/y/ and/w/).5 Table 1 below presents the consonantal system of al-Issa Arabic in terms of phonation, place of articulation and manner of articulation.

Table 1.

Consonant system of al-Issa Arabic.

Place Manner	Labial	Labio-dental	Interdental	Apico-dental	Dental-alveolar	Alveolar	Palato-alveolar	Palatal	Velar	Pharyngeal	Glottal
Plosives	b			t d					g		ʔ
Emphatic Plosives	ḅ			ṭ
Fricatives		f	ṯ ḏ			s z	š		x ġ	ḥ ᶜ	h
Emphatic Fricatives			ḏ̣			ṣ
Affricates							č
Nasal	m					n
Emphatic Nasal	ṃ
Lateral						l
Emphatic Lateral						ḷ
Flap					r
Emphatic Flap					ṛ
Glides	w							y	w

Examining Table 1 reveals that the consonantal system differs in many respects from that of CA and of other neighbouring dialects. First, the reflex of the voiceless uvular stop ∗q is a voiced velar plosive/g/ in all word positions in al-Issa Arabic; thus, the words ∗qabr ‘grave’ and ∗baqar ‘cows’ are produced respectively as 'gabir and 'bigar. The realization of ∗q as/g/ is reported in all Bedouin JA dialects, e.g. Bani Ḥasan (Irshied 1984), Bani Ṣaxar (Palva 1980), Wadi Ramm (Mashaqba 2015), as well as rural JA dialects (gāl dialects), e.g., the dialects of Ḥōrān, which include dialects spoken twelve km south of Damascus as far as Ajloun Governorate in the north west of Jordan (Cantineau 1946); it distinguishes them from ᶜAmmāni Arabic (ʔāl dialects), where ∗q is realized as/ʔ/ by female speakers, e.g. ∗qarya > ʔarya ‘village’, ∗ṭarīq > ṭarīʔ ‘street’ (Al-Wer 2007). This feature also distinguishes al-Issa Arabic from some regional dialects such as that of the Druze in al-Azraq in Jordan (El-Zain 1981), Palmyra Arabic in Syria, and Swēda Arabic spoken in southern Syria (Cantineau 1934), in which/q/ is retained(cf. the appendix to view the geographical distribution of the dialect).

Like many Bedouin and rural JA dialects, the emphatic interdental fricative ∗ḏ̣ and the emphatic dental plosive ∗ḍ have merged into/ḏ̣/ in all contexts, as shown in (1).

(1)	CA form	Al-Issa Arabic form	Gloss
	ḍarab	'ḏ̣arab	he hit
	rakaḍ	'ričaḏ̣	to run
	māḍi	'māḏ̣i	sharp
	maḍḍāha	maḏ̣'ḏ̣āha	to spend time
	ᶜaḍḍ	ᶜaḏ̣ḏ̣	to beat

Additionally, the voiceless velar stop ∗k undergoes a historical process of affrication into/č/. While such a process is conditioned in other Bedouin JA dialects, e.g., Bani Hasan (Irshied 1984) and ᶜAbbādi Arabic (Sakarna, 1999), the affrication process is unconditioned in al-Issa Arabic, similar to some central dialects of Palestine (Naïm 2011), as shown in (2):

(2)	CA form	Al-Issa Arabic form	Gloss
	kabš	'čabiš	male sheep
	kīs	čīs	bag
	ḥaka	'ḥača	to talk
	šakl	'šačil	form
	bakaw	'bičaw	they cried
	y-fakkir	'yfaččir	he is thinking
	duyūk	dyūč	cooks
	fakk	fačč	to untie

As the data in (2) shows, the historical ∗k is attested as/č/ in all word positions: initially, medially and finally. Although ∗k has been retained totally in some rural JA dialects, e.g. Wadi Mousa Arabic (Huneety 2015), urban JA dialects, e.g., Ammani Arabic (Al-Wer 2007), and some Bedouin JA dialects, e.g., Wadi Ramm Arabic (Mashaqba 2015), ∗k has been affricated into/č/ under certain conditions in other dialects; for instance, in Bani Hasan Bedouin, the word ∗kabš ‘maple sheep’ is realized as čabiš but the word ∗bank ‘bank’ is realized as banik.

An archaic feature of the dialect under investigation is the realization of the palate-alveolar ∗j as/y/ in all word positions, e.g. ∗dajāj > di'yāy ‘chicken’, ∗jāb > yāb ‘he brought’, ∗šajar ‘trees’ > 'šiyar, ∗daraj > 'diray ‘stairs’. The process is referred to as yaʔyaʔah, and differentiates al-Issa Arabic from all other JA dialects. This feature has been reported in the Tamim dialect spoken in the north of Najd, the Rashidi dialect in the Southern Sands (Southern Najdi) spoken in the Empty Quarter (Ingham 1986), and in the Bahrini dialect (Holes 1980).

3.2. Vocalic system

Similar to the vocalic system of CA and applying the minimal pair test, the vowel system in al-Issa Arabic comprises a set of eight vowel phonemes: the three short vowels/i/,/a/ and/u/, their long counterparts’/ī/,/ā/,/ū/, and the diphthongs/aw/ and/ay/ (for details on CA diphthongs, refer to Iványi 2006). In most JA dialects, the reflex of the diphthongs ∗ay and ∗aw are the two mid vowels/ē/ (/${\mathcal{ı}}^{-}$/ in Bani Hasan Bedouin), and/ō/ (/ū/ in Bani Hasan Bedouin), e.g. Ḥōrān Arabic (Cantineau 1946), ᶜAbbādi Arabic (Sakarna, 1999), ᶜAmmāni Arabic (Al-Wer 2007). These diphthongs,/aw/ and/ay/, are always monophthongized in the Honaine dialect (spoken in Algeria by two tribes: Beni Abed and Beni Kallad) as/ū/ and/${\mathcal{ı}}^{-}$/ respectively (Aïd 2015). A salient feature of al-Issa dialect is the retention of historical diphthongs/aw/ and/ay/ in all contexts, as in (3. The third column includes examples that form minimal pairs to test the phonemic profile of/ay/ and/aw/ in al-Issa dialect:

(3)	Example	Gloss	Minimal pair test	Gloss
	bayt	home	bāt	he slept
	'ḥayṭna	our wall	'ḥaṭna	he put us
	čayfha	how is she?	'čafha	her palm
	sayf	sword	sūf	Suuf village
	ṣayf	summer	ṣūf	wool
	fawz	winning	fāz	he won
	lawn	colour’	lān	he accepted
	ṣawm	fasting	ṣūm	fast! (imp.)

In Wadi Ramm Arabic (Mashaqba 2015), ∗ay and ∗aw are retained in some particular contexts, especially after gutturals (as in xayl ‘horses’, ḥaydar ‘Haidar, proper name’, hawša ‘fight’, wiᶜayna ‘we got up’, sihayt ‘I forgot’) and emphatic consonants (as in ṣayd ‘hunting’ and ṭayr ‘bird’). Palva (1976) reported that in ᶜAjārma Arabic, a Bedouin JA dialect spoken in the central part of Amman, ∗ay and ∗aw are partially monophthongized into [ey] and [ow] in the context of previous back and emphatic consonants, e.g. xeyl ‘horses’, howša ‘fight’, but fully monophthongized elsewhere into [ē] and [ō]. Abu Haider (1991) and Watson (2002) show that diphthongs are retained in the Rab${\mathcal{ı}}^{-}$ᶜa dialect in Iraq and San'ani Arabic, respectively. Retention of CA diphthongs produces a resemblance to CA, which is not present in other Arabic dialects.

3.3. Syllable structure and syllable repair processes

The syllable is a complex phonological unit that is divided into two constituents: an onset and a rhyme, which comprises the nucleus, the most sonorant element, plus a coda. Onset is an obligatory component in Arabic syllable structure, and thus in the case of onsetless syllables the prosthetic glottal stop/ʔ/ surfaces before utterance-initial syllables (cf. Watson 2002; Huneety 2015; Mashaqba et al., 2019).

Like many Arabic dialects, e.g. Palestinian Arabic (Abu Salim 1982), Jerusalem Arabic (Rosenhouse 2006) and Wadi Mousa Arabic (Huneety 2015), the core syllables of al-Issa Arabic are: CV, CVC, CVV, CVVC and less frequently C₁C₂VVC CVC₁C₁. In terms of weight, CV syllables are light, CVV syllables are heavy and CVC₁C₁ and CVVC are superheavy. The weight of a CVC syllable depends on its position in the word; it is heavy in initial and medial positions, but light in the word-final position after the final C is deemed extrametrical (McCarthy 1979; Hayes 1979, 1980, 1995). Five of these types occur unconditionally, i.e. initially, medially and finally: CV, CVC, CVV, CVVC, C₁C₂VVC whereas CVC₁C₁ syllables occur finally, where the final coda cluster is a geminate, e.g. ʔis.'tadd ‘to retrieve’, and in monosyllabic words, e.g. ḥadd ‘border’. For details on geminate representation in JA, refer to Mashaqba et al. (2021).

Similar to rural JA dialects, e.g., Wadi Mousa Arabic spoken in the south (Huneety 2015), and ᶜAjloun Arabic spoken in the north (Al-Sughayer 1990), the onset position can be filled by a maximum of two consonants, regardless of sonority which involves a gradual increase of sonority from the onset towards the peak (Clements 1990). Onset clusters are the result of some phonological processes such as syncope as in slāḥ ‘weapon’, frāš ‘furniture’, rfūf ‘shelves’ (cf. §3.3.1), and trisyllabic elision, as in ∗ba.ga.ra ‘cow’ > 'bga.ra (cf. §3.3.2).

C₁C₂C₃ clusters are prohibited in al-Issa dialect regardless of their sonority hierarchy. The definite article l-assimilates to a following [+coronal] producing a geminate as in (4a), but it surfaces as l-before [-coronal] sounds as in (4b). When the definite article l-is attached to an onset cluster where C₁ is [-coronal], this results in a C₁C₂C₃ cluster. To repair this, l-of the definite article is deleted and the glottal stop surfaces in order to avoid the vowel-initial utterances that are not permitted in Arabic dialects (cf. Watson 2002; Huneety 2015), as in (4c). In case when l-attaches to a word-initial coronal cluster, the impermissible cluster of three consonants is broken by a process of degemination, as in (4d):

(4)	a.	∗al-dār	ad'dār	the home
		∗al-zariᶜ	az'zariᶜ	the crops
	b.	∗al-mizraᶜa	al'mizraᶜa	the farm
		∗al-fazᶜa	al'fazᶜa	the help
	c.	∗ʔal-frāš	ʔaf'rāš	the furniture
		∗ʔal-brūj	ʔab'rūj	the towers
	d.	∗ʔal-grūš	ʔag'rūš	the money
		∗ʔal-ṣdūr	ʔaṣ'dūr	The chests

The data in (4b) shows that deletion of the definite article l-takes place to avoid an impermissible C₁C₂C₃ cluster. Below, we will present another strategy for repairing triconsonantal clusters in the dialect: epenthesis of the short high [i] (cf. 3.3.2).

Final-coda clusters are not permitted and they are broken through inserting the short high vowel [i] no matter what the sonority hierarchy, as shown in (5) (cf. 3.3.2):

(5)	Underlying form	Al-Issa Arabic	Gloss
	jahl	'yahil	ignorance
	farš	'fariš	furniture
	gamḥ	'gamiḥ	wheat
	libs	'libis	clothes
	dars	'daris	lesson
	kart	'čarit	card
	misk	'misič	musk

Below is an examination of syllable repair processes in al-Issa Arabic.

3.3.1. Syncope

Syncope aims to reduce the number of monomoraic syllables and increase the number of bimoraic ones (Broselow 1992; Watson 2002). The process affects the short high vowels/i/ and/u/ in open, non-final unstressed syllables followed by a bimoraic ultimate syllable (i.e., Ci,CāC and Cu.CūC) unless the ultimate syllable contains the high front/ī/ (Ci,CīC), as shown in (6a). Moreover, the unstressed high front vowel is syncopated when a vowel-initial morpheme is concatenated to a nominal stem, as in (6b):

(6)	a.	CA/ historical form	Al-Issa Arabic	Gloss
		∗fi.'rāš	frāš	furniture
		∗či.'tāb	čtāb	book
		∗su.'hūl	shūl	plains
		∗bu.'rūj	brūy	towers
		∗wi.'nīn	wi.'nīn	moaning
		∗ṣi.'ḥīḥ	ṣi.'ḥīḥ	correct
		∗ši.'ḥīḥ	ši.'ḥīḥ	scarce
	b.	∗fāliḥ-'īn	fāl.'ḥīn	They m. are successful.
		∗jāmiᶜ-'āt	yāmᶜ'āt	universities
		∗mīxir-'ayn	mīxr'ayn	two columns of tent

As seen in (6), the historical short high vowel is syncopated in an unstressed syllable in the context of a following bimoraic syllable. This syllabification repairs the structure of the phonological word in Arabic as it prefers maximizing bimoraicity (Watson 2002). It can be accounted for by assuming that the dialect avoids vulnerable monomoraic syllables which cannot construct a foot on their own and thus are left stranded because they would not be able to parse into the next prosodic level within the phonological word (cf. Broselow 1992; Broselow et al., 1997).

3.3.2. Epenthesis

Epenthesis is the process of inserting a vowel to break up a forbidden consonant cluster (Crowley 1997; Matthews 2007). The main function of adding this intervening vowel is to meet the syllable structure requirement and syllabify stray consonants (Hall 2006). Following Kiparsky (2003), Arabic dialects are typologically classified according to the position of the epenthetic vowels in C₁C₂C₃ clusters into: VC-dialect (C₁VC₂C₃) as in Iraqi Arabic (Itô 1989), CV-dialect (C₁C₂VC₃) as in Cairene Arabic (Kiparsky 2003), and C-dialect (C₁C₂C₃), e.g. Moroccan Arabic (Kiparsky 2003). Thus, the word ḥabl-ha ‘her rope’ is produced as ḥabil-ha in VC-dialects, ḥabli-ha in CV-dialects and as ḥabl-ha in C-dialects. Al-Issa Arabic is classified as a VC-dialect because in medial C₁C₂C₃ clusters the intervening vowel is inserted between C₁ and C₂, as in ∗samᶜ-hum > samiᶜ-hum ‘their hearing’. The VC dialect group is attested in Levant dialects and in Bedouin dialects, such as Bani Hassan Arabic (Irshied and Kenstowicz 1984) and Wadi Mousa Arabic (Huneety 2015).

Al-Issa dialect has a strong ban against three consonant clusters, and thus the intervening vowel [i] is added when any of the imperfect markers, e.g. y-, t-, -n, concatenates to stems with an initial cluster, e.g. y-rḥal > 'y[i]rḥal ‘he decamps’, t-lᶜab > 't[i]lᶜab ‘she plays’, n-čtib > 'n[i]čtib ‘we write’; in such cases, the epenthetic vowel forms the nucleus of the new syllable where C₁ occupies the onset of the syllable, C₂ the coda, and C₃ the onset of the following syllable. At the phrasal level, an epenthetic vowel is required when a closed syllable is followed by a word-initial C₁C₂ cluster, as in wiriᶜ kwaiyyis > 'wiriᶜ [i]kwaiyyis ‘a boy good boy’. Loanwords with complex C1C2C3 clusters are adapted in al-Issa by breaking up that cluster into CiCC as fulsčāb > ful[i]s'čāb ‘foolscap’ > and sčrāb > s[i]č'rāb ‘scrap’. Based on Kiparsky (2003) typology, this lends more support to classifying al-Issa Arabic as a VC-dialect.

The epenthetic vowel is [i] in many JA dialects, e.g. ᶜAbbādi Arabic (Sakarna, 1999), Wadi Ramm Arabic (Mashaqba 2015) and Wadi Mousa Arabic (Huneety 2015); it is the high back vowel [u] in Ḥōrān dialects (Cantineau 1946; Al-Damen 2007; Al-Ghazo 1987), and the short high vowel [a] in Sudanese Arabic (Hamid 1984) and Sanᶜani Arabic (Watson 2002). In contrast, in many Arabic dialects, e.g. Najdi Arabic (Ingham 1994), al-Karak Arabic (Btoosh 2006) and Maᶜāni Arabic (Rakhieh 2009), the insertion of the epenthetic vowel affects C₁C₂ clusters with a reverse sonority, i.e. those that fail to adhere to the sonority hierarchy principle, which requires that a peripheral segment should be more sonorous than the segment closer to the nucleus, e.g. ḥaml > ḥamil ‘meat’, jism > jisim ‘body’, badw > baduw ‘Bedouins’, ᶜabd > ᶜabid ‘slave’. However, epenthesis in C₁C₂ clusters with a falling sonority is blocked, as in samᶜ ‘hearing’.

The most frequently recorded type of epenthesis in Al-Issa Arabic is the insertion of the short high vowel [i] in CVC₁C₂ nominal stems in order to produce phonotactically correct forms. The process takes place irrespective of the sonority hierarchy principle which involves a gradual decrease in sonority in complex codas, as in (7):

(7)	∗samᶜ	'samiᶜ	hearing
	∗ʔakl	'ʔačil	eating
	∗ᶜilm	'ᶜilim	knowledge
	∗ḥabl	'ḥabil	rope

The data in (7) shows that al-Issa Arabic disallows final coda clusters. The epenthesis of [i] changes the canonical form CVC₁C₂ into CVC₁VC₂, where the epenthetic vowel functions as a nucleus for the added syllable, C₁ as the onset and C₂ as the coda. According to Kiparsky (2003), final –C₁C₂ clusters occur unlimitedly in CV and C dialects. However in VC dialects, –C₁C₂ clusters are either not allowed, or they are allowed with falling sonority. Given that breaking the final C₁C₂ clusters is a feature attributed to VC-dialects, then al-Issa Arabic is a VC-dialect type. Two pieces of evidence support treating the short high vowel [i] as an epenthetic rather than a lexical vowel. First, where CVC₁C₂ forms are concatenated to vowel-initial morphemes, the epenthesis process is blocked e.g. 'samᶜ-ač ‘your hearing’, 'ḥabl-ač ‘your rope’, where C₁ functions as the onset and C₂ as the coda of the new syllable ('sam.ᶜač and 'ḥab.lač). The second piece of evidence comes from the adaptation of loan words with a CVC₁C₂ template produced by al-Issa Arabic members; in such forms, the coda cluster is broken by an epenthetic vowel, as in 'banič from the word ‘bank’, 'šaniṣ from ‘chance’, and 'šibis from ‘chips’.

Another frequent type of epenthesis in al-Issa Arabic involves the low back vowel [a] in CaGC nominal stems whose C₂ is a guttural (G = guttural), as in (8a). Interestingly, the short high vowel [i] shows up in al-Issa Arabic nominal stems whose C1 or C3 is a guttural, as in (8b). This is different from Zawaydeh Bedouin Arabic (Sakarna, 2002) and Wadi Ramm Bedouin Arabic (Mashaqba 2015), where epenthesis is blocked in complex clusters whose third C is a guttural, e.g., damᶜ ‘tear’, wasx ‘dirty’, milḥ ‘salt’.

(8)	a.	∗laḥm	'laḥam	meat
		∗baxt	'baxat	luck
		∗baḥr	'baḥar	sea
		∗mahr	'mahar	dowry
	b.	∗ḥarb	'ḥarib	war
		∗xalṭ	'xaliṭ	mixing
		∗zarᶜ	'zariᶜ	crops

Another case of epenthesis is exhibited in imperative verbs that are morphologically derived from imperfective forms of C₁C₂VC, e.g. ∗lᶜab ‘play (imp.)’, ∗dhan ‘paint (imp.)’. To repair such clusters in the dialect, the C₁C₂ cluster is broken by an epenthetic [i], e.g. ilᶜab ‘play (imp.)’, išrab ‘drink (imp.)’, igra ‘study (imp.)’. Since Arabic permits no vowel-initial utterances, the glottal stop [ʔ] is inserted, resulting in 'ʔilᶜab ‘play (imp.)’, 'ʔišrab ‘drink (imp.)’, ʔigra ‘study (imp.)’ When the imperative verb is preceded by a word-final consonant, the final coda of the previous word takes the position of the onset instead of the glottal stop, e.g. taᶜāl ʔilᶜab ‘come to play (imp.)’ > ta.ᶜā.lil.ᶜab.

3.3.3. Raising

The term raising defines the process of vowel shift that affects tongue height, thus raising a vowel, for example, from mid to high in the context of a following high vowel (Trask 1996). Benkirane (2008) defines this process as fronting and raising the short low vowels/ā/ and/a/ towards the close front vowels/ī/ and/i/. The process has been reported in many Arabic dialects. For example, in the North Syrian group, the low vowel/ā/ is raised towards [ē] in the context of a preceding/i/, e.g. ∗kitāb > ktēb ‘book’ (Versteegh 2001); in Beirut Arabic, the low vowels/ā/ and/a/ are raised towards [e] and [ē], as in madrasa ‘school’ > madrase, lisān ‘tongue’ > lsēn (Naïm 2006: 276). In the Jalbūn dialect spoken in Palestine, the feminine morpheme marker –a is raised toward [e] except when preceded by emphatic or guttural/back consonants, as in ∗samača > samače ‘fish’, makatta > mačatte ‘ashtray’ vs qaddāḥa > kaddāḥa ‘lighter’, xalāṣa > xalāṣa ‘placenta’ (Herin 2013: 105).

Similar to many Bedouin dialects of Arabic, e.g. Negev Arabic (Blanc 1970) and ᶜAbbādi Arabic (Sakarna, 1999), al-Issa Arabic raises the short low vowel/a/ to [i] in a non-final open syllable. This raising rule affects various verbal forms, including verb Forms I, VII and VIII, changing them from CaCaC into CiCaC, n-CaCaC into n-CiCaC, and iC-t-aCaC into iC-t-iCaC, as seen in (8):

(8)	Underlying form	Surface form	Gloss
	masaḥ	'misaḥ	to clean
	katab	'čitab	to write
	saqa	'siga	to water
	n-kasar	'nčisar	to break

The same process applies to nominal stems of the pattern CaCaC, changing them into CiCaC, e.g., ∗samak > 'simač ‘fish’, ∗bagar > 'bigar ‘cow’. When the trisyllabic elision process is active, the raising process can affect the second vowel, changing the pattern from C₁C₂aC into C₁C₂iC, e.g. 'simač ‘fish’ > 'smič-u ‘his fish’, 'filat ‘to escape’ > 'flit-u ‘they escaped’ (cf. data in 3.3.3 for the interaction between raising and trisyllabic elision). Evidence in support of/a/ as an underlying vowel in CaCaC patterns before it raises into [i] comes from the fact that raising of/a/ is blocked when C₁ or C₂ is a guttural or an emphatic, as in 'ᶜazam ‘to invite’, 'laxam ‘to hit’, 'ġasal ‘to wash’, 'ḥasab ‘to calculate’, 'šaᶜar ‘hair’. This is different from al-Zwaidah Arabic (Sakarna, 2002), where the raising process is blocked only when C₂ is a guttural, but is active elsewhere. Thus where the words ∗ᶜaraj ‘to walk lamely’ and ᶜali (proper name) are realized as ᶜaray and ᶜali in al-Issa Arabic, they are produced as ᶜiraj and ᶜili in al-Zawidah Arabic (Sakarna, 2002). A further supporting piece of evidence comes from the production of loanwords of the pattern CaCaC, e.g. 'fa.raz ‘to classify’, 'ba.laf, a Persian word meaning to ‘trick’, which underwent a raising process, producing 'fi.raz and 'bi.laf; yet it fails to affect the loan word 'hačar ‘hacker’, since C₁ is a guttural.

It is worth mentioning that the raising process does not exist in many rural JA dialects, e.g. Maᶜāni Arabic (Rakhieh 2009) and Wadi Mousa Arabic (Huneety Arabic); however, in Bdūl Arabic spoken by Bedouins in Petra (Bani Yassin and Owens 1984) the low short vowel/a/ is raised toward/u/ rather than/i/ in a few cases, e.g. ruma ‘to throw’.

3.3.4. Trisyllabic elision

Trisyllabic elision refers to the process of deleting the short vowel in the context of a following non-final short open syllable (Irshied 1984). The low short vowel is elided in underlying ∗CaCiC when followed by a vowel-initial suffix, e.g. ∗laᶜib-u > lᶜab-u ‘they (m.) played’ (cf. Irshied 1984 for Bani Hasan Bedouin; Mashaqba 2015 for Wadi Ramm Arabic). It also affects the short low vowel in an antepenultimate light syllable that is followed by another monomoraic light syllable, as in ∗ba.ga.ra > bgara ‘cow’, ∗ša.ja.ra > šjara ‘tree’. The process has been attested in many Arabic dialects, including Najdi Arabic (Abboud 1979) and Bani Hasan Arabic (Irshied 1984).

Basically, trisyllabic elision in al-Issa Arabic affects Cv.Cv.Cv(C) stems which become C₁C₂v.Cv(C), as in (9). Recall that the raising process targets the short low vowel/a/, as in čatab > čitab ‘to write’ which in turn undergoes a trisyllabic elision when the word becomes trisyllabic by concatenating to a vowel-initial morpheme (cf. 3.3.3). This process basically affects the low short vowel/a/, and/i/ which results from the raising process. Trisyllabic elision involving the raising process has also been observed in a few words affecting the short back vowel/u/ as in (9e). Further investigation should be carried out on the frequency of each vowel occurrence in JA dialects.

(9)				Gloss
	a.	gaᶜad + aw	'gᶜad-aw	they (m.) sat
		gaᶜad-at	'gᶜad-at	she sat
		gaᶜad-an	'gᶜad-an	they (f.) sat
	b.	čitab-aw	'čtib-aw	they (m.) wrote
		čitab-an	'čtib-an	they (f.) wrote
		čitab-at	'čtib-at	she wrote
	c.	maᶜaz
		maᶜaz-u	'mᶜaz-aw	his goats
		maᶜaz-i	'mᶜaz-i	my goats
		maᶜaz-ak	'mᶜaz-ač	your goats
	d.	baġal-i	'bġal-i	my mule
		baġal-u	'bġal-aw	his mule
		baġal-ak	'bġal-ač	your mule
	e.	nu.gaṭ-i	'nga.ṭ-i	drops
		bu.kas-i	'bča.si	my boxes
		du.wal-ak	'dwa.law	his countries
		ju.mal-ak	'yma.lač	your sentences
	f.	ba.ga.ra	'bga.ra	cow
		ša.ja.ra	'šya.ra	tree
		baṣala	'bṣa.la	Onion

The data in (9) shows that this process affects CaCaC stems when concatenated to a vowel-initial suffix, which surfaces as C₁C₂aC-V(C). Moreover, the trisyllabic elision targets the Ca.Ca.Ca stems which do not include suffixes as in (9f) above; in such a case, the short low vowel in the antepenultimate syllable is elided to produce a CCaCa structure.

3.3.5. Gahawa syndrome

The gahawa syndrome is a process whereby gutturals are avoided in syllable-finally through changing positions with the following vowel, i.e. they are syllabified through inserting a vowel at the onset of an added syllable (Rosenhouse 2006). According to the gahawa syndrome process, an underlying CaG.CV(C) is attested as Ca.Ga.CV(C), where G =/h/,/ḥ/,/ᶜ/,/x/,/ġ/), as in (10).

(10)	∗gahwa	ga. 'ha.wa	coffee
	∗naᶜya	na. 'ᶜa.ya	ewe
	∗maᶜzūm	ma.ᶜa. 'zūm	invited
	∗baġla	ba.'ġa.la.	female mule
	∗laḥ.ma	la.'ḥa.ma	meat
	∗tax.dim	ta.'xa.dim	she serves

As seen in the above examples, the guttural consonants are banned in the coda position; to repair this, the guttural takes the onset position and the short vowel [a] is inserted, resulting in CV.Ga.CV(C). The process has been attested in some Bedouin dialects, e.g. Negev Arabic (Blanc 1970), Bdūl Arabic (Bani Yasin and Owens 1984), ᶜAjārma Arabic (Palva 1976), Al-Zawaida Arabic (Sakarna, 2002), and Wadi Ramm Arabic (Mashaqba 2015). It has also been attested in some sedentary Egyptian dialects (de Jong and Erik 2007). Although emphatic consonants (/ṣ/,/ṭ/, and/ḏ̣/) are marked as secondary guttural consonants (cf. Mashaqba 2015),6 they fail to trigger this process, i.e. there is no ban against emphatics occupying the coda position, as in 'tiṭ.laᶜ ‘she leaves’, 'ʔaṣ.bir ‘I have patience’.

The ban against gutturals in the coda position is overcome differently in other Arabic dialects, e.g. Najdi Arabic (Abboud 1979), Ḥijāzi Arabic (Al-Mozainy 1981) and ᶜAbbādi Arabic (Sakarna, 1999), whereby underlying/CaG.CV(C)/ is realized as/CGa.CV(C)/, thus omitting the short low vowel of the first syllable, resulting in an onset cluster. The process is known as the ghawa syndrome, or CV metathesis. See the examples in (11):

(11)	∗gahwa	ghawa	coffee
	∗naᶜja	nᶜaja	ewe
	∗maᶜzūm	mᶜazūm	invited
	∗baġla	bġala	female mule
	∗laḥ.ma	lḥa.ma	meat

As seen above, syllable final gutturals are avoided in these dialects by a metathesis process, resulting in an onset cluster CG. For example, in the word ∗gah.wa ‘coffee’, the glottal fricative/h/ occupies the coda position in the penultimate syllable; to repair this,/h/ shifts position with the short vowel/a/, resulting in a permissible onset cluster gh-.

3.4. Emphasis spread

Emphatic consonants in Arabic are characterized by pharyngealization, a secondary articulation associated with their primary articulation. Such consonants spread/assimilate this feature to neighbouring vowels and consonants, in what is known as emphasis spread (e.g., Kahn 1975; Watson 1999; 2002; Davis 1995). Two types of emphatics occur in Arabic dialects: primary and secondary emphatics. The former type defines the set of pharyngealized coronals (/ṭ/,/ṣ/,/ḏ̣/,/ḍ/) which spread emphasis bi-directionally to neighbouring consonants and vowels, minimally in the syllable domain and maximally in the phonological word (Watson 2002; Huneety 2015). This group of consonants contrasts with the set of plain consonants only in [+/-emphatic]:/t/,/s/,/ḏ/, and/d/ respectively. See data in (12) taken from al-Issa Arabic testing the minimal contrast of plain vs. primary emphatic consonants, and recall that/ḏ̣/,/ḍ/ merged into/ḍ/ in al-Issa Arabic (cf. §3.1):

(12)	rab'baṭ-ni	he ties me	rab'bat-ni	she brought me up
	ṭāb	to recover	tāb	to repent
	šāṭ	to get angry	šāt	to kick with leg
	'bāṣ-hum	their bus	'bāṣ-hum	he kissed them
	ṣabb	to pour	sabb	to curse
	maṣṣ	to suck	mass	to touch
	ḏ̣arr	to harm	ḏarr	ants
	'ḥaḏ̣wa	acceptance	'ḥaḏwa	horse shoe
	laḏ̣ḏ̣	to kick	laḏḏ	to be delicious

Secondary emphatics (non-primary) define the set of consonants (e.g.,/ḅ/,/ḷ/,/ṛ/) which are emphatic next to the low vowels/a/ and/ā/ (Davis 2009; Huneety and Mashaqba 2016a); refer to data in (14) for non-primary emphatics in al-Issa Arabic. The most evident effect of emphatics on the neighbouring segments is the lowering of F2 (cf. Al-Masri and Jongman 2004; Mashaqba 2015). Since emphasis spreads both regressively and progressively, some authors regard emphasis as a prosodic feature (Zemánek 2006: 205).

In this section, the study examines the domain and the directionality of emphasis in al-Issa Arabic. It also examines the influence of the emphatics on adjacent vowels and consonants. The primary source of emphasis in al-Issa Arabic is the set of pharyngealized segments:/ṭ/,/ṣ/ and/ḏ̣/, which spread emphasis to other neighbouring segments, phonemically contrast respectively with their plain counterparts/t/,/s/ and/ḏ/, and can occur in all word positions, as in (13):

(13)	Word-initially	Word-medially	Word-finally
	ṭāb to recover	ma'ṭār airport	liṣṣ theaf
	ṭāl to bring	'bāṣ-hum he stays long	naṭṭ to jump
	'ḏ̣allal to shadow	'māḏ̣i past	yaḏ̣ḏ̣ to hurt
	ḏ̣aym oppression	'n-ṣawwir we take photos	ḥayṭ wall

Additionally, the dialect has a set of secondary emphatics: the lateral/ḷ/, the voiced bilabial nasal/ṃ/, the dental-alveolar flap/ṛ/ and the bilabial plosive/ḅ/. Such emphatics exhibit phonemic contrast in a few words in the contiguity of the short high vowels/a/ and/ā/, as in (14).

(14)	Plain	Emphatic
	'walla to appoint	'ẉaḷḷa by God
	'b-alla to wet something	'b-aḷḷa really!
	'daxal to enter	'daxaḷ to seek asylum
	xāl beauty spot	xāḷ uncle
	ḥabb to love	ḥaḅḅ to kiss
	baḥḥ to wash dishes	ḅaḥḥ nothing
	mayy Mai [proper name]	ṃayy water
	šām to waive his right	šāṃ I smell
	xamm non-word	xaṃṃ to grab
	'barra to exonerate	'baṛṛa outside

This contrasts with Juffēn Arabic (Huneety and Mashaqba 2016a), where the set of secondary emphatics comprises the lateral emphatic/ḷ/, which surfaces with the different lexemes of the word Allah or in the context of a following/a/,/ā/ or/u/ preceded by any of the back consonants/x/,/g/ or/ġ/, e.g., x̣āḷ ‘maternal uncle’, and gāḷ ‘to say’. For further readings on emphatic/ḷ/ in Arabic, refer to Ferguson (1956).

Previous studies on emphasis show that it is a gradient and differs cross-dialectically (Watson 2002, 2007; Mashaqba 2015). Arabic dialects differ with respect to the domain of emphasis spread and which segments block the flow of emphasis. Like many Arabic dialects, emphasis is bi-directional; the minimum domain is the syllable, and it may extend over the word. Data in (15) provides examples of rightward and leftward spread of emphasis in al-Issa Arabic; elements affected by emphasis spread are underlined.

(15)	Leftward emphasis	Rightward emphasis	Bi-directional emphasis
	šaṣṣ to look	ṭamm to fill a hole with soil	'ma.ṭar rain
	xayṭ thread	'ṣal.la to pray	'xa.ṭar danger
	'fā.yiḏ̣ overfilled	'ṭa.lab order	'mis.ṭa.ra line
	baṭṭ duck	'ṭar.aḥ to knock down	'm-ṭaw.wiᶜ old man
	'ru.baṭ to tie	ḏ̣arr to cause harm	'saṭ.ṭar to line up
	'ba.laṣ to flee	'ṣa.max to hit the head	'ba.ṭal hero
	naṭṭ to jump	ṭall to look	'ru.ṭab dates
	ḥaṭṭ to put	ḏ̣amm to hug	'na.ḏ̣ar eyesight

As can be seen from the above lists, leftward emphasis is absolute and affects the stem of the phonological word. This is similar to many Arabic dialects where leftward emphasis is unrestricted, e.g. ᶜAbbādi Arabic (Sakarna, 1999), Wadi Mousa Arabic (Huneety 2015) and Wadi Ramm Arabic (Mashaqba 2015). Interestingly, in rightward emphasis, an emphatic segment spreads emphasis within the tautosyllabic segments. In a case where C₁ of the geminate is tautosyllabic, it spreads emphasis to the adjacent syllable, e.g. 'ṣalla ‘to pray’, 'ḥaṭ.ṭaw ‘they put’. In other JA dialects, rightward emphasis may extend over the whole word but is blocked by the high front vocoids/i/ and/u/ in Jerusalem Arabic (Card 1983),/i/,/u/,/y/,/š/ and/w/ in northern Palestinian Arabic (Davis 1995),/i/,/y/,/j/, and/š/ in southern Palestinian Arabic (Davis (1995)/i/,/ī/ and/y/ in Wadi Ramm Arabic (Mashaqba 2015). In ᶜAbbādi Arabic (Sakarna, 1999), emphasis is absolute and is not blocked by any segments.

In leftward and rightward emphasis, an emphatic segment fails to spread emphasis into adjacent prefixes and suffixes unless they are tautosyllabic, as shown in (16) below:

		Gloss
(16)	a.	'l-ṭā.rig for Tariq
	b.	'n-ḥiṭ.la.hum we put something for them
	c.	'y-ṭaw.wil to stay long
	d.	'ṭla.ᶜ-aw they (m.) went outside

As seen in the above examples, emphasis reaches the prefixes l-, n-, and y- (a-c) because they are tautosyllabic with the emphatic segments. However, it fails to affect suffixes -la.hum in (b) and -aw in (d) because the suffixes are heterosyllabic with the emphatic segments. The spread of emphasis in al-Issa Arabic never reaches a previous or following word. This contrasts with Ṣanᶜāni Arabic (Watson 2002), where in phrases invoking God, emphasis can reach a previous word, as in in-šāḷḷa ‘God willing’, wa-ḷḷah ‘by God's name’, where in both cases the emphatic ḷ influences the previous words and make them emphatic.

3.5. Stress assignment

Stress is the suprasegmental feature of a syllable that is produced with a greater amount of energy than unstressed syllables. Three factors define a stressed syllable: a longer duration, greater intensity, and a higher pitch than other syllables within the word (De Jong and Zawaydeh, 1999; Zuraiq and Sereno 2007; Ladefoged and Johnson, 2011; Kager 2009; Mashaqba, 2015; Huneety and Mashaqba 2016b; Mashaqba and Huneety 2018; Huneety et al., 2020). Examining stress patterns in al-Issa Arabic shows that as in many Arabic dialects, e.g. ᶜAjārma Arabic (Palva 1976), northern JA (Bani-Yasin and Owens 1987). ᶜAbbādi Arabic (Sakarna, 1999), ᶜAjloun Arabic (Abu Abbas 2003) and Bedouin JA in the north (Huneety and Mashaqba 2016b), we find that assignment of stress depends on the weight of the syllable (light, heavy or superheavy) and the distance of the syllable from the right edge of the word. The following is a summary of the stress patterns in al-Issa Arabic:

• 1A final superheavy syllable CVC₁C₁ (where C₁C₁ are identical) or CVVC attracts stress, as in min.'šār ‘saw’, baš.'kīr ‘bathrobe’, mi.'zāt ‘offensive’, ʔal.'ḥadd ‘the border’, and mᶜa.'lān ‘business owner’
• 2If the word lacks an ultimate superheavy syllable, stress is assigned to the penultimate syllable if it is heavy, as in 'luz.ma ‘close cousins’, ʔi.'rub.ṭan ‘tie me (imp.)’, 'tḥaš.far ‘to exert the best endeavour’, mas.'mū.ḥa ‘allowed’.
• 3Stress goes to the antepenultimate syllable if the word lacks a final superheavy syllable and a penultimate heavy syllable, e.g. ṭan.ni.'šū.la.hum ‘they ignored them’, 'či.ta.bit ‘I write’, 'ḥiṭ.la.hum ‘put something to them’, and to the penultimate in disyllabic words, e.g. 'wi. riᶜ ‘boy’, 'či.tab ‘to write’. 'ṭaw.waḥ ‘to throw’.

Examining the interaction of stress and epenthesis, we find that epenthetic vowels are visible for stress, i.e. they can bear stress like lexical vowels. In the examples given below, an epenthetic vowel is added to avoid an impermissible C₁C₂C₃ cluster; this epenthetic vowel forms a syllable that is visible to stress. According to the stress algorithms given above, examples (17a), (17b) and (17c) lack any superheavy syllable; since the penultimate syllable with the epenthetic vowel is heavy, it takes the main stress, as in data in (17):

17	a.	yabl-ha	ya.'b[i]l-ha	before her
	b.	samᶜ-hum	'sa.m[i]ᶜ.hum	their hearing
	c.	čitab-t-ha	či.ta.'b[i]t.-ha	I wrote it

This is different from some Arabic dialects, e.g., Iraqi Arabic (Watson 1999) and Maᶜāni Arabic (Rakhieh 2009), where epenthetic vowels are always invisible to stress and therefore stress is assigned as if a syllable had no epenthetic vowels. Thus in the word ka.'ta.b[i]t.ha ‘I wrote it’, stress is placed on the antepenultimate syllable; although the penultimate syllable is heavy, it fails to bear stress since it has an epenthetic vowel; this suggests that stress takes place before epenthesis.

In the case of gahawa syndrome, the inserted syllable that includes the guttural consonant attracts the main stress in the absence of an ultimate heavy syllable, as in ga.'ha.wa ‘coffee’, na.'ᶜa.ya ‘ewe’. This contrasts with CV.CV.CV words, where stress falls on the antepenultimate syllable, as in 'či.ta.bit ‘I wrote’ and 'mi.sa.ḥit ‘I erased’.

Interaction between syncope and stress confirms that the deletion process targets unstressed segments which reflects the ability of speakers to identify strong segments of the word and then preserve them. The syncopated syllable is never a stressed one; it fails to take place when the short high vowel is stressed, as in 'fi.him ‘to understand’, 'či.sab ‘to earn’, as it fails to affect the short low vowel in the same context, i.e. in open, unstressed positions, as in ḥa.'l${\mathcal{ı}}^{-}$m ‘forbearing’, ᶜa.'r${\mathcal{ı}}^{-}$s ‘groom’. The melodic process of raising involves changing the short low vowel/a/ quality into the short high front vowel [i] which does not affect the word prosody and does not trigger stress shift. For example, the underlying forms ∗čatab ‘to write’ and ∗masaḥ ‘to erase’ surface as 'čitab and 'misaḥ where no stress shift is reported.

Thus, a penultimate (C)CV syllable attracts stress in disyllabic words in the absence of a final heavy or superheavy syllable, e.g. 'rsa.law ‘they send’, and in polysyllabic words as a result of gahawa syndrome, e.g. la.'ḥa.ma ‘meat’. In the absence of an ultimate and penultimate heavy syllable, an antepenultimate CV syllable always attracts stress regardless of the weight of preceding syllables. Again, a CVC syllable takes its weight from its position in the word; where it is heavy in initial and medial positions, it is equal in weight to a CV syllable in final positions after the final C is deemed extrametrical (McCarthy 1979; Hayes 1995).

Drawing on the metrical model advanced by Hayes (1995), we find that similar to many Eastern Arabic dialects, e.g. ᶜAjloun Arabic (Abu Abbas (2003) and Wadi Mousa Arabic (Huneety 2015), the dialect under investigation follows a moraic trochaic system ('μ μ), and therefore in words like 'li.fa ‘he came’ and 'či.tab ‘he wrote’, stress goes to the leftmost syllable given that the last consonant in 'či.tab is extrametrical. This contrasts with Zalabiah and Zawaideh dialects in the south of Jordan which have an iambic foot (μ ′μ); thus stress goes to the rightmost syllable in words like li.'fa ‘he came’ and ki.'tab ‘he wrote’. (For details, refer to Mashaqba and Huneety 2018).

Additionally, the direction of foot parsing works from left to right and there is a strong ban against the construction of single moras in the dialect at either edge of the word, and therefore they are left unfooted. For example, in the word ʔa.wā.dim ‘people’, foot parsing that goes from left to right over the antepenultimate syllable ʔa. This syllable is monomeric, and therefore it fails to construct a foot by itself not can it construct a foot with the following bimoraic syllable wā given that al Issa Arabic has a strong ban against degenerate foot. The penultimate syllable wa: is heavy after the long vowel contributes two moras and it constructs a foot that attracts the main stress. The ultimate CVC syllable dim is monomoraic and therefore cannot construct a foot by itself; it is left unfooted as a result.

The End Right Rule (ERR) is in place to account for the selection of the rightmost visible foot. Thus, an ultimate superheavy syllable, i.e. CVCC or CVVC, receives the main stress following ERR. Following Watson (2002), a superheavy syllable is made up of a heavy canonical syllable (bimoraic) and an extrasyllabic consonant. For example, in the word bar.mīl ‘barrel’, the ultimate CVVC syllable mīl ‘barrel’ comprises a canonical syllable that constructs a foot that attracts the main stress in conformity with ERR, whereas the extrasyllabic syllable is left unsyllabified.

The following is a summary of the metrical parameters of stress in al Issa Arabic:

• (a)Consonant Extrametricality: C → ⟨C⟩/___] word
• (b)Foot Construction: Form moraic trochees from left to right.
• (c)Degenerate feet: Forbidden absolutely.
• (c)Foot Extrametricality: Foot → ⟨Foot⟩/___] word
• (d)Word Layer Construction: End Rule Right

4. Conclusion

This study identifies the major phonological aspects of al-Issa Arabic, a Bedouin JA spoken in the north of the country that had not previously been investigated. Analyzing data from 60 participants shows that al-Issa Arabic exhibits some linguistic characteristics that make it stand out from fellow/neighbouring Jordanian and regional dialects. The major linguistic features of al Issa-Arabic, which are not exhaustive, are summarized below:

• •obligatory gahawa syndrome
• •application of trisyllabic elision
• •retention of/aw/ and/ay/
• •unconditioned affrication of ∗k into/č/.
• •the realization of the palate-alveolar ∗j as/y/ in all word positions
• •syllabification phenomena include gahawa syndrome, [i] epenthesis regardless of sonority hierarchy, syncope, and trisyllabic elesion.
• •sound shift of/a/ > [i] in an open syllable
• •emphasis spreads bidirectionally within the phonological word; leftward emphasis is absolute, but rightward emphasis spreads emphasis within the tautosyllabic segments
• •trochaic moraic stress ('Ca.CaC)
• •in terms of Kiparsky (2003) classification, al-Issa Arabic is a VC dialect.
• •In terms of Clevenald's typology of JA dialects, al-Issa Arabic belongs to yigūl group.

Further examination of the major morphological and syntactic aspects of al-Issa Arabic is recommended. A basic lexicon of this dialect should be documented by collecting the major vocabulary within thematic categories. Further comprehensive examination of JA dialects should be taken into consideration, especially those that have not yet been investigated, e.g., Badarin Bedouin Arabic, North West Bedouin Arabic, Majarr Rural Arabic, and Rumtha Rural Arabic.

Declarations

Author contribution statement

Anas Al Huneety, Bassil Mashaqba: Conceived and designed the analysis; Analyzed and interpreted the data; Contributed analysis tools or data; Wrote the paper.

Riyad Abu Hula, Baraah Khalid Thnaibat: Analyzed and interpreted the data; Wrote the paper.

Funding statement

This research did not receive any specific grant from funding agencies in the public, commercial, or not-for-profit sectors.

Data availability statement

The data that has been used is confidential.

Declaration of interests statement

The authors declare no conflict of interest.

Additional information

No additional information is available for this paper.

Appendix A. Supplementary data

The following is the supplementary data related to this article: