https://www.openbookpublishers.

com

©2024 Fabio Gasparini, Kamala Russell and Janet C. E. Watson. Copyright of individual
chapters are maintained by the chapter author(s).

This work is licensed under an Attribution-NonCommercial 4.0 International (CC BY-NC

4.0). This license allows you to share, copy, distribute, and transmit the text; to adapt
the text for non-commercial purposes of the text providing attribution is made to the
authors (but not in any way that suggests that they endorse you or your use of the
work). Attribution should include the following information:

Fabio Gasparini, Kamala Russell and Janet C. E. Watson, Diversity across the Arabian
Peninsula: Language, Culture, Nature. Cambridge, UK: Open Book Publishers, 2024,
https://doi.org/10.11647/OBP.0411

Further details about CC BY-NC licenses are available at

http://creativecommons.org/licenses/by-nc/4.0/

All external links were active at the time of publication unless otherwise stated and
have been archived via the Internet Archive Wayback Machine at
https://archive.org/web

Any digital material and resources associated with this volume will be available at
https://doi.org/10.11647/OBP.0411#resources

Semitic Languages and Cultures 28

ISSN (print): 2632-6906

ISSN (digital): 2632-6914

ISBN Paperback: 978-1-80511-337-9

ISBN Hardback: 978-1-80511-338-6
ISBN Digital (PDF): 978-1-80511-339-3

DOI: 10.11647/OBP.0411
Cover image: Photo by Rabah Al Shammary, titled ‘Wild Acacia tree, Ha’il, Arabian
Peninsula’, July 26, 2021; https://unsplash.com/photos/green-tree-on-brown-sand-
under-blue-sky-during-daytime-e-UPgjjEwCM?utm_content=creditCopyText&utm_
medium=referral&utm_source=unsplash.
Cover design: Jeevanjot Kaur Nagpal

The main fonts used in this volume are Charis SIL and Scheherazade New.
 ASPECTS OF THE PHONOLOGY AND
MORPHOLOGY OF SAUDI VARIETIES OF
ARABIC

Stuart Davis, Wafi Alshammari, Musa Alahmari, and

Mamdouh Alhuwaykim

1.0. Introduction
Various Saudi subvarieties of Arabic are known to display certain
unusual linguistic features with respect to aspects of their pho-
nology and morphology. Such features may be archaic features
of Arabic that have long disappeared in other varieties or sub-
strate features, but other unusual features may be innovations.
An example of an archaic feature is the persistence of a lateral
fricative pronunciation of historical ḍād still found in the south-
western area of ʿAsīr and Saudi Tihāmah, as documented in such
works as Watson and Al-Azraqi (2011) and Al-Wer and Al-
Qahtani (2016). This is best understood as an archaic feature that
is consistent with Sībawayh’s description of the sound (Watson
and Al-Azraqi 2011, 426). An example of a substrate feature is
the nasal definite article which occurs in Faifi Arabic (Alfaifi and
Behnstedt 2010; Alfaifi and Davis 2021) and other varieties of
southwestern Saudi Arabic (Prochazka 1988; Behnstedt 2016).

© 2024 Stuart Davis, et al., CC BY-NC 4.0 https://doi.org/10.11647/OBP.0411.04

138 Davis, Alshammari, Alahmari, and Alhuwaykim

Watson (2018) views the nasal definite article in these varieties

as reflecting a non-Arabic Semitic substrate (see also Al-Jallad
2021). While various Saudi varieties preserve ancient features,
innovation is also present. One example of an innovation in some
Saudi varieties is the presence of word-initial consonant clusters.
While some varieties, such as Urban Hijazi Arabic, do not allow
initial clusters, quite a few varieties of Saudi Arabic do allow such
clusters, such as Najdi Arabic (Abboud 1979; Ingham 1994) and
the southwestern Saudi dialect described in Alahmari (2018).
That the presence of word-initial clusters is an innovation can be
gleamed from the observation that they may be quite limited in
some varieties, as in a subvariety of Faifi Arabic discussed in
Alfaifi and Davis (2021); word-initial clusters seem to first arise
diachronically from the extension of the process of high vowel
deletion to word-initial syllables, an environment where deletion
is blocked in many dialects (such as in Urban Hijazi).
With this as background, in this chapter we will highlight
some unusual features of the phonology and morphology of var-
ious Saudi varieties of Arabic that are either unexamined or have
not been studied in detail, and to consider whether they are ar-
chaic features or reflect internal innovation. An important aspect
of our article, though, will be the detailed description of the phe-
nomena that we will be considering. The descriptions are based
on native speaker intuition in consultation with other speakers of
the same dialect. In §2.0 of this chapter we focus on an unusual
productive morphological augmentative witnessed in Haʾili Ara-
bic of the northern Najdi region (Alshammari and Davis 2019),
which includes forms like klɑɑb ‘dog (augmentative)’ for kalb
 Phonology and Morphology of Saudi Varieties of Arabic 139

‘dog’ (where bold indicates pharyngealisation), šwɑɑʿir ‘poet

(augmentative)’ for šaaʿir ‘poet’, and srɑɑwiil ‘pants (augmenta-
tive)’ for sirwaal ‘pants’, where augmentative indicates largeness
and sometimes awkwardness (as in Sifianou 1992). The Haʾili Ar-
abic pattern will be described in detail, focusing on its morpho-
phonological realisation. In §3.0 we describe and analyse the 2nd
person masculine singular possessive pronoun, focusing on a
southwestern Arabian variety (Alahmari 2018). In this variety,
the 2nd person masculine singular possessive suffix alternates be-
tween -k, -ka, and -ak as witnessed by examples such as kutub-k
‘your books’, malik-ka ‘your king’, and baab-ak ‘your door’. As far
as we are aware, this three-way allomorphic variation is unusual.
In §4.0, we discuss final degemination along with stress shift,
which appears to be fairly widespread among Najdi and north-
western Saudi Arabic varieties, especially as witnessed in disyl-
labic words that end in an underlying geminate, as illustrated by
the difference between ḥagg ‘a truth’ and íl-ḥag ‘the truth’, with
stress on the definite article, or in the elative, as exemplified by
áxaf ‘lighter’ with initial stress, where most dialects have axáff.
Focusing on a dialect spoken in Sakaka City (Alhuwaykim 2018),
the details of the degemination pattern are presented and ana-
lysed, including the lack of degemination (and stress shift) in de-
rived contexts (e.g., sikát-t ‘I kept silent’). We then analytically
unify degemination with the occurrence of word-final vowel
shortening (and stress shift) in disyllabic words that end in a long
vowel. We conclude in §5.0 with a summary of the paper.
140 Davis, Alshammari, Alahmari, and Alhuwaykim

2.0. The Morphological Augmentative in Northern

Najdi/Haʾili Arabic
In this section we describe in detail a unique productive morpho-
logical augmentative pattern found in Haʾili Arabic as spoken by
the Shammar group. The Shammar group falls under the subdia-
lect of Northern Najdi in the division of Najdi Arabic found in
Ingham (1994). The data presented in this section are mostly
from Alshammari and Davis (2019) and are supplemented by the
native intuition of the first author of that work (who is the second
author of the present chapter). A pertinent discussion of the mor-
phological augmentative can be found in Assuwaida (1997), who
relates the formation of the augmentative to the older dialect of
Tayy, which goes back to the pre-Islamic era. The uniqueness of
the Tayy dialect is emphasised by Al-Jallad (2013) and the dis-
tinctiveness of the Arabic of the Shammar group from other
Northern Najdi varieties is emphasised by Ingham (1982).
In presenting the morphological augmentative we will also
show the corresponding diminutive forms. Haʾili Arabic possesses
a morphological diminutive and there is an analytical issue as to
whether the Haʾili augmentative is derived directly from the cor-
responding diminutive or, like the diminutive, is derived directly
from a corresponding base noun. That is, for example, is the aug-
mentative for ‘dog’, klɑɑb, derived from the diminutive kleyb or
from the noun kalb ‘dog’? In our description of the augmentative,
though, we will maintain that the diminutive serves as the base
for the augmentative and we will provide reasons for this posi-
tion. The issue of the relationship between an augmentative and
a diminutive is briefly discussed by Sifianou (1992, 157) for
 Phonology and Morphology of Saudi Varieties of Arabic 141

Modern Greek, where she mentions that the occurrence of an

augmentative in a language suggests the presence of a diminu-
tive. The diminutive in Haʾili Arabic is similar to the pattern
found in Classical Arabic and in other varieties of Peninsular Ar-
abic (e.g., Holes 1984), but with some diachronic phonological
changes largely predictable based on the Classical Arabic dimin-
utive; see Zewi (2006) and Gadoua and Davis (2019) for over-
views of the diminutive across different Arabic varieties and see
Alshammari and Davis (2019) for a description of the diminutive
in Haʾili Arabic.
In (1)–(4) of Table 1 below, we show the diminutive and
augmentative of words whose base form is a monosyllabic noun.
The base noun is given in the leftmost column with the word
pattern indicated as either CVCC in (1), CVGG in (2), where GG
is a geminate, CVVC in (3), and CCVVC in (4); the diminutive is
shown in the second column, the corresponding augmentative in
the third column, and the English gloss in the rightmost column
(note also that bold indicates pharyngealisation; a full stop indi-
cates a syllable boundary; long vowels are transcribed as a se-
quence of two identical vowel letters).
142 Davis, Alshammari, Alahmari, and Alhuwaykim

Table 1: Diminutives and Augmentatives of monosyllabic word forms

(1) CVCC word Diminutive Augmentative Gloss
(a) kalb kleyb klɑɑb ‘dog’
(b) gird greyd grɑɑd ‘monkey’
(c) wajh wjeyh wjɑɑh/wjaah ‘face’
(2) CVGG word Diminutive Augmentative Gloss
(a) xadd xdeyd xdɑɑd ‘cheek’
(b) ḥagg ḥgeyg ḥgɑɑg ‘right’
(3) CVVC word Diminutive Augmentative Gloss
(a) bɑɑb bweyb bwɑɑb ‘door’
(b) nɑɑr nwey.r-ah nwɑɑ.r-ah ‘fire’
(c) dɑɑr dwey.r-ah dwɑɑ.r-ah ‘clay house’
(d) rɑɑs rweys rwɑɑs ‘head’
(e) fɑɑs fweys fwɑɑs ‘hatchet’
(4) CCVVC word Diminutive Augmentative Gloss
(a) grɑɑd grɑy.yid grɑɑ.yid ‘tick’
(b) ḥjaab ḥjay.yib ḥjɑɑ.yib ‘veil’
(c) ftaat ftay.yit — ‘crumbs’

The augmentative of monosyllabic nouns in (1)–(4) seems to be

based on the diminutive form, with the diphthong in the first
syllable of the diminutive (ey or ay) simply replaced with the long
pharyngealised vowel that we transcribe as long ɑɑ. The remain-
der of the augmentative is exactly the same as the diminutive,
although pharyngealisation from the augmentative vowel may
show limited spreading onto neighbouring phonemes, as indica-
ted by the bold in the transcription (the details of pharyngealisa-
tion spread in Haʾili Arabic is left for future research). The argu-
ment for viewing the augmentative as deriving from the diminu-
tive is based not only on the simplicity of the description, but
also on the observation that irregularities of the diminutive are
encountered in the augmentative, as seen in (3b) and (3c), where
the same suffix that occurs in the diminutive also occurs in the
augmentative. Furthermore, there occur diminutive forms wi-
thout a corresponding augmentative, as in (4c), but not vice-
 Phonology and Morphology of Saudi Varieties of Arabic 143

versa. It should be noted that the presence of pharyngealisation

in the augmentative sometimes leads to phonological contrast, as
in the pair klaab ‘dog, plural’ versus klɑɑb ‘dog, augmentative’,
but sometimes to homophony, as in the example grɑɑd ‘monkey
augmentative’ in (1b), which is also the noun meaning ‘tick’ seen
in (4a).
The next set of data in (5)–(11) of Table 2 illustrates the
augmentative (and diminutive) forms for a wide variety of differ-
ent types of disyllabic nouns (note: * indicates an ungrammatical
form).
Table 2: Diminutives and Augmentatives of disyllabic word forms
(5) CV.CVC word
Diminutive Augmentative Gloss
(a) gɑ.lɑm gleym glɑɑm ‘pen’
(b) ʿi.jil ʿjeyl ʿjɑɑl/ʿjaal ‘calf’
(6) CV.CVC word
Diminutive Augmentative Gloss
(a) šaa.ʿir šwey.ʿir šwɑɑ.ʿir ‘poet’
(b) naa.g-ah nwey.gah/ nwɑɑ.gah/ ‘camel (F)’
nwey.dzah *nwɑɑ.dzah
(7) CVC.CVC word Diminutive Augmentative Gloss
(a) dir.hɑm drey.him drɑɑ.him ‘Dirham’
(b) ʿan.bɑr ʿney.bir ʿnɑɑ.bir/ ‘dungeon’
ʿnaa.bir
(c) ʿil.b-ɑh ʿley.b-ah ʿlɑɑ.b-ɑh ‘can’
(8) CV.CVVC word Diminutive Augmentative Gloss
(a) ki.biir/ kbɑy.yir/ kbɑɑ.yir/ ‘big’
tsi.biir tsbɑy.yir *tsbɑɑ.yir
(b) ki.taab ktay.yib ktɑɑ.yib ‘book’
(c) ṣi.ġiir ṣġɑy.yir — ‘small’
(d) ḥa.nuun ḥnɑy.yin — ‘kind, compassionate’
(9) CV.C-ah word Diminutive Augmentative Gloss
(a) sa.n-ah snay.y-ah. snɑɑ.y-ah/ ‘year’
snaa.y-ah
(b) ma.r-ah mray.y-ah mrɑɑ.y-ah ‘woman’
144 Davis, Alshammari, Alahmari, and Alhuwaykim

(10) CVC.CVVC word Diminutive Augmentative Gloss

(a) sir.waal srey.wiil srɑɑ.wiil ‘pants’
(b) ṣil.ṭɑɑn ṣley.ṭiin ṣlɑɑ.ṭiin ‘Sultan’
(c) xab.bɑɑz xbey.biiz xbɑɑ.biiz ‘baker’
(d) maj.nuun mjey.niin mjɑɑ.niin/ ‘crazy’
mjaa.niin
(11) CCV.C-ah word Diminutive Augmentative Gloss
(a) nxa.l-ah nxey.l-ah nxɑɑ.l-ɑh ‘palm tree’
(b) wri.g-ah wrey.g-ah/ wrɑɑ.g-ah/ ‘paper’
wrey.dz-ah *wrɑɑ.dz-ah

Again, as in Table 1, the augmentative is formed from the dimi-

nutive by replacement of the diphthongal nucleus of the first syl-
lable of the diminutive with the long pharyngealised vowel ɑɑ,
which triggers subsequent pharyngealisation spread. Concerning
the optionality of pharyngealisation in the augmentative in the
forms in (5b), (9a), and (10d), as well as in (1c) above, pharyn-
gealisation can be optionally blocked in the presence of a palatal-
type consonant—although this does not account for the optiona-
lity shown in (7b). The data items in (6b), (8a), and (11b) are
interesting in that these words show optional affrication of velar
stops in the diminutive; the corresponding augmentative cannot
surface with affrication. This suggests an antagonistic relation-
ship between affrication and pharyngealisation. With respect to
semantics, examples like (8c) and perhaps (8d) show that a word
signifying smallness cannot be made into an augmentative, but
(8a) shows that a word that expresses largeness can have a dimi-
nutive form. This difference supports the observation that the
augmentative is based on the diminutive (though semantics may
also contribute to this difference).
A final set of data of base words consisting of three syllables
is shown in (12) below in Table 3, where the augmentative seems
 Phonology and Morphology of Saudi Varieties of Arabic 145

to be formed from the diminutive by replacing the diphthongal

nucleus of the first syllable with the long pharyngealised vowel
ɑɑ and the remaining syllables preserve identity with the corre-
sponding syllables of the diminutive.
Table 3: Diminutives and Augmentatives of trisyllabic word forms
(12) Trisyllabic word Diminutive Augmentative Gloss
(a) siy.yaa.rah swey.rii.r-ah swɑɑ.rii.r-ah ‘car’
(b) mir.jey.ḥah mrey.jii.ḥ-ah mrɑ(ɑ).jii.ḥ-ah ‘swing’
(c) ʿan.kə.buut ʿney.ki(i)b ʿnɑɑ.ki(i)b/ ‘spider’
ʿnaa.ki(i)b
(d) bar.naa.mij brey.nii.mij brɑɑ.nii.mij ‘programme’

The presence of a morphological augmentative in Haʾili Arabic

raises a number of questions. There is a question of the conside-
ration of the semantics and pragmatics of the augmentative in a
more thorough manner than what we present here. Also, there is
the issue of the morphological relationship between the augmen-
tative and the diminutive, where we maintain that the former is
based on the latter, but this would require further and more de-
tailed argumentation than what we have presented. The question
that we want to briefly address here concerns the matter of
whether the occurrence of a productive morphological augmen-
tative is a recent innovation or a preservation of a truly archaic
feature of Haʾili Arabic. Augmentatives are rarely discussed in
Arabic, since they are typically expressed periphrastically, using
a phrase of two or more words as in the Standard Arabic example
bayt kabiir ‘a big house’. Assuwaida (1997) argues for the anti-
quity of the augmentative in Haʾili Arabic by providing evidence
that it is part of the older dialect of Tayy (Haʾil region, modern-
day Saudi Arabia), going back to the pre-Islamic era. He provides
evidence from poetry and place names that display the
146 Davis, Alshammari, Alahmari, and Alhuwaykim

augmentative pattern. Two examples that he gives are ʿuwaɑriḏ̣

and tuwaɑrin (although we note that in the contemporary dialect
the /u/ in the first syllable would not be pronounced, given the
diachronic loss of unstressed high vowels in initial syllables). As-
suwaida (1997) notes that ʿwaɑriḏ̣ is the name of a lone-standing
mountain located in front of twaɑrin, which is a castle in the Aja
mountains in Tayy. Assuwaida’s discussion leads us to conclude
that the morphological augmentative is both an archaic feature
of Haʾili Arabic as spoken by the Shammar group, who descend
from the ancient Arab Tayy tribe, (Al Rasheed 1991; Fattah and
Caso 2009) and a feature that has probably always been geo-
graphically limited within the Arabian Peninsula. It is possible
that it historically derived from the diminutive in pre-Islamic ti-
mes as an ancient innovation, but its persistence into a contem-
porary dialect is truly noteworthy. (see Goitein 1960 for the oc-
currence of diminutives and augmentatives in Yemeni Jewish
Arabic, where diminutives can be used with augmentative mea-
ning).

3.0. The Possessive Suffix /-k/ in a Southwestern

Dialect of Saudi Arabic
The masculine singular second person possessive (or genitive)
suffix has the reflex -ka in standard varieties of the language,
such as Classical Arabic and Modern Literary Arabic (Cowan
1958), as exemplified by the word kitaabu-ka ‘your (MS) book
(NOM)’. In dialects, such as Cairene Arabic, that lack case endings,
the realisation of this possessive suffix shows allomorphy be-
tween -ak and -k, where the latter occurs only after words that
 Phonology and Morphology of Saudi Varieties of Arabic 147

end in a vowel, such as abuu-k ‘your father’ or ʿašaa-k ‘your din-

ner’; all other forms take -ak, such as balad-ak ‘your country’, rigl-
ak ‘your leg’, and kitaab-ak ‘your book’. The allomorph -ak is also
suffixed to feminine words that end in tā marbūṭa, such as ṭɑɑlib-
t-ak ‘your student (F)’, whose non-possessive realisation is ṭɑɑliba.
The most straightforward analysis of the allomorphy in a dialect
like Cairene is to posit /-ak/ as the underlying form of the 2nd
person masculine singular possessive suffix and to have a rule
that deletes the /-a/ of the suffix when it occurs immediately af-
ter a vowel. Feminine words that end in an orthographical tā
marbūṭa would be analysed as either ending in an abstract /t/ or
in a t introduced as part of the possessive construction of femi-
nine word forms.
When we turn to Saudi varieties of Arabic, the allomorphy
between -k and -ak (or -ik) has been discussed by Ingham (1982;
1994), especially with respect to Central Najdi Arabic. Here we
find a pattern and an analysis that is quite different from the Cai-
rene pattern briefly described above. Ingham (1982, 96) notes
that in the central area of the Najdi Arabic region, suffixes like /-
ak/ lose their vowel when they follow a syllable ending in VC
(i.e., a short vowel followed by a single consonant). Examples
include naxal-k ‘your (MS) palms’ and walad-k ‘your (MS) son’,
where the suffix is just -k, but in other phonological environ-
ments the vowel surfaces, as in beet-ak (or beet-ik, depending on
the subvariety) ‘your (MS) house’. This differs not only from Cai-
rene Arabic, but also from Urban Hijazi Arabic, as briefly dis-
cussed in Al-Essa (2019, 161), where the 2nd person masculine
singular possessive suffix is given as -ak, but the corresponding
148 Davis, Alshammari, Alahmari, and Alhuwaykim

2nd person feminine singular possessive suffix is reported as hav-

ing the allomorphs -ik after a word-final consonant and -ki after
a word-final vowel. With this as background, in this section we
consider specifically the allomorphy of the 2nd person masculine
singular possessive suffix in an understudied Southwestern Saudi
Arabian (SSA) variety spoken in the area of Northern Tihama,
which, as reported in Alahmari (2018), shows unusual three-way
allomorphy between -k, -ak, and -ka. We will detail their distri-
bution and provide argumentation for -k being the underlying
form of the suffix (we do not discuss the separate problem of the
2nd person feminine singular possessive suffix, which involves a
different set of issues and does not always pattern like its mascu-
line counterpart).
The distribution of the 2nd person masculine singular pos-
sessive suffix in SSA is straightforward, but interesting because it
attests the three allomorphs -k, -ak, and -ka, which makes it
somewhat different from other varieties. The allomorphs are pre-
dictable depending on the nature of the last syllable of the noun
to which it attaches. The analytical questions are which variant
of the allomorph constitutes the basic or underlying form; how
are the other variants derived; and what might be the diachrony
of the allomorphy? In our view, the basic or underlying form of
the suffix in SSA is just /-k/. To see this, we must consider the
range of data that is shown below. A main observation is that the
allomorphy is predictable based on the nature of the last syllable
of the base. To illustrate this, first consider the base words in (13)
that end in a final CVC syllable (the base word is in the left col-
umn, the 2nd person masculine singular suffixed form is in the
 Phonology and Morphology of Saudi Varieties of Arabic 149

middle column, and the gloss is on the right; the full stop indi-
cates a syllable boundary).
Table 4: Possessive formation for stems ending in CVC
(13) Base word Suffixed form Gloss
(a) /kutub/ ku.tubk ‘your books’
(b) /walad/ wa.ladk ‘your son’
(c) /balad/ ba.ladk ‘your country’
(d) /galam/ ga.lamk ‘your pen’
(e) /maktab/ mak.tabk ‘your desk’
(f) /ṣaaḥib/ ṣaa.ḥibk ‘your friend’

Here we see that the suffixal allomorph -k can attach to words

that end in a single consonant. The result is a final consonant
cluster ending in k. Since SSA allows for a word-final voiceless
consonant to be preceded by any other consonant, the final clu-
ster in word forms like that shown in (13) is phonotactically per-
missible. However, the generalisation for when the allomorph -k
occurs is not that the base word ends in a single consonant, but
that it ends in a CV(C) syllable. This is made clear by the forms
in (14) where a long vowel precedes the final consonant in the
base word (the forms in the rightmost column with an asterisk
are ungrammatical).
Table 5: Possessive formation for stems ending in CVVC
(14) Base word Suffixed form Gloss Ungrammatical
alternative
(a) /baab/ baa.bak ‘your door *baabk
(b) /beet/ bee.tak ‘your home *beetk
(c) /xaal/ xaa.lak ‘your uncle *xaalk
(d) /jiiraan/ jii.raa.nak ‘your neighbours *jii.raank

These forms make clear the role of syllable weight in determining

allomorphy. If we consider the example in (14a), /baab/, the un-
grammatical output *baabk is not disallowed because of the final
150 Davis, Alshammari, Alahmari, and Alhuwaykim

cluster, since the final cluster bk occurs in (13a), kutubk. Rather,

a word-final consonant cluster after a long vowel, as in the
righthand column in (14), is ruled out because it results in an
extra-heavy CVVCC syllable. Since this is not phonotactically per-
mitted, the allomorph that surfaces is -ak.
The role of final syllable weight in the determination of al-
lomorph is also made clear by word forms that end in a consonant
cluster or a word-final geminate. These words always take the
allomorph -ak, as shown in (15) and (16).
Table 6: Possessive formation for stems ending in CVCC

(15) Base word Suffixed form Gloss Ungrammatical

alternative
(a) /bint/ bin.tak ‘your daughter’ *bintk
(b) /uxt/ ux.tak ‘your sister’ *uxtk
(c) /ahl/ ah.lak ‘your family’ *ahlk
(d) /nafs/ naf.sak ‘yourself’ *nafsk

Table 7: Possessive formation for stems ending in CVGG

(16) Base word Suffixed Gloss Ungrammatical
form alternative
(a) /rabb/ rab.bak ‘your god’ *rabbk
(b) /umm/ um.mak ‘your mother’ *ummk
(c) /jadd/ jad.dak ‘your grandfather’ *jaddk
(d) /ʿamm/ ʿam.mak ‘your uncle’ *ʿammk
(e) /maḥall/ ma.ḥal.lak ‘your place’ *ma.ḥallk

In both (15) and (16), the suffixation of the allomorph -k would

lead to an extra-heavy syllable ending in three consonants, as
shown in the rightmost column; such forms are avoided and the
allomorph -ak surfaces. If we reference moraic theory for the ana-
lysis of Arabic along the lines of Watson (2002), Mahfoudhi
(2005), Davis and Ragheb (2014), and Alahmari (2018) specifi-
cally for SAA, a short vowel would comprise one mora, a
 Phonology and Morphology of Saudi Varieties of Arabic 151

geminate consonant one mora, a long vowel would be two moras,

and a coda consonant (other than one in word-final position)
would also constitute one mora. Given this moraic structure, the
generalisation for the occurrence of the allomorph -ak is that it
attaches to a final-syllable that is bimoraic. The allomorph -k can-
not attach to such a syllable since it would create an extra-heavy
(i.e., trimoraic) word-final syllable; such syllables are prohibited
in many Arabic dialects.
Given the patterning whereby the allomorph -ak attaches
to a word-final heavy (bimoraic) syllable, the other allomorph -k
can be analysed as suffixing to a word-final light syllable (i.e., a
monomoraic syllable), as shown in (13). Recall that a word-final
consonant is not moraic, so that a final CVC syllable as in the
word ku.tub ‘books’ would be monomoraic (or light). In the suf-
fixed output of word forms like ku.tub-k ‘your books’ in (13a), the
final syllable becomes bimoraic (heavy), with the vowel u of the
final syllable contributing one mora and the immediately follow-
ing consonant b contributing another; the suffix -k does not add
weight. Further evidence that the suffix -k attaches to word-final
light syllables come from base words that end in a final CV. Con-
sider the data in (17).
Table 8: Possessive formation for stems ending in CV
(17) Base word Suffixed form Gloss
(a) /abu/ a.buuk ‘your father’
(b) /axu/ a.xuuk ‘your brother’
(c) /ʿaša/ ʿa.šaak ‘your dinner’
(d) /ġada/ ġa.daak ‘your lunch’

Here, as in many other dialects of Arabic, the allomorph -k is

suffixed to a vowel-final word, with subsequent vowel lengthen-
152 Davis, Alshammari, Alahmari, and Alhuwaykim

ing of the final vowel. While some analyses have argued that the
final vowel is underlyingly long in such forms (e.g., McCarthy
2005), Alahmari (2018) specifically argues that the final vowels
in the SSA words in (17) are underlyingly short, which is in
agreement with Watson’s (2002) view on the final vowel length
problem in Arabic (the details of which are beyond the scope of
the current chapter).
A set of data related to the word forms in (17) are words
that are marked with the feminine ending tā marbūṭa, shown in
(18), where in isolation the word ends in the short suffixal vowel
-a, but in possessive forms the phoneme t occurs immediately af-
ter the short vowel, which can be viewed as a linker to the fol-
lowing morpheme. When the 2nd person masculine singular pos-
sessive morpheme is suffixed to such a word, the allomorph -k
occurs and clusters with the t, as in (18).
Table 9: Possessive formation with stems ending in tā marbūṭa
(18) Base word Gloss Suffixed form Gloss
(a) ra.ga.b-a ‘neck’ ra.ga.batk ‘your neck’
(b) jad.d-a ‘grandmother’ jad.datk ‘your grandmother’
(c) ʿam.m-a ‘aunt’ ʿam.matk ‘your aunt’

The possessive forms shown in (18) are completely regular in the

sense that the allomorph -k attaches to a final light syllable,
whether we take that syllable to be vowel-final ending in a short
-a (a final CV syllable) or abstractly ending in /C-at/ where the
final /t/ is not pronounced in unaffixed forms.
Given the presentation so far in this section, that the allo-
morphy between -k and -ak is determined by the weight of the
final syllable to which it is suffixed, whereby -k attaches to mon-
omoraic final syllables and -ak to bimoraic ones, the data shown
 Phonology and Morphology of Saudi Varieties of Arabic 153

below in (19) are unexpected, where a third allomorph appears,

-ka, rather than the expected -k.
Table 10: Possessive formation with the suffix -ka
(19) Base word Suffixed form Gloss Ungrammatical
alternatives
(a) /malik/ ma.lik.ka ‘your king’ *ma.likk
*ma.li.k-ak
(b) /samak/ sa.mak.ka ‘your fish’ *sa.makk
*sa.ma.k-ak
(c) /maalik/ maa.lik.ka ‘your owner’ *maa.likk
*maa.li.k-ak
(d) /maslak/ mas.lak.ka ‘your path’ *mas.lakk
*mas.la.k-ak

What the base word forms shown in (19) all have in common is
that the final syllable is CVk, that is, a word-final light (monomo-
raic) syllable ending in k. The expected allomorph based on the
discussion above for a word like samak ‘fish’ in (19b) should be
*samakk, but this does not occur. The clear reason for this is the
avoidance of a derived bimorphemic word-final geminate. While
such geminates can occur in SSA (e.g., sakat-t ‘I fell silent’), they
are avoided for the 2nd person masculine singular possessive suf-
fix (although forms like ma.lik-k ‘your king’ do occur in other
Saudi varieties, such as the northwestern variety discussed in
Alhuwaykin 2018). Given the avoidance of the derived word-fi-
nal bimorphemic geminate-k in (19), one might expect the cor-
rect output for the words in (19) to have the allomorph -ak, re-
sulting in a form like *samak-ak ‘your fish’ for (19b), but recall
from the above discussion that -ak attaches only to word forms
that have a final bimoraic syllable. The base for the hypothetical
*samak-ak would be sa.mak with a word-final light CVC syllable.
Consequently, a third allomorph surfaces only for base words
154 Davis, Alshammari, Alahmari, and Alhuwaykim

that have CVk as its final syllable. This allomorph is -ka, which
we would maintain, is an archaic form that survives in this spe-
cific context.
A final set of data that is supportive of the analysis offered
here comes from base words where the final syllable is bimoraic
with k as the word-final consonant. The pertinent data are shown
in (20).
Table 11: Possessive forms of words with final /k/ in a bimoraic syllable
(20) Base Suffixed Gloss Ungrammatical
word form alternatives
(a) /mulk-k/ mul.kak ‘your posses- *mulkk
sion’ *mulk.ka
(b) /šeek-k/ šee.kak ‘your check’ *šeekk
(borrowing) *šeek.ka
(c) /šakk-k/ šak.kak ‘your doubt’ *šakkk
*šakk.ka
Since the stems to which the 2nd person masculine singular pos-
sessive suffix attaches in (20) consist of a single heavy (bimoraic)
syllable, the allomorph that occurs is -ak, which attaches to word-
final bimoraic syllables. As shown in the rightmost column, the
suffixing of either of the other allomorphs would create extra-
heavy (trimoraic) syllables and would be phonotactically prob-
lematic. Thus, we maintain that the distribution of the allo-
morphs of the 2nd person masculine singular possessive suffix is
for the most part weight-based, with -ak attaching to any base
word whose final syllables is bimoraic and -k attaching to base
words whose final syllable is monomoraic, unless the final sylla-
ble is CVk, in which case the allomorph is -ka.
In considering the questions concerning which allomorph
is the underlying form of the 2nd person masculine singular pos-
sessive suffix (from a synchronic perspective) and what the
 Phonology and Morphology of Saudi Varieties of Arabic 155

diachrony might be, we would maintain, as argued for in Alah-

mari (2018), that in SAA the synchronic underlying form of the
suffix is /-k/, without a vowel. The clearest evidence for this
comes from the data in (13), where base words end in a final CVC
syllable, as in kutub ‘books’. Here, any of the three allomorphs
are phonotactically possible, which includes the correct output
kutubk as well as the unattested *kutubak and *kutubka. The fact
that the correct output is kutubk strongly suggests that the under-
lying form is just /-k/, since there would be no reason for the low
vowel to delete from a synchronic perspective if the underlying
forms were either /-ak/ or /-ka/. We refer the reader to Alahmari
(2018) for a formal analysis of the allomorphy within Optimality
Theory. If we maintain that the underlying form of the suffix is
/-k/, then the rule that inserts the vowel -a in such forms as bint-
ak from underlying /bint-k/ would probably be morphophono-
logical in order that the vowel quality be a. It may be possible to
maintain a purely phonological account of the vowel insertion
for the data shown in (14)–(16) and (20), but the specific argu-
mentation for that is beyond the scope of this chapter.
With respect to the diachrony, that is, which allomorph
may reflect a more archaic form of the language, a case can be
made that the -ka allomorph that appears only on word forms
like in (19), which end in a final CVk syllable, is actually an ar-
chaic form reflecting what is found in Classical Arabic. One could
speculate that the loss of suffixal -ka in Saudi dialects relates to
the loss of an earlier case system in the dialects. Based on this, a
form like malik-ka ‘your king’ in SSA could be viewed as a rem-
nant from a possible earlier form, such as *malikV-ka (where V is
156 Davis, Alshammari, Alahmari, and Alhuwaykim

a vocalic case marker), whereby this case ending has been dia-
chronically lost. The persistence of -ka in SSA in just this one
environment can be understood as due to specific avoidance of a
word-final derived geminate; the allomorph did not change to -
ak since in SSA -ak attaches only to bimoraic syllables. Thus, one
could argue that the presence of the third allomorph -ka is an
archaism among contemporary Saudi varieties. As far as we are
aware, SSA is unusual among contemporary spoken Saudi varie-
ties in maintaining three allomorphs, -k, -ak, and -ka, for the 2nd
person masculine singular possessive suffix rather than two, -k
and -ak (or -ik), as in many other Saudi varieties.

4.0. Final Degemination and Stress Shift in the

Attuwair Dialect of Arabic
The special properties of geminate consonants in Arabic have
been discussed for a variety of phenomena, including stress as-
signment by Watson (2002) for Sanʿani Arabic, first language ac-
quisition by Ragheb and Davis (2014) for Cairene Arabic, and
mora timing by Khattab and Al-Tamimi (2014) for Lebanese Ar-
abic. All these studies demonstrate the saliency of geminate con-
sonants for Arabic phonology. One way that this saliency is man-
ifested is by the attraction of stress to a word-final syllable ending
in a geminate consonant, which occurs in most dialects of Arabic,
and the attraction of exceptional stress to syllables closed by a
geminate in southwestern Arabian dialects of Arabic, which in-
clude Sanʿani Arabic (Watson 2002) and the southwestern vari-
ety of Northern Tihama detailed by Alahmari (2018; 2020). In
both these varieties, primary stress falls on the first syllable of
 Phonology and Morphology of Saudi Varieties of Arabic 157

the word dáw.wart ‘I/you (MS) looked for’, where the first syllable
is closed by a geminate consonant; this pattern is exceptional,
since normally CVCC final syllables attract stress regardless of
the nature of the preceding syllable, as in the example gam.bárt
‘I/you (MS) sat’, where the first syllable is closed but not by a
geminate. The stress-attracting nature of geminate consonants
might be an areal feature of southwestern Arabia, but we think
that it is an understudied part of the stress system of Arabic dia-
lects more generally.
Given the saliency of geminate consonants for Arabic pho-
netics and phonology, it may be somewhat surprising to find
cases of degemination. One common case of degemination in Ar-
abic, discussed in Farwaneh (2009), occurs when an underlying
geminate comes immediately before another consonant; this is
exemplified by Levantine /kull-hum/, which is realised as
kul.hum ‘all of them’. This can be contrasted with Cairene
kul.lu.hum, where epenthesis occurs rather than degemination. A
second type of degemination that occurs in some Arabic dialects
is word-final degemination. This is characteristic of the Sudanese
varieties discussed by Ali (2014; 2015). In the Hamar subdialect
of Sudanese Arabic, not only is there word-final degemination,
but more generally, the final consonant in a word-final cluster
deletes. What is interesting is that the final degemination and
final cluster reduction do not entail a shift of stress. The conse-
quence of this is that the Hamar dialect distinguishes kál.lam ‘he
told’ from kal.lám ‘I told’ (from underlying /kallam-t/) totally on
the basis of stress location.
158 Davis, Alshammari, Alahmari, and Alhuwaykim

With this as background, an unusual feature that is found

in various varieties of Najdi Arabic and in northwest Arabia,
though little discussed, is the degemination of word-final gemi-
nates with concomitant stress shift off the final syllable, as in the
example á.ham ‘more important’, where most other dialects have
a.hámm. However, this does not happen when the word-final
geminate is bimorphemic, as in si.kát-t ‘I kept silent’, where there
is no stress shift or final degemination. In this section, based on
Alhuwaykim (2018), we will detail this phenomenon in one sub-
variety, spoken in the village of Attuwair iṭṭ.weer, which is a Bed-
ouin-influenced dialect within the boundaries of Sakaka City
(Aljouf) in northwestern Saudi Arabia, north of where Northern
Najdi Arabic is spoken, according to Ingham’s (1994) classifica-
tion of Najdi subdialects. In this section, we present the pertinent
data from Attuwair Arabic, offer a descriptive analysis of the phe-
nomenon, and suggest that while it is an innovative feature of
Arabic, it is most likely an older innovation.
In Attuwair Arabic in northwest Saudi Arabia and also in
at least some varieties of Najdi Arabic there is a phenomenon of
the degemination of word-final geminate consonants. However,
before presenting the data that illustrate this, we first consider
what we view as a related phenomenon: the shortening of word-
final long vowels in CVV base words upon prefixation, as exem-
plified in (21).
 Phonology and Morphology of Saudi Varieties of Arabic 159

Table 12: Final shortening of long vowels in CVV words upon prefixa-
tion (with prefix in bold)
(21) Word Gloss Word Gloss
(a) maa ‘water’ (a′) íl.ma ‘the water’
(note: máa-kum)
(b) daa ‘sickness (b′) íd.da ‘the sickness’
(c) zii ‘costume’ (c′) íz.zi ‘the costume
(d) jaa ‘he came’ (d′) máa.ja ‘he did not
come’
(e) fii ‘there is’ (e′) máa.fi ‘there is not’
(f) lii for me’ (f′) máa.li ‘there isn’t for
me’

The data in the left-hand column in (21a–f) show lexical items

that have the shape CVV and carry stress when pronounced in
isolation or as unaffixed. The prefixal data in the right-hand co-
lumn are interesting in that they show that the definite article
prefix in (21a′–c′) and the negative prefix in (21d′–f′) are part of
the prosodic word to which they attach and attract stress. When
this occurs, as shown, the vowel of the base word is short. The
most straightforward analysis is the shortening of a word-final
vowel upon prefixation with concomitant stress shift to the first
syllable. While one can argue whether the data in (21) really re-
flect word-final vowel shortening in disyllabic words (21a′–f′) or
an alternative analysis of vowel lengthening when the word is
realised as monosyllabic in isolation (21a–f) (and when a suffix
is added, as in máa.kum ‘your (PL) water’), a parallel phenomenon
occurs with word-final geminates: they are realised as geminates
in word forms that are monosyllabic (and when a suffix is added),
but they shorten (i.e., degeminate) when word-final in disyllabic
forms. This is shown by the data in (22), when a prefix is added
to a monosyllabic word that ends in a geminate, and in (23),
which shows disyllabic comparative (i.e., elative) forms that
160 Davis, Alshammari, Alahmari, and Alhuwaykim

typically end in a geminate consonant in most other dialects of

Arabic with word-final stress.
Table 13: Degemination of final geminates in CVGG words upon prefi-
xation (with initial stress)

(22) Word Gloss Word Gloss

(a) ʿamm ‘uncle’ (a′) íl.ʿam ‘the uncle’
(note: ʿám.m-i)
(b) hamm ‘distress’ (b′) íl.ham ‘the dis-
tress’
(c) ḥubb ‘love’ (c′) íl.ḥub ‘the love’
(d) ḥagg ‘truth’ (d′) íl. ḥag ‘the truth’
(e) jaww ‘weather’ (e′) íl.jaw ‘the
weather’
(f) fayy ‘shadow’ (f′) íl.fay ‘the
shadow’
(g) damm ‘blood’ (g′) fíi.dam ‘there is
blood’
(h) /jaa-yamm/ (h′) jáa.yam ‘he came to’
(i) /maḥall/ (i′) má.ḥal ‘place, gro-
cery store’

Table 14: Degemination of final geminates in segmentally mono-

morphemic words (elative forms)
(23) Other dialects Attuwair Gloss
(a) a.šádd á.šad ‘stronger’
(b) a.xáff á.xaf ‘lighter’
(c) a.hámm á.ham ‘more important’
(d) a.ʿámm á.ʿam ‘more general’
(e) a.ḥágg á.ḥag ‘more entitled to’

In the left-hand column of (22), the data in (22a–g) show mono-

syllabic word forms that end in a geminate consonant where the
geminate consonant is pronounced. It is clear from the suffixal
forms of these words, such as ʿám.m-i) ‘my uncle’ shown in (22a),
that these base words end in final geminates. However, when a
prefix is placed before these base words, such as the definite
 Phonology and Morphology of Saudi Varieties of Arabic 161

article prefix as in (22a′–f′), the word-final geminate consonant

undergoes degemination and stress shifts to the first syllable. The
data shown in (22g–i) show that degemination (and stress shift)
also occurs with other prefixes besides the definite article. Just
as interesting is the observation on comparative (elative) form of
roots, where the last two consonants are identical, as in (23).
These comparatives typically have final geminates with stress on
the final syllable in most dialects of Arabic, as shown in the left-
hand column of (23). In the Attuwair dialect (and in various Na-
jdi varieties as well) there is degemination of the final consonant,
with stress appearing on the initial syllable.
At first glance, the phenomenon reflected by the data in
(22) and (23), which involves the metrical shortening of the final
syllable and stress shift to the initial syllable, may seem some-
what reminiscent of an observation from Ingham’s (2008, 330)
work on Najdi Arabic about how the definite article attracts
stress. He notes, “The stress rules of the dialect also lead to stress-
ing of the definite article ál- preceding nouns of the form CvC or
CvCvC, as in álbil ‘the camels’, álwalad ‘the boy’, álhawa ‘love’”
(See also Watson 2011, 902 on this). However, while the attrac-
tion of stress to the definite article might explain the data in
(22a′–f′), where final degemination would be seen as a conse-
quence of stress shift, it does not really explain the elative data
in (23) nor the degemination seen in (22i). Moreover, the definite
article in Attuwair Arabic does not always attract the stress, as
can be seen in an example like il.má.ḥal ‘the place, grocery store’,
where stress is on the penultimate syllable, as it would be in a
word like mid.rí.sah ‘school’. Furthermore, if a monosyllabic noun
162 Davis, Alshammari, Alahmari, and Alhuwaykim

ends in two different consonants, then no stress shift occurs when

the definite article is added, as exemplified by il.ʿílm ‘the
knowledge, news’, where stress remains on the final syllable de-
spite the presence of the definite article.
If the motivation for final degemination (22)–(23) and final
vowel shortening (21) is not the stressing of the definite article
(or other prefix), then what is the motivation? Here we suggest
that there is a constraint that is active in Attuwair Arabic against
having a moraic element at the end of a word that would result
in a bimoraic stressed syllable. The practical result of this avoid-
ance would either be the shortening of a final long vowel or
degemination of a final geminate. This would not apply to mon-
osyllabic word forms, like the data in the left-hand column of
(21) and (22), since it would result in a (subminimal) mono-
moraic word form, for example *ma ‘water’ instead of maa or
*ʿam ‘uncle’ instead of ʿamm (see Alhuwykim 2018 for an Opti-
mality Theory analysis that offers a formal analysis with ranked
constraints). The evidence that the constraint in Attuwair Arabic
against having a moraic element at the end of a word that would
result in a bimoraic stressed syllable comes from apparent excep-
tions in which there is no degemination. Consider the two exam-
ples shown below in (24).

(24) No degemination (and no stress shift) if the word-final geminate

is morphologically derived
(a) /sakat-t/ → si.kátt ‘I kept silent’
(b) /malak-k/ → mi.lákk ‘your (MS) king’

Both words in (24) end in a final geminate, but there is no

degemination. The difference between the two words in (24) and
 Phonology and Morphology of Saudi Varieties of Arabic 163

the final geminate in the words in (22) and (23) which undergo
degemination is that in (22) and (23) the final geminate is un-
derlyingly moraic, whereas in (24a–b) the final geminate is de-
rived (sometime referred to as a ‘fake’ geminate). In (24a), the
word-final geminate-t that surfaces comes about through the con-
catenation of two different morphemes: a root-final /t/ followed
by the inflectional 1st person suffix /-t/. Neither of these conso-
nants is underlyingly moraic, so no degemination (i.e., avoidance
of a word-final mora) occurs. Similarly, the word-final geminate-
k in (24b) comes about through the concatenation of the final
/k/ root morpheme for malak with the 2nd person masculine sin-
gular possessive suffix -k. Neither of these is underlyingly moraic,
so there is no word-final mora in (24b). This view that it is a
word-final mora that is lost in the degemination process (and in
the vowel shortening process) is further supported by the lack of
stress shift in words ending in a consonant cluster as shown in
(25).

(25) No stress shift (or mora loss) in words ending in two different
consonants:
(a) la.ʿáb-t ‘I played’
(b) il.ʿílm ‘the knowledge/news’

There is no stress shift or word-final mora loss in the two words

in (25) because the word-final consonant in these words is not
moraic. The second syllable retains its stress because it is a heavy
(bimoraic) syllable, but the word-final consonant is not moraic.
A final type of evidence for the view that stress shift results from
the demoraification of a word-final consonant (if it is moraic), as
in (22) and (23), or the demoraification of a word-final long
164 Davis, Alshammari, Alahmari, and Alhuwaykim

vowel, as in (21), comes from the observation that no word-final

degemination occurs if the final geminate is preceded by a long
vowel. This is shown by the two examples in (26).

(26) No degemination in word-final CVVGG syllables

(a) il.jáadd ‘the serious (M)’
(b) il.ʿáamm ‘the general/public’

The last syllable in each of the examples in (26) is trimoraic.

These final syllables have a long vowel which is two moras and
a final geminate which adds the third mora. If degemination were
to occur, the final syllable would still be heavy because of the
long vowel maintaining the stress. Words in Attuwair Arabic that
have a final CVVC syllable receive stress on that syllable, as illu-
strated by such words as bi.ʿíid ‘distant’ and il.ʿáam ‘the (last)
year’. Consequently, degemination in (26) would still leave stress
on the final syllable, meaning that word-final degemination (and
vowel shortening) only occur if it would result in a stress shift.
To summarise our descriptive analysis, an underlying
word-final long vowel shortens, as in (21), and an underlying
word-final geminate degeminates, as in (22)–(23), only if the out-
come is an unstressed monomoraic final syllable. That is, there is
avoidance of stressing a word-final bimoraic syllable that ends in
a moraic element (see Alhuwaykim 2018 for a formal Optimality
Theory analysis).
There are a number of larger questions that word-final
degemination in Attuwair Arabic relate to. First, as a theoretical
issue, final degemination as illustrated in this section is best un-
derstood on the view that geminate consonants are underlyingly
moraic. This explains why there is no degemination in (24a–b),
 Phonology and Morphology of Saudi Varieties of Arabic 165

when the surface geminate is derived from two different mor-

phemes: there is no underlying geminate in these word forms, so
degemination (i.e., demoraification) cannot occur. This supports
the view that geminates are underlyingly moraic in Arabic, as
maintained by Davis and Ragheb (2014) and Khattab and Al–
Tamimi (2014). Second, is the issue of the geographical isogloss
of final degemination in Saudi Arabian varieties of Arabic and
whether it always occurs alongside final vowel shortening, as in
(21). While our focus has been on Attuwair Arabic in northwest
of Saudi Arabia, somewhat north (and west) of northern Najdi
Arabic as delimited in Ingham (1994), we are aware that word-
final degemination is also characteristic of at least some varieties
of Najdi Arabic. However, we do not know what the isogloss of
final degemination might be and if the specific details as pre-
sented here for Attuwair Arabic are exactly the same or different
in other varieties that witness final degemination. A final matter
concerns whether word-final degemination is an archaic feature
of Attuwair (and Najdi) Arabic or an innovation reflecting an in-
ternal development. While we have no direct evidence bearing
on it, we suggest that final degemination in Attuwair (and Najdi)
Arabic is an innovation specific to an area of Saudi Arabia that
includes Sakaka City and at least parts of the Najdi region, but
degemination is probably not a recent innovation, given that it
seems to be fairly widespread. These matters are in need of fur-
ther research.
166 Davis, Alshammari, Alahmari, and Alhuwaykim

5.0. Conclusion
In this chapter we have presented an abundance of data reflecting
unusual phenomena in understudied dialects of the Arabian Pen-
insula, specifically morphological augmentatives in Northern
Najdi/Haʾili Arabic, the distribution of the allomorphs of the sec-
ond person masculine singular possessive suffix in a southwest-
ern Saudi Arabian dialect (Northern Tihama), and final
degemination and stress retraction in Attuwair, Sakaka City Ara-
bic. For each case we have presented detailed data illustrating
the phenomenon and have offered a descriptive analysis. We
have also discussed the question as to what extent the phenom-
ena presented here reflect archaic features, be it substrate or
older forms of Arabic, or just reflect Arabic-internal develop-
ments. In all the cases presented, data come from the intuitions
of native speaker linguists in consultation with other speakers of
the same dialect. While some may consider this manner of data
compilation less than rigorous, it is nonetheless important for na-
tive speakers to carry out such work because data can be gath-
ered quickly and can lead to research questions regarding the
verification of the data, their status as internal developments or
archaic forms, and their contemporary sociolinguistic manifesta-
tions.

References
Abboud, Peter. 1979. ‘The Verb in Northern Najdi Arabic’. Bulle-
tin of the School of Oriental and African Studies 42/3: 467–
99.
 Phonology and Morphology of Saudi Varieties of Arabic 167

Alahmari, Musa. 2018. ‘An Optimality-Theoretic Analysis of

Some Aspects of the Phonology and Morphology of a South-
western Saudi Arabian Arabic Dialect’. PhD dissertation, In-
diana University.
———. 2020. ‘Stress Assignment and Foot Construction in a
Southwestern Saudi Arabian Dialect’. In Perspectives on Ar-
abic Linguistics XXXII, edited by Elly van Gelderen, 11–33.
Amsterdam: John Benjamins.
Al-Essa, Aziza. 2019. ‘Phonological and Morphological Varia-
tion’. In The Routledge Handbook of Arabic Sociolinguistics,
edited by Enam Al-Wer and Uri Horesh, 151–68. London:
Routledge.
Alfaifi, Abdullah, and Peter Behnstedt. 2010. ‘First Notes on the
Dialect of Ǧabal Fayfā (Jazan Province/Saudi Arabia)’.
Zeitschrift für Arabische Linguistik 52: 53–67.
Alfaifi, Abdullah, and Stuart Davis. 2021. ‘An Examination of the
/m-/ Definite Article in Upper Faifi Arabic’. Zeitschrift für
Arabische Linguistik 73: 36–52.
Alhuwaykim, Mamdouh. 2018. ‘Aspects of the Phonology of
Northwestern Saudi Arabic Dialect: An Optimality-Theo-
retic Analysis’. PhD dissertation, Indiana University.
Ali, Abdel-Khalig. 2014. ‘Syllabification and Phrasing in Three
Dialects of Sudanese Arabic’. PhD dissertation, University
of Toronto.
———. 2015. ‘Prosodic Domains of Syllabification Sudanese Ar-
abic’. In Perspectives on Arabic Linguistics XXIX, edited by
Hamid Ouali, 33–54. Amsterdam: John Benjamins.
168 Davis, Alshammari, Alahmari, and Alhuwaykim

Al-Jallad, Ahmad. 2013. ‘Arabia and Areal Hybridity’. Journal of

Language Contact 6: 220–42.
———. 2021. ‘The History of the am- Definite Article: South Ara-
bian or Arabic?’. Zeitschrift für Arabische Linguistik 73: 53–
70.
Al Rasheed, Madawi. 1991. Politics in an Arabian Oasis: The Ra-
shidis of Saudi Arabia. New York: I. B. Tauris Publishers.
Alshammari, Wafi, and Stuart Davis. 2019. ‘Diminutive and Aug-
mentative Formation in Northern Najdi/Ha’ili Arabic’. In
Perspectives on Arabic Linguistics XXXI, edited by Amel
Khalfaoui and Youssef Haddad, 51–73. Amsterdam: John
Benjamins.
Al-Wer, Enam, and Khairia Al-Qahtani. 2016. ‘Lateral Fricative
ḍād in Tihāmat Qaḥtān: A Quantitative Sociolinguistic In-
vestigation’. In Perspectives on Arabic Linguistics XXVII, ed-
ited by Stuart Davis and Usama Soltan, 151–69. Amster-
dam: John Benjamins.
Assuwaida, Abdul Rahman. 1997. An-Nakhatu at-Taʾiyyatu fi al-
Lahjati al-Haʾiliyyah [The Tayy Flavour in the Haʾili Dia-
lect]. Haʾil, Saudi Arabia: Dar al-Andalus li-l-Nashr wa-l-
Tawzī. [Arabic]
Behnstedt, Peter. 2016. Dialect Atlas of North Yemen and Adjacent
Areas. Leiden: Brill.
Cowan, David. 1958. Modern Literary Arabic. Cambridge: Cam-
bridge University Press.
Davis, Stuart, and Marwa Ragheb. 2014. ‘Geminate Representa-
tion in Arabic’. In Perspectives on Arabic Linguistics XXIV–
 Phonology and Morphology of Saudi Varieties of Arabic 169

XXV, edited by Samira Farwaneh and Hamid Ouali, 3–19.

Amsterdam: John Benjamins.
Farwaneh, Samira. 2009. ‘Toward a Typology of Arabic Dialects:
The Role of Final Consonantality’. Journal of Arabic and Is-
lamic Studies 9/4: 82–109.
Fattah, Hala, and Frank Caso. 2009. A Brief History of Iraq. New
York: Hermitage.
Gadoua, Abdulhamid, and Stuart Davis. 2019. ‘Diminutive For-
mation in a Libyan Dialect with Some Phonological Impli-
cations’. In Perspectives on Arabic Linguistics XXXI, edited by
Amel Khalfaoui and Youssef Haddad, 31–49. Amsterdam:
John Benjamins.
Goitein, Shlomo D. 1960. ‘The Language of Al-Gades: The Main
Characteristics of an Arabic Dialect Spoken in Lower
Yemen’. Le Muséon 73: 351–94.
Holes, Clive. 1984. Colloquial Arabic of the Gulf and Saudi Arabia.
London: Routledge and Keegan Paul.
Ingham, Bruce. 1982. North East Arabian Dialects. London: Kegan
Paul International.
———. 1994. Najdi Arabic: Central Arabian. Amsterdam: John
Benjamins
———. 2008 ‘Najdi Arabic’. In Encyclopedia of Arabic Language
and Linguistics III, edited by Kees Versteegh, 326–34. Lei-
den: Brill.
Khattab, Ghada, and Jalal Al-Tamimi. 2014. ‘Geminate Timing in
Lebanese Arabic: The Relationship between Phonetic Tim-
ing and Phonological Structure’. Laboratory Phonology 5:
231–69.
170 Davis, Alshammari, Alahmari, and Alhuwaykim

Mahfoudhi, Abdessatar. 2005. ‘Moraic Syllable Structure and

Edge Effects in Arabic’. In Perspectives on Arabic Linguistics
XVII–XVIII, edited by Mohammad Alhawary and Elabbas
Benmamoun, 27–48. Amsterdam: John Benjamins.
McCarthy, John. 2005. ‘The Length of Stem-final Vowels in Col-
loquial Arabic’. In Perspectives on Arabic Linguistics XVII–
XVIII, edited by Mohammad Alhawary and Elabbas Benma-
moun, 1–26. Amsterdam: John Benjamins.
Prochazka, Theodore, Jr. 1988. ‘Gleanings from Southwestern
Saudi Arabia’. Zeitschrift für Arabische Linguistik 19: 44–49.
Ragheb, Marwa, and Stuart Davis. 2014. ‘On the L1 Development
of Final Consonant Clusters in Cairene Arabic’. In Perspec-
tives on Arabic Linguistics XXVI, edited by Reem Khamis-
Dakwar and Karen Froud, 263–81. Amsterdam: John Ben-
jamins.
Sifianou, Maria. 1992. ‘The Use of Diminutives in Expressing Po-
liteness: Modern Greek Versus English’. Journal of Pragmat-
ics 17/2: 155–73.
Watson, Janet C. E. 2002. The Phonology and Morphology of Ara-
bic. Oxford: Oxford University Press.
———. 2011. ‘Dialects of the Arabian Peninsula’. In The Semitic
Languages, edited by Stefan Weninger, Geoffrey Kahn, Mi-
chael Streck, and Janet Watson, 897–908. Berlin: de Gruy-
ter.
———. 2018. ‘South Arabian and Arabic Dialects’. In Arabic His-
torical Dialectology: Linguistic and Sociolinguistic Approaches,
edited by Clive Holes, 316–34. Oxford: Oxford University
Press.
 Phonology and Morphology of Saudi Varieties of Arabic 171

Watson, Janet C. E,. and Munira Al-Azraqi. 2011. ‘Lateral Frica-

tives and Lateral Emphatics in Southern Saudi Arabia and
Mehri. Proceedings of the Seminar for Arabian Studies 41:
425–31.
Zewi, Tamar. 2006. ‘Diminutive’. In Encyclopedia of Arabic Lan-
guage and Linguistics I, edited by Kees Versteegh, 637–40.
Leiden: Brill.