The Emergence of a Mixed Type Dialect: The Example of the Dialect of the Bani ˁAbbād Tribe (Jordan)

Abstract

The present article aims at questioning the status of the šāwi dialect of the Bani ^ʕAbbād tribe by providing a new analysis of the main distinctive phonological, morphological, and syntactical traits which may hint at dialect mixing. The data provided by the field research, based on a functional framework that relies on descriptive linguistics and a typological approach, show that this dialect is deeply affected by a koineizing tendency due to increasing contacts with the populations of the neighboring areas (especially ^ʕAmmān and Salṭ) which, in turn, leads to the gradual loss of its authentic features. Finally, this paper discusses whether the dialect of the Bani ^ʕAbbād should still be considered as belonging to the yigūl group (recently renamed Central Bedouin ygūlu) of the Syro-Mesopotamian sheep-raising tribes or if a new typology of mixed type dialects should eventually be adopted for the dialects displaying important markers of both Bedouin and sedentary types.

1. Introduction

The Traditional Classification of Jordanian Dialects in the Light of Recent Developments

The Hashemite Kingdom of Jordan is characterized by a considerable diversity of regional dialects.

Nevertheless, in some cases the typological classifications of the Arabic dialects spoken in this country prove to be problematic, because there seem to be a considerable number of transitional and mixed type dialects.

As stated by Sawaie (2011, p. 499): “records of the linguistic situation in Trans-Jordan in the early part of the 20th century are not available. Consequently, it is hard to state with certainty which dialects dominated then”.

The only available source on the Arabic dialects spoken in this area in this period is Bergsträsser’s (1915) Sprachatlas von Syrien und Palästina.

In 1936 Cantineau drew a classification of several nomadic dialects in his Études sur quelques parlers de nomades arabes d’Orient. In the first part of his study, he distinguished the dialects of the camel-rearing tribes from the small-cattle ones and divided the latter into two groups: the atrochaic and the trochaic dialects (Cantineau 1936, p. 114). However, after extending the scope of his research to some other tribes and completing some data he had previously collected, in the second part of his study Cantineau (1937, p. 110) classified the dialects into four groups (which he labelled as a classement rationnel): group A, B, C and Bc.

Further investigations were reported in 1963 when for the first time R. Cleveland typologically classified the Jordanian dialects through phonological, morphological, and syntactic characteristics and drew “the most general outlines of the situation illustrated by a very limited number of dialectal characteristics” (Cleveland 1963, p. 56).

In his work, he divided the Arabic dialects spoken in the area into four groups:

(1) Yigūl–(2) Bəgūl-(3) Bəkūl-(4) Bəˀūl.

The groups were named according to the pronunciation of “he says” since this feature “indicates both an important phonetic and morphological characteristic” (Cleveland 1963, p. 171).

From 1969, Palva dedicated a considerable number of publications to the linguistic situation of Jordan and in 1984 he set a more complete classification of the Jordanian dialects in his “A General Classification for the Arabic Dialects Spoken in Palestine and Transjordan”, where he added some new criteria of analysis with regard to those applied by Cleveland, namely the reflex of the sequences CvCaCv- and –aXC (where X is a guttural), the gender distinction in the 2nd and 3rd persons plural in personal pronouns and verbs, the use of the adverbs here and now, and the occurrence of the compound negation mā…š.

The new classification he outlined divides the Jordanian dialects into three main groups:

• The urban Palestinian dialects (biˀūl), which typologically belong to the Levantine urban dialects;
• The rural dialects, which are in turn grouped into:
  • the Galilean dialects (biqūl);
  • the central Palestinian dialects (biḳūl), more conservative than the Galilean dialects;
  • the south Palestinian dialects (bigūl), closely related to the previous ones, they have some features that show a greater Bedouin (especially Negev) influence;
  • the north and central Transjordanian dialects (bigūl), closely related to the Horan dialects;
  • the south Transjordanian dialects (bigūl), that are influenced by the Hijazi Bedouin dialects of Arabia Petraea and represent a mixed dialect type;
• The Bedouin dialects, which are in turn divided into:
  • The dialects of Negev (bigūl), which show some features typical of the sedentary dialects (namely the b-imperfect). The dialects of this group typologically belong to the Sinai type, and they exhibit some similarities with the Bedouin dialects of Arabia Petraea;
  • The dialects of Arabia Petraea (yigūl), they display some affinities with the Hijazi dialects;
  • The dialects of the Syro-Mesopotamian sheep-rearing tribes (yigūl), spoken in Transjordan, “(they) belong to the same type as the rest of the dialects of the sheep-rearing tribes in the Syrian and Mesopotamian peripheries of the Syrian Desert” (Palva 1984, p. 372).
  • The dialects of the North Arabian Bedouin type, yigūl, spoken in Transjordan by the Sirḥān, the Bani Ṣaxar and the Bani Xālid.

Herin et al. (2022) have recently implemented Cantineau, Cleveland and Palva’s groupings with new elements and obtained the following categories1:

(I)

Sedentary bigūlu: Mu^ʔābi and Balgāwi-Ḥōrāni;

(II)

Southern Bedouin ygūlu (Ḥwēṭāt, Bdūl, Zawāyda, etc…);

(III)

Southern Bedouin bigūlu (mostly Nagab and Sinai);

(IV)

Central Bedouin ygūlu (^ʕAǧārma, ^ʕAdwān, ^ʕAbābīd, etc…);

(V)

Northern Bedouin ygūlūn: ^ʕNizi, Šammari, Bc (Misā^ʕīd), Šāwi, Ca (Bū ^ʕĪd et ^ʕĪdīn in Lebanon, so far unattested in Jordan)

Their “taxonomy” has the merit to account for the changes caused by recent dialect contacts between speakers of Bedouin sub-groups varieties but does not take into consideration the effects of the increasing dialect mixing between the Bedouin and the sedentary population, which contributed to blur the linguistic boundaries set by this dichotomy. In particular, some Bedouin tribes such as the Bani ˁAbbād (or ˁAbābīd), who live in the vicinity of the big urban centers, start to assimilate the speech habits of neighboring communities and lose the authentic badawī features that characterize their dialects.

During the last century, most of the Jordanian tribes, except for some living in the southern districts of the country (like the Bdūl), abandoned their nomadic lifestyle and settled in some definite areas in semi-sedentary or sedentary conditions.

Many economic and social factors have contributed to intensifying the contacts between Bedouin and sedentary variations and the tendency to dialect levelling or koineization. Among these, the most relevant factors were: the increasing contacts with the inhabitants of the cities of Salṭ and ˁAmmān and those with the refugees coming from Palestine and Syria, the growing access to education, the proliferation of television series produced in Egypt, Syria, and Lebanon, and in some cases even inter-marriage with sedentary people (Rosenhouse 1984).

This is particularly evident in the dialect of the ˁAbābīd, in central Jordan: their dialect is deeply affected by a levelling process toward sedentary dialects that has left traces in phonology, morphology and syntax that will be discussed in detail in the section Results.

In the light of these findings, the aim of this study is to question the status of the šāwi yigūl dialect type that is geographically and historically attributed to this vernacular and to point out that the adoption of a more refined classification would better account for mixed dialects which display important markers of both Bedouin and sedentary types.

2. Materials and Methods

This article is based on a corpus consisting of 20,000 words2 of transcribed unmonitored interviews3 of both men and women.

The speakers selected for this qualitative analysis belong to the Bani ˁAbbād tribe, are distributed across different ages (the youngest girl was 5 years of old at the time of the interviews and the oldest man was 94 years old) and display various educational and socio-economic backgrounds.

The data resulted from two fieldwork campaigns carried out in July and August 2016 and from January until August 2017.

In order to limit the impact of my presence on the oral productions of the informants, the interviews4 were carried out by Jordanians5: a boy from Karak, and one belonging to the semi-nomadic tribe of the ˁAǧārma, as well as an ˁAbbādi girl.

2.1. The Tribe

The Bani ˁAbbād are a confederation that is divided into two main groups: al-Ǧburiyya and al-Ǧrumiyya (Peake 1958, p. 166). As noted by Shryock (1997, p. 40), “over time the ˁAbābīd have been internally fragmented and politically weak. […] This lack of consensus is commonly attributed to the diverse genealogical origins of the tribe’sclans”.

According to Oppenheim (1943, p. 227), the areas most densely populated by the Bani ˁAbbād were Māḥiṣ, Wād is-Sīr, ˁArāg al-ˀAmīr, ˁĀrḏ̣a and al-Ġōr. However, nowadays some branches of the tribe also live in Bader al-Ǧadīda, aṣ-Ṣəbīḥi, aš-Šūna aǧ-Ǧanūbiyya, al-Karāma, Wādi aš-Šitā́, Yarga, Marǧ al-Ḥamām, ˁĪra, and Nāˁūr, in the periphery of ˁAmmān and in part in Salṭ (Figure 1 and Figure 2).

Most of the land in the Balga governorate belongs to the ˁAbābīd and indeed, a common Jordanian saying recites ˁAbbād min sīl ˁa-s-sīl (ˁAbbād from river to river), alluding to the fact that their tribal territory extends from the Zarqa Torrent to the Jordan River.

The exact number of members of the tribe is hard to estimate since there are no official censuses available. Nevertheless, according to the data recorded during my fieldwork the tribe could be composed of approximately 350,000 people.

Shryock (1997, p. 43) reports that until the 1950s most of the ˁAbbādi clans were still living in tents (byūt aš-šaˁr in Arabic). They used to spend the winter in the Jordan Valley (locally called Ġōr or Aġwār), where they grazed sheep and goats, whereas during the spring and summer months they moved to its crest (the šifa), where the temperatures are cooler.

Unlike the desert tribes, the Bani ˁAbbād had always been engaged in farming and in the late 1960s some members of the tribe started to prefer a more sedentary way of life. However, according to Shryock (1997, p. 46) even “in the midst of this rapid change, the Balga tribes [have always] consider[ed] themselves fully Bedouin”and their identity is strictly linked with the values of karāma (generosity) and ˁaṣabiyya (inter-tribal solidarity).

As stated by Sakarna (1999, p. 8) “today, the people of ˁAbbādi tribe are no longer nomads and mostly live in settlements. They work as government employees, military individuals, farmers, and in other kinds of occupations”.

2.2. Functional Framework

The research method applied in this work relies on descriptive linguistics which is “based on the empirical observation of regular patterns in natural speech” (François and Ponsonnet 2013, p. 184).

The present linguistic analysis is built on a corpus of narrative transcribed texts (20,000 words) taken from the recordings of unmonitored speech. After collecting this body of data, it was segmented through the program ELAN, and analyzed in order to “identify the distinctive component of the system and the principles that underline its organization” (François and Ponsonnet 2013, p. 184).

The functional framework used for my analysis combines typological and discourse-based approaches, since this “provides the tools necessary to address questions of the meanings underlying language variation” (Brustad 2000, p. 7).

In order to efficiently describe and explain the phonological and morphological variation of the dialect of the Bani ˁAbbād, the methodology adopted was to synthesize those concepts that are able to most efficiently account for the data recorded during my fieldwork. It especially relied on the doctoral thesis of Herin on the dialect of the city of Salṭ (which to date constitutes the only Jordanian vernacular exhaustively described), and the work of Haspelmath and Sims (2010), Understanding Morphology.

While examining the syntactical structures of the data corpus collected, I referred in particular to Eléments de syntaxe générale and Syntaxe Générale, une introduction typologique, Tome 1 and 2 by Creissels (1995, 2006a, 2006b, respectively), and The Syntax of Spoken Arabic by Brustad (2000).

3. Linguistic Analysis

Although in 1999 Sakarna analyzed some phonological aspects of the dialect of the Bani ˁAbbād, in particular those of the As-Sakarna branch, no in-depth study had been carried out on the dialect of the Bani ˁAbbād tribe when in 2016 I started to gather the material for my master’s thesis Le dialecte des Bani ˁAbbād: Analyse des traits phonologiques, morphologiques et syntaxiques discriminants6 (Torzullo 2018).

According to the information reported in Cantineau (1937), the above-mentioned classifications provided by Cleveland (1963) and Palva (1984), and the description of the dialect of the semi-nomadic al-ˁAǧārma tribe (Palva 1976), my initial assumption was that the dialect of the Bani ˁAbbād typologically belonged to the Jordanian ygūl-group which is part of the so-called Šāwi dialects (in French also called petits-nomades).

My findings show that this dialect has a number of characteristics belonging to the small-cattle nomads described by Cantineau, such as the gender distinction in the 2nd and 3rd pl. persons, but they point out that it is deeply affected by a koineizing tendency that caused some significant changes in the structure of the vernacular and a loss of some authentic features.

This phenomenon is due to the increasing contact with the sedentary population of the neighboring areas of ˁAmmān and Salṭ, especially due to economic and educational reasons: most of the members of the tribe have access to higher education, which brings them to study in the capital, where the nearest universities are to be found. Furthermore, many ˁAbābīd work in these two cities and consequently have daily exchanges with people who do not speak their same variety of vernacular, and in some cases do not even understand their original dialect. These circumstances force them to opt for a more common variety, or at least to put aside some of the original traits of their vernacular and adopt some madani (urban) linguistic features, in order to facilitate the communication with the outsiders of the tribe.

If, on the one hand, this convergence to a “common linguistic style” (Giles and Ogay 2007, p. 296) for the sake of intelligibility has improved the effectiveness of their contacts, on the other, this “accommodative code variation” (Giles et al. 1973, p. 179) has led over time to a penetration of sedentary elements in the phonology, morphology, and syntax of their speech and to the disappearance of distinctive šāwi marks, mainly among the youngest members of the tribe.

3.1. Contact Induced Changes in the Contemporary Dialect of the Bani ˁAbbād

3.1.1. Phonology and Phonotactics

Reflexes of OA /q/

Old Arabic (OA) /q/ is usually realized as a voiced velar stop/g/in the dialect of the Bani ˁAbbād. Ex: gawi ‘strong’, garye ‘village’, manṭega ‘region’, dagīga ´minute´, galb ´heart´, giṣṣa ´story´.

However, it does not occur as a palatal variant /ǧ/ in the contiguity of front vowels, as it certainly was in the 1960s and 70s in the yigūl dialect of the ˁAǧārma tribe who also lives in the Balga District.

In the corpus under analysis there are only four instances of this phenomenon: two of them, ˁeǧib ‘young children’ and ǧiddāmi ´in front of me´, were produced by the oldest members of the tribe interviewed, a woman and a man of 90 and 94 years old respectively, and the latter two, ǧider ´cauldron, copper pot´ and šiǧǧ ‘part of the tent reserved to men’, were obtained through elicitation.

The words ǧiddāmi and ǧider occur along with their variants geddāmi and gider where the velar is not realized as affricate. This alternation can be considered as an example of “stylistic contrasts plain colloquial vs. koineized colloquial” (Palva 1976, p. 10).

Reflexes of OA /k/

The affricate /č/ of OA /k/ shows a regressive character within this dialect. It rarely appears in spontaneous speech, and when occurring it is only used by the oldest speakers: iḥčilha ‘tell her’, tḥači ‘you (f.) speak’, čīl ‘measure, weigh’, čīf ‘how’, bičīla ‘type of Bedouin dessert’, čänna ‘daughter-in-law’, hēč ‘thus’, čalb ‘dog’, yčammilen ‘they (f.) complete’, čam ‘how much’, čiṯīr ‘a lot’, mičān ‘place’.

Given the number of instances of this variant it is possible to conclude that these lexemes are “vestigial variants” and represent some “fossilised traces of an earlier dialect system when such forms were more general” (Trudgill 1999, p. 321).

Even the morphological contrast between -k and -č, which allows the opposition of the pronominal suffix of the 2nd m. sing. person and the pronominal suffix of the 2nd f. sing. person, is poorly attested7. It occurs in the speech of the older and less-educated speakers, while the youngest and those who work or study in ˁAmmān replace the allomorph -č with -ki.

The recessive status of the affricate /č/ is particularly evident within a family I interviewed: the use of the affricate /č/ gradually disappears over the three generations of women who live in the same household8:

• The grandmother regularly uses it in the contiguity of a front vowel, and she uses čān even to express the verb ‘to be’.
• The mother uses the affricate /č/ especially to contrast the pronominal suffix of the 2nd m. sing. person and the pronominal suffix of the 2nd f. sing. person; she never uses čān to express the verb ‘to be’, and the only word where she employs / č / is čänna ‘daughter-in-law’, while reciting a poem.
• The daughter never uses the affricate /č/.

This example clearly showcases the shift from a purely Bedouin trait of the dialect to a more mixed type one by sedentarization.

Reflexes of the Old Diphthongs /ay/ and /aw/

The monophthongisation in /ī/ and /ū/ of the old diphthongs /ay/ and /aw/ does not appear as a prominent trait in this dialect.

In Studies in the Arabic Dialect of the Semi-Nomadic әl-ˁAǧārma Tribe, Palva (1976, p. 19) reports: “one of the most striking characteristics distinguishing the dialects of the nomadic type from those of the sedentary type in the Syro-Palestinian dialect area is the fluctuation /ē/-/ī/ and /ō/-/ū/”.

This feature is briefly mentioned by Cantineau in Les parlers arabes du Ḥōrān (Cantineau 1946, p. 156), who refers to it as occurring in sporadic cases, while Bettini (2006, p. 30) only accounts for a fluctuation between the realizations /ē/ and /ī/ of /ay/ (which appears as a common trait in the dialects analysed in Contes Féminins de La Haute Jézireh Syrienne) and a long vowel /ō/ for the diphthong /aw/.

Behnstedt (1997, pp. 62–63), on the other hand, does not report this fluctuation at all in the Map31 of his Sprachatlas von Syrien.

In the corpus under analysis there are only four occurrences of monophthongisation in /ī/ and /ū/ occurring in natural speech: čīl ‘measure, weight, čīf ‘how’, ḏ̣īf ‘guest’ and hūšāt ‘fights’, together with a sentence reported by a man describing the Bedouin traditions concerning the drinking of coffee:

(1)	Ygūl-lak	awwal	finǧān	la- ḏ̣- ḏ̣īf
	to _say.IPFV.3msg-DAT.2msg	first	cup	for-ART-guest
	ṯāni	la-s-sīf	aṯ- ṯāliṯ	la-l- ḥīf
	second	for-ART-sword	ART-third	for-ART-injustice
	One says: the first cup is for the guest, the second to talk about war, and the third to discuss injustice.

In addition to these instances, two examples of a fluctuation between the realizations /ē/ and /ī/ were obtained by elicitation: līl ̴ lēl ‘night’ and zīt ̴ zēt oil´.

According to the tendencies that appear in the data, the realizations /ē/ and /ō/ of the old diphthongs /ay/ and /aw/ gained ground in the dialect of Bāni ˁAbbād, to the detriment of the most archaic ones. So, it is possible to observe: bēt ´house, tent´, bēn ´between´, ġēr ´other´, xēr ´good, well´, xēl ´horse´, ˁēla ´family´, ˁēn ´eye´, tōr ´bull´, yḥōšen ´they (f.) plough´, zōǧa ´wife´, xōf ´fear´, fōg ´above´, ṭōr ´cave´.

This state of the affairs suggests that the above-mentioned statement by Palva (1976) concerning the monophthongisation in /ī/ and /ū does not currently hold true for this dialect. The few instances of the phonemes /ī/ and /ū/ (</ay/ and /aw/) may thus represent the last traces of an older monophthongisation system where such forms were more widespread, and that in the last decades have been replaced by more sedentary forms.

Cases of Trochaism

According to Cantineau (1936, p. 114), trochaism is a typical trait of some petits nomades dialects, namely those of the ^ʕŌmūr, Ṣlūt, Bani Xāled and Sirḥān.

Dialects characterized by a trochaic rhythm maintain the older Arabic inflectional -a- as well as the -a of the feminine morpheme between a long syllable and a pronominal suffix (Palva 1976, p. 25), i.e., CvCC or C$\stackrel{-}{\mathrm{v}}$C + pron. suff. > CvCCA + pron. suff. & C$\stackrel{-}{\mathrm{v}}$CA + pron. suff.

In my data this phenomenon is poorly attested (see

Table 1. Trochaic occurrences in verbs.

yčammilen ‘they (f.) finish’ 3	yġammirin ‘they (f.) collect’ 1
ytālifu ‘they (m.) gathered’ 1	rākib-u ‘he mounted it’ 2

and

Table 2. Trochaic occurrences in nouns and prepositions.

ahl-a-ha ‘her family’ 6   waǧh-a-ha ‘her face’ 1   geddām-a-na ‘in front of us’ 1

ˁašān-a-ha ‘because of her’1   ˁomr-a-ha ‘her age’ 1

), and it only rarely appears in the speech of the oldest speakers:

Given the number of instances of this trait, it is possible to conclude that nowadays the dialect of the Bani ˁAbbād belongs to the atrochaic group even if the vestigial variants found in the corpus suggest that a trochaic rhythm characterized the vernacular at an earlier stage. This change in the rhythm of the syllable is most likely due to the influence of the surrounding sedentary dialects that display an atrochaic pattern.

3.1.2. Morphology

Personal Pronouns

The independent forms of personal pronouns in the dialect of the Bani ˁAbbād are illustrated in

Table 3. Independent personal pronouns.

1. s.	ʔana̴ ʔani	1. p.	ʔiḥna̴ ʔaḥna ̴ ʔəḥna
2. m. s.	ʔinta̴ ʔinte	2. m. p.	ʔintu
2. f. s.	ʔinti	2. f. p.	ʔintin
3. m. s.	huwwa̴ hū	3. m. p.	humma̴ humm
3. f. s.	hiyya̴ hī	3. f. p.	hinna̴ hinn ̴ hunna

:

It is possible to observe for the 1 s. an alternation between the forms ^ʔana̴^ʔani.

According to the data concerning the dialect of the ^ʕAǧārma, Palva (1976, p. 27) states: “it is impossible to decide whether [the first variant] is a traditionally genuine form or a loan from the neighbouring dialects [Bani Xaled, Sirḥān, etc.]”.

The second form, ^ʔani, is regarded by Younes and Herin (2016, p. 4) and Isaksson (1999, p. 59) as characteristic of the šāwi tribes, in free variation with the form ^ʔāni, already mentioned by Cantineau (1937, p. 173). Behnstedt (1997, p. 501), B. Herin (2010, p. 46) and Al Tawil (2019, p. 138) report that this form is commonly found in the Ḥōrān (both Syrian and Jordanian) and that it also marginally appears in the dialect of Salṭ.

As for the dialect of the ^ʕAbābīd, it appears only in sporadic instances:

-

in the negative copula māni ´I am not´ (6)9;

-

in the expression: māni ^ʕārif ´not knowing´(1), ^ʔana, māni sāmi^ʕ ´not listening´(1), ^ʔani ma ḥibbo al-gatū ´I don´t like the cake´ (1).

The 1 p. also exhibits two variants: ^ʔaḥna ̴ ^ʔəḥna (49) and ^ʔiḥna (20). The form ḥənna attested in the neighboring tribe of the ˁAǧārma and phonetically associated with the gahawa syndrome is not attested. Palva (1976, p. 26) affirms that the form ^ʔəḥna is “a stylistic variant” of ḥənna, and that their respective use depends on the form found in the tribe to which the ˁAǧārma address their speech.

It is possible to suppose that ḥənna may also have been employed in the dialect of the Bani ˁAbbād but that due to the contact with the adjacent sedentary dialects it disappeared, while the borrowing ^ʔiḥna penetrated in the vernacular and started to become more popular, especially among younger speakers.

Regarding the bound pronouns, it is worth mentioning that due to the recessive character of -č, the original šāwi morphological contrast between the pronominal suffix of the 2nd m. sing. person -k and the pronominal suffix of the 2nd f. sing. person -č, is not consistent.

The occurrences of the feminine bound pronouns’ forms -eč and -č are in total 12. The only instance that occurs in free speech is xuwayāteč ´your (f.) friends (m.)´ (2), while the other words containing this bound pronoun are to be found in optative sentences, in a few words occurring in some lines of poems and in some traditional quotes: salāma tsallmeč ‘(may you) live in good health’ (1), (Aḷḷah) yis^ʕid ^ʕumreč ‘may God make you (f.) happy for all your (f.) life’ (1), salāmteč ‘I wish you (f.) good health’(1), ^ʕammeč ‘your (f.) paternal uncle (1)’, yxāfeč ‘you (f.) are afraid’ (3), la-ḥāleč ‘alone (f.)’ (1), ^ʕīneč ‘your (f.) eye’ (1), ^ʔaḥebbeč ‘I love you (f.)’ (1).

Thus, even if as noted by Palva (1976, p. 47), “there is [still] a (bedouinizing) tendency which is actualized in certain speech situations associated with traditional culture”, in this case religious formulae, poetry and citations of famous quotes, the allomorph -č is nowadays being replaced by the sedentary form -ki.

Interrogative Pronouns

The interrogative pronouns occurring in the data collected are reported in

Table 4. Interrogative pronouns.

Who…?	mīn; min
What…?	šū ̴ šu; wēš; əš ̴ʔēš
Which…?	ʔayy(a)
How…?	čīf; šlōn; kēf ̴ kīf
Where…?	wēn
When…?	mata; mita; ʔēmta; waymat
Why…?	lēš; lwēš
How much…?	gaddēš ̴ geddēš; čam; kam

:

The traditional Bedouin forms (in bold in Table 4) are very marginal in the dialect of the Abābīd: min (3), wēš (1), čīf (1), lwēš (6), and they are to be found only in the recordings of the older speakers.

The variants mita; ^ʔēmta; waymat for ´when´and čam for ´how much´ were obtained only through elicitations and were never used by the speakers in natural speech.

As for šlōn, its use seems to be limited to the question šlōn-ak/ič? ´how are you´ and it does not occur in other positions.

Sedentary forms (which are to be considered as loans from the dialects of ^ʕAmmān and Salṭ) are the most attested ones in the corpus, and in many cases, they have already completely replaced their genuine nomadic counterparts.

Verbs C1=ˀ

The weak verbs C1=ˀ ˀaxaḏ-yāxuḏ ‘to take’ and ˀakal-yākul ‘to eat’ are as a rule reinterpreted as III w/y verbs in šāwi dialects (Younes and Herin 2016, p. 10) and realized as kala (ou čala) et xaḏa.

In respect to this trait, Cantineau (1936, p. 87) writes: ”cette conjugaison caractérise d’une façon remarquable les parlers de nomades moutonniers et les oppose aux parlers de sédentaires syro-palestiniens, qui ont toujours ^ʔakal, ^ʔaxaḏ”.

However, as illustrated in

Table 5. Inflexion of perfective.

1. sg. xaḏēt ̴ ʔaxaḏt	kalēt ̴ ʔakalt	1. pl. xaḏēna ̴ ʔaxaḏna	kalēna ̴ ʔakalna
2. m. sg. xaḏēt ̴ ʔaxaḏet	kalēt ̴ ʔakalet	2. m. pl. xaḏētu ̴ ʔaxaḏtu	kalētu ̴ ʔakaltu
2. f. sg. xaḏēti ̴ ʔaxaḏti	kalēti ̴ ʔakalti	2. f. pl. xaḏēten ̴ ʔaxaḏten	kalēten ̴ ʔakalten
3. m. sg. xaḏa ̴ ʔaxaḏ	kala ̴ ʔakal	3. m. pl. xaḏu ̴ ʔaxaḏu	kalu ̴ ʔakalu
3. f. sg. xaḏat ̴ ʔaxaḏat	kalat ̴ ʔakalat	3. f. pl. xaḏēn ̴ ʔaxaḏen	kalēn ̴ ʔakalen

and

Table 6. Inflexion of imperfective.

1. sg. āxuḏ ̴ ōxuḏ	ākil ̴ ōkil	1. pl. nāxuḏ ̴ nōxuḏ	nākul ̴ nōkil
2. m. sg. tāxuḏ ̴ tōxuḏ	tākul ̴ tōkil	2. m. pl. tāxḏu ̴ tōxḏu	tāklu ̴ tōklu
2. f. sg. tāxdi ̴ tōxḏi	tākli ̴ tōkli	2. f. pl. tāxḏin/en ̴ tōxḏin	tāklin/en ̴ tōklin
3. m. sg. yāxuḏ ̴ yōxuḏ	yākul ̴ yōkil	3. m. pl. yāxḏu ̴ yōxḏu	yāklu ̴ yōklu
3. f. sg. tāxuḏ ̴ tōxuḏ	tākul ̴ tōkil	3. f. pl. yāxḏin/en ̴ yōxḏin	yāklin/en ̴ yōklin

, the dialect of the Bani ^ʕAbbād shows the coexistence of two patterns for these two verbs, in both perfective and imperfective: one typically Bedouin (xaḏa–kala/yāxuḏ-yākul) and one sedentary (^ʔaxaḏ- ^ʔakal/yōxuḏ-yōkil).

The forms reported by Cantineau (1936, p. 87) were fully confirmed by elicitation (even with some minor differences), while the instances contained in the recorded corpus are mixed:

(2)	ʕalwēh ani min ʕarab	I wish I was of the people
	Hadla u la min xaḏa garāyibha	of Hadla and not of those who took her relatives
	La (a)ḥuṭṭ ʔana min aḏ- ḏahab ganṭār	so I [could] put a quintal of gold
	W aṭlaʕ min al-lōm ʔaxṭobha	and get rid of the blame and betroth her
(3)	rāḥ ʕand aš-šēx  ṭalabha, iši ḥabbha al-muhimm, banāt garāybo w an-nās illi yagrabūlo mā  ḥabbu inno yḥibbha laʔanno  ṯāri u mrattab  ḥabbu inno yōxiḏ  minhum u mā xaḏa minhum.

He went to the sheikh to ask for her hand, what is important is that he loved her, the girls of his relatives and the people who were related to him, did not like [the fact] that he loved her, because he was rich and wealthy, [so] they wanted him to take one of them, [but] he did not choose one of them.

(4)	axaḏu	humma	al-ǧanūb	kāmel
	to_take.PFV.3mpl	they.m	ART-south	complete
	They took over the whole South.

(5)	gabel	al-ʕeǧib	bass	yākol	xobez
	before	ART-children	but	to_eat. IPFV.3msg	bread
	min	ʕa- ṣ- ṣayǧān	maʕa	šāy
	from	on-ART-ṣayǧān	with	tea
	Before the children only ate bread made on the ṣāǧ10 with tea.

(6)	al-aġnām	yōklu	min-ha	akṯar	xuḏār	mā	fi
	ART-sheep	to_eat.IPFV.3mpl	of.3fsg	most	vegetables	NEG	there were
	They ate mostly sheep (meat), there were no vegetables.

The traditional paradigms still represent the majority of occurrences in the texts (52.5%), however, it is possible to observe a growing use of the sedentary forms (47.5%).

It is interesting to notice that the morphological variation between the two forms occurs only in natural speech but not when reporting some highly traditional tales or poems.

In this respect, Henkin (2010, p. 219) reports: “of all the oral registers, it is vernacular that reacts most significantly to dialectal and demographic variables […]. In contrast, the traditional registers of oral literature, including oral narrative and oral poetry […] are less affected by everyday communicational needs”.

The use of one variety over the other depends on sociolinguistic factors, namely age and level of education. The use of sedentary paradigms is increasing, and they are replacing the traditional forms mostly in the speech of the youngest generations, who have greater contact with the inhabitants of the biggest urban centers.

Prepositions

The preposition ˁugub ‘after’ represents a vestigial variant in the dialect of the Bani ˁAbbād, who nowadays use the form baˁd.

According to Herin (2010, p. 132) “ dans le parler traditionnel, baˁd- signifie ‘encore’ et non ‘après’ qui est une influence des parlers urbains (^ʕAmmān et Palestine)”.

In the corpus under analysis, ˁugub occurs only twice, in the speech of the oldest people interviewed, i.e., a woman and a man of 90 and 94 years old, respectively:

(7)	hāḏ̣	faddān	ad-darrāsāt	ʕugub
	DEM.msg	pair of oxen	ART-harvester	after
	gabel	yudrusu	ʕa-l-ḥamīr	w	al-bugar
	Before	to_thresh.IPFV.3mpl	on-ART-donkeys	and	ART-cows
	After one threshed with a pair of oxen, [but] before with some donkeys and cows.

(8)	Mīn	ḥakam	ʕugubu-hum	al-inglīz
	who	to_rule.PFV.3msg	after-3mpl	ART-English
	Who ruled after them? The English.

This preposition was probably more widespread some decades ago when its sedentary counterpart ba^ʕd had not penetrated so deeply in the vernacular. The current status of ˁugub shows once again the linguistic changes in progress in this dialect, in the direction of a de-bedouinization of the most conservative and traditional forms, which facilitates communication with those outside the tribe.

Conditional Conjunctions

The most frequently attested conditional conjunction in the data collected is ˀiḏa (39):

(9)	ʔiḏa	ʔiǧa	ḏ̣ēf	min	barra
	if	to_come.PFV.3msg	guest	from	outside
	ma	ysawwū-lo	akel	ˁādi	yiḏbaḥū-lo
	NEG	to_make.IPFV.3mpl-for-3msg	food	simple	to_slaughter.IPFV. 3mpl-for-3msg
	xarūf	yiḏbaḥū-lo	naʕǧe	ʕanz
	ram	to_slaughter.IPFV. 3mpl-for-3msg	sheep	goat
	If a guest comes from far away, one does not prepare for him a normal meal, one slaughters for him a ram or a sheep or a goat.

(10)	ʔiḏa	enta	ma	hazzēt
	if	you	NEG	to_shake.PFV.2msg
	al-finǧān	yḏ̣all	yṣubbū-l-ek
	ART-cup	to_keep.IPFV.3msg	to_pour.IPFV.3msg-for-2msg
	If you don’t shake the cup, one keeps pouring you (coffee).

However, this form does not belong to the original ^ʕAbbādi repertoire, but it is a progressive sedentary variant that supplanted the traditional form to introduce conditional clauses, čān, which has never been employed by the speakers.

3.1.3. Syntax

Genitive Exponent

According to Younes and Herin (2016, p. 12) the most frequent form of genitive exponent found in most Šāwi dialects is giyy.

This form, first reported by Cantineau (1946, p. 204), was also attested by Cleveland (1963, p. 61), who defined it as characteristic of the Jordanian Bedouin dialects.

In the dialect of the ^ʕAbābīd there are no instances of this local variant, and the only occurrences of a genitive exponent in natural speech is the more general and widespread Levantine taba^ʕ, which has to be regarded as a borrowing from the adjacent urban dialects:

(11)	kānu	yḥarrkū	haḏ̣ōla
	to_be.PVF.3mpl	to_stir.IPVF.3mpl-3msg	DET.pl
	bi-l-maġrāfe	tabaʕat11	al-xašab
	in-ART-ladle	tabaʕ.fsg	ART-wood
	They stirred those with the wooden ladle

(12)	al-gider	ṭanǧara	kbīra	yḥuṭṭu
	ART-cauldron	pan	big	to_put.IPVF.3mpl
	bī-ha	al-ǧarīša	tabaʕat	al-gameḥ
	in-3fsg	ART-grain	tabaʕ.fsg	ART-wheat
	The cauldron is a big pan where they put the grain of wheat

The forms giyy and šiyy12 were productive at an earlier stage, as reported by a speaker of ^ʕArāg al-^ʔAmīr, but today they are only used in the field of trade:

(13)	Dār gīti dār tabaʕti, hal-kalime kānat mawǧūda, gīti aw šīti, yaʕni mulki w ili, ā zayy hēk ā, al-banāt giyyāti aw ḥalālāti yaʕni ili  ḥatta ʕan al-ġanam ygūlu-lo giyyāti, kānat mawǧūde hassāʕʕ bistaʕmalūha at-tuǧǧār ʕenna, maṯalan ygūl lak ʔana gayyāti 3 alāf ʔana gayyāti 10 alāf, yaʕni flūsi.

Dār gīt13i or dār taba^ʕti ‘my house’, gīti or šīti, this word existed, it means of mine, mine, like this yes, al-banāt giyyāti ‘my daughters’ or ḥalālāti ´my cattle´ for example, one also used to say al-ġanam giyyāti ‘my sheep´, this existed but now businessmen use it, to say my 3 or 10 thousand, to indicate my money.

It is possible to conclude that the progressive variant taba^ʕ superseded the equivalent traditional forms of genitive exponents in the dialect of the Bani ^ʕAbbād.

A similar scenario is also attested by Procházka (2018), Younes (2014) and Younes and Herin (2013) for the Šāwi Bedouin dialects spoken in Syria14 and for those spoken in Lebanon by the Abu ^ʕĪd and the ^ʕAtīǧ.

This state of the affairs confirms the analysis of Palva (1982, p. 28), who states that “the genitive exponent belongs to the features in the dialects that are particularly exposed to koineization”.

Negation

The negation in the dialect of the Bani ˁAbbād has also undergone a process of sedentarization.

It is possible to notice that the negative form miš has gained ground in the non-verbal negation at the expense of the genuine negative particles mā and mū (Palva 1976, p. 42; Al Tawil 2021, p. 23). These last variants represent, respectively, 16.45% and 13.92% of non-verbal negation in the data:

(14)	ʕašāʔir	al-ǧanūb	lāʔ	mā
	tribes	ART-South	NEG	NEG
	bēn-na	u	bēn-hum	mašākil
	between-1pl	and	between-3mpl	problems
	There were no problems between us and the tribes of the South.

(15)	gabel	yinsawi	šaḥge	mū	dabka
	before	to_make.PASS.3msg	šaḥge15	NEG	dabka
	Before one [used to] dance the šaḥge and not the dabka.

In total, 63.29% of this type of negation is constituted by the item miš and the remaining 5% by muš, which has to be regarded as an older borrowing as it is already attested in Ḥōrān (Cantineau 1946, p. 389):

(16)	hāy	əl-muṣṭalaḥāt	hāy	alān	ixtafat
	DEM.3fsg	ART-terms	DEM.3fsg	now	to_change.PFV.3fsg
	min	ǧīlna	miš	mawǧūde	yaˁni
	from	generation-1pl	NEG	present.3fsg	it is to say
	These terms here have now disappeared from our generations, they are no more present.

(17)	ḏahab	ṭalaʕ	zayy	ḥǧār	miš	ḏahab
	gold	to_appear.PVF.3msg	like	stones	NEG	gold
	The gold appeared in the form of stones not gold.

(18)	muš	al-kull	kwayyis	muš	al-kull	mirtāḥ
	NEG	ART-all	good	NEG	ART-all	wealthy
	We do not all have a good financial situation; we are not all wealthy

It is also interesting to point out that the negative structure mā fi-š began to spread in the speech of the youngest speakers of the tribe16:

(19)	mā	fi-š	maʕo	maṣāri
	NEG	there is-NEG	with-3msg	money
	He has no money.

(20)	mā	fi-š	maḥākim	mā	fi-š	muḥāmiyyīn
	NEG	there were-NEG	tribunals	NEG	there were-NEG	lawyers
	[Before] there were no tribunals nor lawyers.

According to Palva (1976, p. 42) the nominal negation miš and the structure mā fi-šš are to be considered as K-forms borrowed from the neighbouring sedentary dialects and not as genuine Bedouin negation.

However, the high frequency of the use of miš suggests that this negation has already been integrated into the dialect of the Bani ^ʕAbbād and that the other particles, i.e., mā, mū, which used to be employed in non-verbal negations, are to be regarded as vestigial variants.

b-Imperfect

Cantineau (1936, p. 83) affirms: “Le b-préfixe de l’inaccompli, si caractéristique des parlers des sédentaires syro-palestiniens, fait entièrement défaut dans les parlers des nomades”. This statement, however, does not hold true anymore for the dialect of the Bani ^ʕAbbād.

In fact, the use of the morpheme b- prefixed to the imperfect is no longer excluded from this Bedouin dialect, and it is possible to find 53 occurrences of this trait in the corpus:

(21)	humma	bistagbilu	aḏ̣-ḏ̣yūf
	they	to_welcome.IPFV.HAB.3mpl	ART-guests
	They welcome the guests.

(22)	bitḥuṭṭi	al-laḥme	w	al-snōbar
	to_put.IPFV.HAB.2fsg	ART-meat	and	ART-pine nuts
	You add the meat and the pine nuts.

(23)	ayyām-ha	ən-nās	mā	maʕā-ha
	days-DEM	ART-people	NEG	with-3fsg
	drāsa	mā	btaʕrif	iši
	education	NEG	to_know.IPFV.HAB.3fsg	thing
	In those days people were not educated, they didn’t know anything.

The adoption of the b-imperfect to mark habitual actions which take place in the present can be considered as an ongoing process in the dialect of the ^ʕAbābīd, and thus it is not rare to find some speakers using both forms (with and without b- prefix) in one and the same sentence:

(24)	Fa yitǧawwaz az-zalama…ˀiḏa kān fi ʿindhum dār biskin bi-nafs ad-dār ygullak waladna mā biṭlaʕ barra yaˁni muḥarram inno yiṭlaʕ yiskin xārǧ al-manṭega illi humma sāknīn fīha… bi- ẓ- ẓabṭ... hāy al- ʕāda.

Then the man gets married…if they have a house, he lives in the same household, one says: our boy does not go elsewhere. That is to say that it is forbidden (for him) to go live outside the area where they live… Exactly… this is the custom.

(25)	zulum	yṭlubu	al-ʕarūs	byaʕṭī-hum
	men	to_ask.IPFV.HAB.3mpl	ART-bride	to_give.IPFV.HAB.3msg-3mpl
	Some men ask for the hand of the bride, he grants it to them.

Thus, it seems that, at this stage, this syntactic feature is used in free variation with the imperfect without b-, like among the Negev and Sinai Bedouins (Palva 1994, p. 462; Blanc 1970; De Jong 2000).

The b-imperfect can also be used to mark the progressive aspect of an action. However, in the corpus under analysis there is only one instance of this type of structure:

(26)	bitḏakkar	waḥda	barḏ̣o	qiṣṣa	min	al-qiṣaṣ
	to_remember.IPFV.PROG.1sg	one.f	also	story	between	ART-stories
	I also remember another story [now].

This sedentary trait has penetrated mostly the speech of the younger members of the tribe who have the tendency to conform more to the vernaculars of ^ʕAmmān and Salṭ, and so it reveals “a generational variation amongst the receptiveness of the borrowings” (Younes 2017, p. 136).

Future

The most prominent way in Šāwi varieties to express future and volition is the use of the pseudo-verbs widd- and rād–y(i)rīd ´to want´17 (Younes and Herin 2016, p. 11). The first particle is well attested in the dialect of the Bani ^ʕAbbād while the verb rād is completely absent from it.

It is worth noticing that the pseudo-verb widd- occurs together with the sedentary variant bidd- all along the corpus:

(27)	wədd-ak	trūḥ	ʕ-al-ḥakīm
	want-2msg	to_go.IPFV.2msg	to-ART-doctor
	You will go to the doctor.

(28)	bidd-i	ōkil	tīn
	want.1sg	to_eat.IPFV.1sg	figs
	I will eat figs.

Since the number of instances of these two forms is almost equivalent18, it is possible to affirm that the Bedouin and sedentary varieties of this pseudo-verb coexist in this vernacular. However, their use remains dichotomous: it is considered as equally eligible by the speakers of the new generation, while the older ones (55 years and older) seem to prefer the genuine Bedouin item over the madani one.

4. Discussion

4.1. Bedouin or Sedentary?

In the light of the traits analyzed above, and the common division of the Arabic dialects available, the classification of the dialect of the Bani ^ʕAbbād is not straightforward.

According to the historical background and geographical position in the Balga district, it should typologically belong to the Bedouin Jordanian ygūl-type (part of the broader Šāwi dialects’continuum) or the recently theorized Central Bedouin ygūlu group. However, can it really be unproblematically labelled as such? Is it appropriate to still refer to this vernacular as a Bedouin dialect?

Such classifications which are based on typical, but nonetheless very few features, in fact do not account for all the discrepancies found in this vernacular and do not accurately allow to represent the actual state of the dialect.

As pointed out in the section Linguistic Analysis, this dialect diverges in many aspects from the “essential structure” (Palva 1969, pp. 14–15) of the ygūl group and it remains inaccurate to identify this dialect of the Jordan Valley with the šāwi label.

In fact, in addition to the elements already mentioned throughout the article, it is also possible to observe from Table 6 the lack of the formative /n/ in the 3. m. pl. of the imperfective. This constitutes a noticeable difference (yāxḏu vs. yāxḏūn) with the typical inflextion found in šāwi dialects (Younes and Herin 2016, p. 8) and adds supplementary intricacy in cataloguing this vernacular. Furthermore, other genuine šāwi characteristics have been replaced by sedentary forms, while some typically madani items are being acquired in addition to the original Bedouin ones.

All the features taken into account for this analysis hint at highlighting that this dialect is in a situation of dialect contact and in an ongoing process of linguistic change.

Holes (1995, p. 278) notes that during the last 70 years “Jordan has experienced a long-term drift from the countryside and the desert into its towns and cities”, and massive migrations from neighboring countries, namely Palestine and Syria. As in other Middle Eastern states, “these social and political developments have had, and continue to have profound effects in spoken Jordanian Arabic varieties” (Holes 1995, p. 278), which in some cases tend to overlap more and more.

As Romaine (2010, p. 321) states:

“When groups in contact need to communicate, they have a number of possible choices. One is to use a lingua franca they both share […]. A second option is for one or more parties to learn the other group’s language(s). In cases involving no substantial imbalances of power between the groups, stable multilingualism may result. However, where bilingualism is asymmetrical and the more powerful group imposes its language on a subordinate group, contact often leads to language shift or loss”.

The second scenario applies in this case. The members of the Bani ^ʕAbbād have day-to-day and long-lasting relations with the speakers of sedentary adjacent dialects and in order to facilitate the communication and mutual intelligibility avoid the use of traditional Bedouin features and adopt forms that can be easily understood by outsiders to the tribe19.

However, the linguistic patterns of accommodation and borrowings are not identical for all the people of the community (Palva 1982).

Traditional Bedouin features are disappearing faster from the speech of the young members of the tribe, and there was already a major tendency for women to make use of sedentary characteristics in the everyday vernacular 50 years ago (Palva 1976), since they appear more feminine and sophisticated than the respective Bedouin ones which relate to rough and tough masculinity (Holes 1995).

The borrowing of sedentary speech-habits does not mean that the Bani ^ʕAbbād “give up their language willingly, but continue transmitting [it], albeit in changed form over time” (Romaine 2010, p. 321).

The members of the tribe are aware of this state of the affairs and in fact one of the older members declares in one of the recordings:

-

Fi al-muṣṭalaḥāt hāḏ̣i ya^ʕni wlādna alli ^ʕomrhum ṯalāṯīn sana u xamsa u ^ʕašrīn sana mā kānu ya^ʕrifūha li^ʔannha iltaġat min zamān ā, iltaġat min zamān.

-

Ṭabb al-luġa al-^ʕabbādiyya ʕam bitrūḥ?

-

Mu^ʕḏ̣amha hī trūḥ hassā^ʕ ṣārat madaniyya ma^ʕḏ̣omha hassā^ʕ madaniyya… kull iši ^ʕindana ya^ʕni al-luġa al-^ʕabbādiyya mā ḏ̣all ma^ʕāha galīl ǧiddan.

-

There are these terms [that] our children whose age is 30 or 25 years old do not know because they changed a while ago, yes, changed [a lot of] time ago.

-

Well, so where is the ^ʕAbbādi language going?

-

Most of it is disappearing, now it has become madani, most of it has become madani…all of our things, I mean there is very little left of the ^ʕAbbādi language.

These changes by sedentarization20 result in a hybrid form of dialect which may be considered as a dialect compromise (Holes 1995) built on Bedouin and sedentary forms, which is not simple to define.

4.2. Emergence of a Mixed Dialect Type by Sedentarization

The members of the linguistic community in Jordan do not have equal command of all the varieties in use in the country. Thus, Bedouin speakers often accommodate to the more common speech patterns of the city dwellers for the sake of intelligibility. Over time this convergence materially affects the distinctiveness of a dialect in a situation of language contact (Gumperz 1969, p. 436)

Already in 1992 Palva illustrated the issue of the typological division into Bedouin and sedentary dialect types in Jordan basing his analysis on the vernaculars spoken in the cities of Salṭ and Karak: “both dialects are labelled as bəgūl dialects. […] However, as a matter of fact (they) display many typically Bedouin features, so markedly different from Syro-Palestinian sedentary dialects” (Palva 1992, pp. 53–54).

In this study of the dialect of the Bani ^ʕAbbād the problem is the reverse, i.e., a historically Bedouin-type dialect that shows traits of sedentary dialects, and that thus exhibits important markers of both dialect types.

Due to the new madani established features it is not possible to label this dialect only as Bedouin, and according to the retention of some šāwi traits it cannot be defined as a sedentary dialect either.

The traditionally dichotomous split of Arab dialects in terms of Bedouin versus sedentary (Versteegh 1984; Rosenhouse 1984, 2006; Cadora 1992), is no longer sufficient to describe and classify those varieties like the Bani ^ʕAbbād dialect, i.e., languages which have experienced relatively high degrees of contact to the extent that change is additive (Trudgill 2010, p. 301).

As pointed out by Lentin (1994) and Watson (2011), the mere use of this bipolar division is inaccurate and gives rise to a number of difficulties especially when classifying dialects in a situation of language contact.

Due to the linguistic developments that occurred in the dialect of the Bani ^ʕAbbād it seemed functional to define it as a mixed type, in order to better underline its particular nature.

However, the emergence of mixed type dialects, such as this one, brings attention to the need to conceptualize new ways of grouping dialects and that further criteria of analysis need to be adopted, especially morphological and syntactic ones.