See discussions, stats, and author profiles for this publication at: [Link]

net/publication/325895348

The Amazigh influence on Moroccan Arabic: Phonological and morphological

borrowing

Preprint · June 2018

DOI: 10.13140/RG.2.2.12030.69447

CITATIONS READS

0 841

1 author:

Mohamed Lahrouchi
French National Centre for Scientific Research
83 PUBLICATIONS 286 CITATIONS

SEE PROFILE

All content following this page was uploaded by Mohamed Lahrouchi on 21 June 2018.

The user has requested enhancement of the downloaded file.

 The Amazigh influence on Moroccan Arabic:
Phonological and morphological borrowing

Mohamed Lahrouchi
CNRS & Université Paris 8

1. Introduction

Centuries of coexistence with Amazigh have deeply impacted the Moroccan varieties of
Arabic in all aspects of the lexicon and grammar. One of the major features these varieties
have developed in contact with Amazigh is the loss of vowel length contrast: From Classical
Arabic (henceforth, CA), they have retained three short vowels /i, a, u/, and they have
developed a short central vowel [ə], which Caubet (2007: 03) historically views as a merge of
/a/ and /i/, in contrast with /ŭ/ found in forms like skŭt ‘shut up!’ and xŭbz ‘bread’1 (see also
Watson 2002: 21). In mainstream Moroccan Arabic (henceforth, MA), /u/ is realized as a
secondary feature on labial, velar and uvular consonants, similarly to Amazigh akwr ‘steal’,
gwmr ‘hunt, fish’ and alqwnajn ‘rabbit’ (Heath 1997: 209). Another feature that MA has
developed in contact with Amazigh is sibilant harmony, a long distance process whereby /s/
and /z/ agree in voice and anteriority with post-alveolar /ʃ/ and /ʒ/: e.g. /zuʒ/ > [ʒuʒ] ‘two’,
/ʃms/ > [ʃəmʃ] ‘sun’, /zlliʒ/ > [ʒəlliʒ] ‘tiles’ (see Caubet 2007, and Zellou 2010). Morphology
is no exception to this trend: MA has borrowed the Amazigh feminine marker as a derivative
prefix, used to form abstract nouns often referring to professions (cf. Guay 1918, Colon 1947,
Chtatou 1997, Tilmatine 1999, Caubet 2007, and Zellou 2011). The examples in (1) illustrate
the phenomenon.
(1) Base noun Profession noun
bəqqal ‘grocer’ tabəqqalt
sərraʒ ‘saddler’ tasərraʒt
bərraħ ‘crier’ tabərraħt
bənnaj ‘mason’ tabənnajt
χərraz ‘shoemaker’ taχərrazt

Atypical morphological borrowing of this type, which converts a flectional affix into a
derivational one will be studied along with the aforementioned phonological features. In

I am very grateful to Karim Bensoukas, the editors and the anonymous reviewers for their valuable feedback.
All remaining errors or omissions are of course my own.
1
According to Caubet (2007: 03), Moroccan Arabic has five vowels, "three long or medium ones: /ā/, /ī/, /ū/,
and two short or ultra-short ones: /ə/, /ŭ/." The phonemic status of the two short vowels is subject to debate, as
we will show in this paper.

1
section 2, it is argued that the loss of vowel length contrast results from the loss of CA short
vowels in line with the proposal put forth in Lowenstamm (1991). The schwa-like vowel will
be analysed as an “epenthetic” vowel, whose distribution is entirely predictable from a
syllabic point of view. Sibilant harmony is discussed in section 3. A comparison with
Amazigh will prove necessary for understanding the phenomenon. Section 4 turns to
morphological borrowing: The nature and the behaviour of the circumfix /ta-…-t/ will be
examined therein. Section 5 concludes the paper.

2. Segmental aspects

As members of the Afroasiatic language family, MA and Amazigh share several phonological
features, part of which is naturally reflected in their phonemic inventory. Of particular interest
is the vowel system of MA, which has lost vowel length contrast in contact with Amazigh.
There is a broad consensus among scholars that the three full vowels of MA are historical
reflexes of long vowels and diphthongs in Classical Arabic, and that the schwa has replaced
the Classical Arabic short vowels (see, among others, Heath 1997, Elmedlaoui 2000, Watson
2002, Caubet 2007, and Lahrouchi 2018a). The examples in (2) clearly illustrate this idea.

(2) Classical Arabic Moroccan Arabic

i. waːfaqa wafəq ‘he agreed’
kitaːb ktab ‘book’
lisaːn lsan ‘tongue’
ħabiːb ħbib ‘beloved’
ii. fahima fhəm ‘he understood’
labisa lbəs ‘he wore’
samiʕa sməʕ ‘he heard’
kataba ktəb ‘he wrote’
iii. qaws qus ‘arch’
lawn lun ‘colour’
bajt bit ‘room’
sajf sif ‘sword’

It is worth noting that the distribution of schwa is limited, as opposed to the remaining
vowels; it appears only in closed syllables. Furthermore, it alternates with zero in many
contexts, including pre- and post-vocalic CC clusters (see Benhallam 1989/1990, Kabbaj

2
1990, Kaye 1990, Boudlal 2006, Bensoukas & Boudlal 2012, Lahrouchi 2018a, among
others).
The situation just depicted in MA also arises in Amazigh, as the examples in (3) show
in the Tamazight variety. In the items in (3a), the long vowels of Classical Arabic surface as
short in Tamazight, while the short vowels are deleted, and replaced when necessary by
schwa. The latter vowel is also found in Tamazight native words (3b).
(3)
a. Classical Arabic Tamazight (Taïfi 1991)
ʔalbaħr ləbħərˤ ‘the sea’
balaɣa bləɣ ‘to grow up’
ħadˤana ħdˤən ‘to supervise, place under wardship’
saːfara safər ‘to travel’
ʔalbaːb lbab ‘the door’
ʔalfiːl lfil ‘the elephant’
b. Tamazight native words
lməd ‘learn’
kməz ‘scratch’
rməd ‘gnaw’
ttər ‘ask, request’

Although there is little doubt that the loss of vowel length in MA and the use of schwa
are substrata of the Amazigh language, studies diverge as to the way this can be analysed,
especially with regard to the status of schwa. Several scholars, such as Heath (1997), Caubet
(2007) and Watson (2002), claim that schwa is the reflex of Classical Arabic short vowels,
suggesting that it has a phonemic status (Caubet 2007: 03, Aguadé 2010)2, while others argue
that it is a mere epenthetic vowel used to break up illicit consonant clusters (cf. Benhallam
1980, 1989/1990, Boudlal 2006, 2009, Bensoukas & Boudlal 2012, Lahrouchi 2018a, among
others). One piece of evidence in favour of the latter hypothesis lies in the distribution of
schwa and its alternation with zero. If schwa were the result of the merger of Classical Arabic
short vowels, it should be able to occur in open and closed syllables as well, just like short
vowels do in Classical Arabic. Actually, MA bans any occurrence of schwa in open syllables:
a form like *kətəbə (from Classical Arabic kataba ‘he wrote’) is excluded. Furthermore, the

2
Based on minimal pairs like ħǝbb ‘he kissed’ ≠ ħŭbb ‘love’, ħǝkk ‘he rubbed’ ≠ ħŭkk ‘small box’, nǝqṛa ‘I will
read’ ≠ nŭqṛa ‘silver’, ħǝrrˤ ‘more spicy than’ ≠ ħŭrrˤ ‘free’, nǝssˤ ‘text’ ≠ nŭssˤ ‘half’ and xǝdˤrˤa ‘green (fem.)’
≠ xŭdˤrˤa ‘vegetable’, Aguadé (2010: 101) argues that the vowels /ŭ/ are /ə/ are phonemic in MA.

3
position of schwa within the word varies depending on the nature of the affixes added to the
base: For instance, in the 3rd masculine singular ktəb ‘he wrote’, schwa appears between the
last two consonants, whereas in the corresponding feminine kətbat ‘she wrote’ it shifts
leftward between the first two consonants. This is a notable difference when comparing the
schwa of MA to that of French, which can be omitted but never moved: quelques semaines
‘some weeks’ can be realized as [kɛlkəsəmɛn] or [kɛlkəsmɛn], but not as *[kɛləksəmɛn].3
The behaviour of schwa in MA, just sketched out, makes its distribution largely
predictable. Standard syllable-based analyses view it as an epenthetic vowel used to avoid
complex codas (cf. Benhallam 1980, Boudlal 2001, and Echchadli 1986, among others).4 An
alternative is provided in Kaye (1990), which argues that schwa-zero alternations in MA are
better analysed in terms of empty vocalic positions, of which the properly ungoverned ones
surface as schwa (see also Lahrouchi 2018a). The pair qləb ‘he upturned’ / qəlbu ‘they
upturned’, represented below in (4), illustrates this kind of analysis (PG = Proper
Government):
(4)
a.
PG

q l ə b
| | | |
C V1 C V2 C V3

b.
PG

q ə l b u
| | | | |
C V1 C V2 C V3

These representations, which use the strict CV approach to syllable structure

(Lowenstamm 1996; Scheer 2004), clearly show that empty vocalic positions cannot be

3
One could argue that in Colloquial French this form is viable without any schwa. Actually, this is possible to
the extent that the deletion of schwa does not yield an illicit consonant cluster. To my knowledge, the four-
consonant cluster that results from the omission of schwas is simplified by deleting the lateral consonant, leading
to the form [kɛksmɛn] instead of *[kɛlksmɛn].
4
Except for some category-specific words which allow complex codas in the final position: e.g. kəlb ‘dog’, qəlb
‘heart’, ħənʃ ‘snake’. The reader is referred to Echchadli (1986) for a thorough discussion of this issue (see also
Boudlal 2001, 2006/2007).

4
adjacent. Any sequence of two empty positions must realize one, depending on the
government relation contracted with the neighbouring vowels (see Kaye et al. 1990 on
government phonology). In (4a), V3 is licensed to remain empty by virtue of being word-final.
Properly ungoverned V2 surfaces as schwa and then governs V1. In (4b), the third plural
marker –u associated to V3 properly governs V2, allowing it to remain phonetically inaudible.
Then V1, no more governed, surfaces as schwa. The same mechanism underlies the
distribution of schwa in all Amazigh varieties 5 , apart from Tashlhiyt where syllabic
consonants arise (see Dell & Elmedlaoui 1985, 2002, Boukous 1987, Jebbour 1996, 1999,
Bensoukas 2001, Ridouane 2008, Lahrouchi 2018a, among others). The forms represented in
(5) show how the schwa-zero alternations are handled in the Tamazight variety.
(5) lməd ‘learn (aorist)’ / ləmdəχ ‘I learnt’
a.
PG

l m ə d
| | | |
C V1 C V2 C V3

b.
PG

l ə m d ə χ
| | | | | |
C V1 C V2 C V3 C V4

The distribution of schwa in Tamazight proceeds in much the same way as it does in
MA. The word-final vocalic position is licensed to remain empty, therefore allowing the
preceding one to be phonetically realized. In (5a), properly ungoverned V2 surfaces as schwa
and then governs V1. The same position properly governed by V3 remains silent in (5b), while
V1 realizes as schwa. The difference between the bare form in (5a) and the inflected one in
(5b) ultimately lies in the distribution of empty nuclei, among which the ungoverned ones
surface at the phonetic level.
According to Lowenstamm (1991: 959), the situation depicted in MA reflects the loss
of the ability to associate peripheral vowels to non-branching positions. Consequently, only

5
According to Kossmann (1997: 50), certain varieties of Amazigh have phonemic ("structural") schwa as well as
an epenthetic one, relying on a few minimal pairs such as nətc ‘me’ / ntəc ‘we eat (aorist)’ and certain affixes
like the nominal plural suffix -ən whose vowel has a fixed position.

5
Classical Arabic long vowels, attached to two positions, remained in MA (although they
surface as short). The consonant clusters resulting form the loss of the short peripheral vowels
are then simplified by means of schwa epenthesis. The same reasoning has been extended to
the Amazigh vocalic system (see among others Bendjaballah 1999, 2005, Lahrouchi 2001,
2003, Lahrouchi & Ségéral 2010, and Ben Si Saïd 2014).
The hypothesis that branching preserves the phonetic realization of segments is further
evidenced by the vestige of Classical Arabic short /u/ in MA. Interestingly, this vowel is kept
either as short [ŭ] or as a labial feature in the vicinity of a labial, velar or uvular consonant.
This feature, which has been largely studied (see Heath 1997, Boudlal 2001, 2009, Caubet
2007, Zeroual et al. 2001, and Bensoukas & Boudlal 2012), is viewed as a substratum of the
Amazigh language, since it is not found in the Middle-Eastern varieties of Arabic. A rapid
comparison of the MA examples in (6a) and the Amazigh (Tashlhiyt variety) ones in (6b)
shows strong similarities in this respect.
(6) Singular Plural
a. muʒa mmwaʒ ‘wave’
w
baliza bb aləz ‘suitcase’
w
futˤɑ ff atˤe ‘towel’
mula mmʷalin ‘owners’
qoffa qʷfaf ‘straw basket’
b. aɣjjul iɣwjjal ‘donkey’
agru igwra ‘frog’
w
agajju ig jja ‘head’
aqammu iqʷmma ‘face (slang)’
akbur ikʷbar ‘squirrel’

The vowel /u/ of the singular forms is kept in the plural as a labial feature on the labial
velar, or uvular consonants. The retention of this feature on this specific subset of consonants,
to the exclusion of any other type of consonants, suggests that they share labiality. Several
studies, mainly couched within Element Theory, argue that labial and velar consonants
contain a labial element |U| (see Harris 1990, 1994, Backley 2011, 2012, and Scheer 1996)6. It
is precisely the sharing of this element that allows the short /u/ in the singular forms to remain

6
Note, however, that Scheer (1996) distinguishes labiality |B| from velarity |U|. The first element is found in
labial consonants, and the second in velar consonants.

6
as a secondary articulation on the preceding consonant in the plural.7 The same reasoning
underlies the presence of the secondary labial feature in many other forms of MA, some of
which clearly show that they lost the short /u/ of Classical Arabic (7a). Similar forms which
inherit the labial feature from an underlying /u/ are found in Amazigh (7b).

(7)
a. Moroccan Arabic Classical Arabic
xwəbz xubz ‘bread’
skwət ʔuskut ‘shut up!’
skwən jaskunu ‘he lives’
w
sk at sukuːt ‘silence’
w
g ʕəd ʔuqʕud ‘sit down!’

b. Amazigh (Tashlhiyt variety)

Aorist Preterit
knu ikwna
gru igwra
qlu iqwla
xlu ixwla
agwm jugm
w
ak r jukr

It is worth mentioning that in Amazigh, labiovelarization is blocked for labial

consonants: For instance, the preterit forms of bnu ‘build’ and ftu ‘go away’ are ibna and ifta,
without labiovelarized /b/ and /f/. This phonotactic restriction on the distribution of the labial
feature is due, according to Lahrouchi & Ulfsbjorninn (2018), to its being redundant on an
expression that already contains headed labial element |U| (labials are headed by |U|, but not
velars and uvulars).8 The same phenomenon is responsible, we argue, for labial dissimilation
in reciprocal forms, a phenomenon which turns the prefix /m-/ into [n]: e.g. /m-gibil/ >
[ngibil] ‘face each other’, /m-simiħ/ > [nsimiħ] ‘apologize to each other’ (see Elmedlaoui
1992, Lasri 1991, Alderete 1997, Lahrouchi 2003, 2018b, and Bensoukas 2014). Unlike MA,
Amazigh does not allow more than one headed |U| consonant per word.

7
The reader is referred to Honeybone (2005) on the idea of “sharing makes us stronger”. See also Bucci (2013)
on Coratino, an Italian dialect in which unstressed labial and palatal vowels resist reduction when sharing their
feature with a neighbouring consonant.
8
The idea that labial consonants are headed by the labial element |U| is also made explicit in Backley (2012 :
81), Kaye (2000), and Scheer (1996).

7
 The next section deals with sibilant harmony, another property that MA shares with
Amazigh. We shall see that the sibilants contained within a specific domain undergo
anteriority harmony (in both languages) and voicing harmony (only in Amazigh).

3. Phonological activity: sibilant harmony

In Amazigh, the causative prefix agrees in voicing and anteriority with the sibilant consonant
in the verb stem (see Bensoukas 2004, Elmedlaoui 1992, Dell & Elmedlaoui 2002, Kossmann
1997, and Lahrouchi 2001, 2003, 2018b). This is shown in the examples in (8).
(8) Tashlhiyt variety
Verb Causative
a. mun smun ‘pick up’
gudi sgudi ‘put in a pile’
nkr ssnkr ‘wake’
nu ssnu ‘cook’
b. kuʃm ʃkuʃm ‘paralyze’
ħuʒʒu ʃħuʒʒu ‘visit Mecca’
nʒm ʒʒnʒm ‘save’
nz zznz ‘sell’
In (8a), the causative prefix surfaces as a voiceless alveolar sibilant, whose quantity
variation is regulated by the prosodic shape of the verb stem (see Jebbour 1999, Dell &
Elmedlaoui 2002, and Lahrouchi 2018b). This contrasts with the examples in (8b), where the
causative prefix is realized as a voiced or voiceless palato-alveolar sibilant, depending on the
nature of the sibilant in the stem. Furthermore, we note that voicing harmony is blocked
whenever a voiceless consonant intervenes between the voiced sibilant and the causative
prefix such as in ʃħuʒʒu.
MA has developed the same phenomenon in contact with Amazigh, as shown in the
examples in (9) below. The reader is referred to Harris (1942), Harrell (1962), Heath (1987),
and Zellou (2010), for further details and analysis.
(9) Moroccan Arabic
zuʒ ~ ʒuʒ ‘pair, two’
zzaʒ ~ ʒʒaʒ ‘the glass’
sərˤʒəm ~ ʃərˤʒəm ‘window’
sfənʒ ~ ʃfənʒ ‘doughnut’
sfərˤʒəl ~ ʃfərˤʒəl ‘quince’
ssəmʃ ~ ʃʃəmʃ ‘the sun’

8
 This long distance harmony, applying and changing the word-initial sibilant into [ʃ] or
[ʒ], is optional. Accordingly, the alternations in (9), indicated by a tilde, are in free variation,
and certain speakers may use one or the other type of forms, depending on their language
register.

It is worth adding that sibilant harmony is limited to the stem, except for the definite
article, which undergoes the process by assimilating to the first consonant of the stem (e.g. /l-
sfərˤʒəl/ > [ʃʃfərˤʒəl] ‘quince’).9 In contrast, in Amazigh the process targets the causative
prefix, when immediately preceding the stem (8b). In case another prefix occurs in the
immediate vicinity of the stem, the causative prefix remains unchanged. This is exemplified
in (10) by the verb ʃ-ħʃʃm ‘be ashamed’ whose causative prefix immediately preceding the
stem undergoes harmony. When the reciprocal prefix is added, the causative /s-/ remains
unchanged.

(10) Verb Causative Reciprocal Causative + Reciprocal

ħʃʃm ʃ-ħʃʃm n-ħiʃʃim s-n-ħiʃʃim ‘shame each other’

The careful reader will have noticed that the reciprocal prefix /m/ dissimilate into [n],
in reaction to the stem labial consonant; a phenomenon we have been discussing in the
previous section. Based on this type of alternations, Lahrouchi (2001, 2003, 2018b: 15) has
proposed that sibilant harmony and labial dissimilation occur within a specific domain, which
consists of the stem template, preceded by an initial empty site of the form CV. Complex
combinations of the causative and reciprocal prefixes allow only the inner prefix, which has
access to the initial site, to interact with the segmental content of the stem. The representation
in (11) helps in understanding the proposal (the reader is referred to the original work for
details and analysis).

(11) a. /s-ħʃʃm/ >> [ʃħʃʃm] b. /s-m-ħiʃʃim/ >> [snħiʃʃim]

ʃ ħ ʃ m s n ħ i ʃ i m
| | / \ | | | | | / \ | |
vb[CVCVCVCVCV] CVvb[CVCVCVCVCV]

The causative prefix undergoes sibilant harmony in (11a) since it has access to the
initial CV (underscored), which falls within the domain of the verb. In (11b), it is the

9
The anonymous reviewerz rightly noticed that the sibilants do not undergo voicing agreement in MA. All the
cases of sibilant harmony we could think of involve anteriority only, turning underlying /s/ and /z/ into [ʃ] and
[ʒ], respectively.

9
reciprocal prefix that attaches to the initial CV, resulting in its dissimilation into [n] and
preventing at the same time the causative prefix from harmonizing with the stem sibilant.

Before concluding, we turn to morphological borrowing. The next section outlines a

specific type of nouns, which uses the Amazigh gender marker to derive so-called profession
names and abstract “quality” nouns.

4. Morphological borrowing

Amazigh has had such an impact on the phonological system of MA that it is very difficult for
natives of other Arabic varieties (especially Eastern ones) to understand it, mainly due to the
loss of vowel length contrast. This feature has probably also contributed to restructuring the
morphological shape of words, particularly with regard to case marking. The CA markers for
the nominative, accusative and genitive cases, which happen to be short vocalic suffixes, have
all disappeared in MA, just like any other short vowels: For example, the CA forms arraʒul-u
(nominative), arraʒul-a (accusative) and arraʒul-i (genitive) have all merged into one form in
MA, namely rrˤaʒəl ‘the man’.

Like MA, Amazigh has no case marking. Nouns in this language are inflected for free
and construct states, but not for case (see El Moujahid 1997, Ennaji 2001, Guerssel 1987,
1992, 1995, Ouhalla 1988, 1996, Lahrouchi 2013, among others). The initial vowel in
masculine forms alternates between the free state a- and the construct state u- (e.g. a-frux / u-
frux ‘boy’), while in the feminine free state ta- alternates with construct state t- (e.g. ta-fruxt /
t-frux-t ‘girl’).

This brings us to another type of borrowing in MA, namely the use of the circumfix
/ta-…-t/ as a derivational affix. In Amazigh, this affix consists of two ingredients: the
feminine marker /t/, circumfixed to the stem, and the nominal marker a-, which some scholars
analyse as a kind of “portmanteau” morpheme, realizing definiteness and/or case (Guerssel
1987, 1992, Ouhalla 1988, and El Hankari 2014). We put aside the issue of definiteness and
case marking in Amazigh, for it goes beyond the scope of this paper. The readers’ attention is
drawn to the morphological pieces of the aforementioned circumfix, and their role in
Amazigh as opposed to MA. In addition to being a gender marker, /t-…-t/ is used in
diminutives as well as in the formation of abstract and profession nouns, which Chtatou
(1997: 113) terms resultative nouns. Examples are given in (12) below.

10
(12) Tashlhiyt variety

a. Base noun Diminutive

aɣʷrab taɣʷrabt ‘wall’

akur takurtt ‘ball’
agajju tagajjut ‘head’
amzzˤuɣ tamzzˤuɣt ‘ear’
afarnu tafarnut ‘type of oven’

b. Agentive noun Abstract noun/profession

anddˤam tanddˤamt ‘poet, poetry’
agzzar tagzzart ‘butcher, butchery’
abnnaj tabnnajt ‘builder, masonry’
afllaħ tafllaħt ‘farmer, agriculture’
abrraħ tabrraħt ‘crier, auction’

The base nouns in (12a) are all masculine. Adding the feminine marker allows
deriving the corresponding diminutives, which naturally shift their gender into feminine. The
same mechanism yields abstract nouns and profession names from agentive nouns, some of
which have their bases borrowed from Arabic. This is for instance the case for tabnnajt and
tafllaħt. The base of the latter coexists with a native agentive noun, namely amkraz
‘ploughman’.
MA borrowed the Amazigh feminine marker along with the initial vowel a-, the
typical nominal marker, restricting their use to abstract nouns and profession names, to the
exclusion of diminutives, which undergo another type of formation (see Boudlal 2001, and
Lahrouchi & Ridouane 2016). The examples in (13), part of which is taken from Harrell et al.
(1963), illustrate the phenomenon in MA.
(13)
a. Base noun Profession
χərraz taχərrazt ‘shoemaker, shoe repairing’
fəllaħ tafəllaħt ‘farmer, agriculture’
nəʒʒarˤ tanəʒʒarˤt ‘carpenter, carpentry’
ʃəffarˤ taʃəffarˤt ‘thief, thievery’
xəbbaz taxəbbazt ‘baker, bakery’

11
 bəqqal tabəqqalt ‘grocer, grocery’

b. Base noun Abstract noun / quality

kəddab takəddabit ‘liar, lying’
ħrˤam taħħrˤajmit ‘forbidden, cunningness’
məɣrib tamɣrabit ‘Morocco, act of being Moroccan’
ʕəzri taʕəzrit ‘bachelor, bachelorhood’
rˤaʒəl tarˤʒlit ‘man, bravery’
fdˤuli tafdˤulit ‘intrusive, meddling interference’

As noticed by Zellou (2010: 235), this morpheme was borrowed into MA as an

unanalysed complex form whose components do not undergo the same morphological
alternations as in the source language. Indeed, the vowel of the prefixed part of this circumfix
remains unchanged in MA, while it obeys the general (morpho-phonological) rules of the
Amazigh grammar. For example, it undergoes deletion in the construct state (e.g. tabqqalt
‘grocery’ / n tbqqalt ‘of grocery’), and it alternates though very rarely with i- in the plural
(e.g. tanddˤamt (sg) / tinddˤamin (pl) ‘poetry’, taħħrajmit (sg) / tiħħrajmijin (pl)
‘cunningness’). Plurals of this type are even rarer in MA; the only case I have been able to
find is taħħrˤajmijat (< taħħrˤajmit (sg)), formed by means of –at suffixation, just like any
other external plural.

Another property which deserves to be mentioned relates to variation. To my

knowledge, examples of the type in (13) can be realized without schwas: a form like nəʒʒarˤ
along with its corresponding profession name tanəʒʒarˤt can be in free variation with nʒʒaˤr
and tanʒʒarˤt, respectively. Likewise, ʃəffarˤ and taʃəffarˤt can be realized as ʃffarˤ and taʃfarˤt.
This is also confirmed in the phonetic forms provided by Chtatou (1997: 113). The omission
of schwas in these variants would then stand for another phonological feature that MA has
borrowed from Amazigh, which allows complex consonant clusters without any vowel,
especially in the Tashlhiyt variety. Schwas are optionally deleted in the vicinity of obstruents,
be they voiced or voiceless, while they tend to be preserved when adjacent to a sonorant (e.g.
χərraz / taχərrazt (13a)). Extending their hypothesis on Tashlhiyt Berber syllable structure,
Dell & Elmedlaoui (2002) analyse the consonant clusters of the type just described in MA as
containing syllabic consonants.

Cross-linguistically, there is a tendency for morphological borrowing of the type

discussed above to preserve as much as possible the original form of the morphemes in the
host language (see Zellou 2011: 236 and references therein). It is worth noticing here that

12
despite its morphological unanalysability, /ta-…-t/ has evidently preserved its gender
specification in MA: all forms of the type in (13) are marked as feminine. Languages differ as
to degree of adaptation of the borrowed morphemes: For instance, Latin words like corpus,
formula and copula form their plural in English as corpora, formulae (or formulas), and
copulae (or copulas), respectively, whereas in French they remain invariable in the singular
and plural. Italian words like panini and spaghetti, which end in a plural marker, remain
unchanged in French and English, irrespective of their number. They are adapted as
unanalysed morphemes, just like MA does with the Amazigh morpheme /ta-…-t/.

5. Conclusion
In this article, I have outlined some of the intricate phonological and morphological features
which Moroccan Arabic has developed in contact with Amazigh. Based on previous work, I
have argued that Moroccan Arabic has lost the Classical Arabic short vowels, resulting in a
vocalic system where length is no longer contrastive. The language has also developed a short
central vowel, referred to as schwa, which is used to break up illicit consonant clusters. I have
shown that the distribution of this schwa is predictable, and that its alternation with zero is
better analysed within a strict CV analysis where empty vocalic positions interact laterally to
derive the surface consonant clusters. Only ungoverned empty positions surface as schwa. In
relation to this issue, I have also examined labiovelarization, a feature that MA shares with
Amazigh. I have shown that a floating /u/ surfaces as a secondary feature on the consonants
which already contain a labial feature (element |U|).
Sibilant harmony, another feature which MA borrowed from Amazigh, has been
analysed as a long distance process that is confined within a specific domain, consisting of the
stem template, plus an empty initial CV. This empty initial site allows for the Moroccan
Arabic definite article and the Amazigh causative prefix to harmonize with the stem sibilant
when no other morpheme immediately precedes the stem. Complex combinations of causative
and reciprocal prefixes in Amazigh show that only the inner prefix, which has access to the
initial site, interacts with the stem consonants: Labial dissimilation triggers the reciprocal
prefix when it immediately precedes a stem which contains a labial consonant (causative +
reciprocal + stem); the reverse order (reciprocal + causative + stem) allows the causative
prefix to agree in voicing and anteriority with the stem sibilant.
As for morphological borrowing, I have focused on the circumfix /ta…-t/, which MA
borrowed as an unanalysed complex form, whose components do not show any alternations.
In Amazigh, this complex morpheme consists of a feminine marker t, circumfixing the stem,

13
plus an initial vowel a-, which is generally analysed as a noun marker. The circumfix is also
used to derive diminutives and profession abstract nouns. In Moroccan Arabic, its function is
limited to the formation of profession and abstract nouns.

Long-lasting contact between Amazigh and Arabic has deeply impacted the structure
of the Moroccan variety, extending over all aspects of grammar, including phonetics,
phonology and morphology. Nowadays, the impact goes the other way around, with a
growing influence of Arabic on Amazigh, which operates mainly at the lexical level (lexical
borrowing), while other components of grammar are barely affected.

References
Alderete, John (1997). Dissimilation as local conjunction. In Proceedings of North East
Linguistic Society 27, Kiyomi Kusumoto (ed.), 17-31. Amherst: Graduate Linguistic
Student Association. [Reprinted in John McCarthy (ed.), Optimality Theory in
phonology: Selected readings, 394-406. Oxford: Blackwell]
Backley, Phillip (2011). Element theory, an introduction. Edinburgh: Edinburgh University
Press.
Backley, Phillip (2012). Variation in Element Theory. Linguistic Variation 12.1, 57-102
Benhallam, Abderrafi (1980). Syllable structure and rule types in Arabic. Ph.D. dissertation,
University of Florida.
Benhallam, Abderrafi (1989/1990). Moroccan Arabic syllable structure. Langues et
Littératures 8: 177-191.
Bendjaballah, Sabrina (1999). Trois figures de la structure interne des gabarits: activité
morphologique du niveau squelettal des représentations phonologiques en berbère,
somali et bedja. Ph.D. dissertation, Université Paris 7.
Bendjaballah, Sabrina (2005). Longueur phonologique des voyelles en kabyle. Etudes et
documents berbères 22: 47-69.
Ben Si Said, Samir (2014). De la nature de la variation diatopique en kabyle: étude de la
formation des singulier et pluriel nominaux. Ph.D. dissertation, University of Nice.
Bensoukas, Karim (2001). Stem Forms in the Nontemplatic Morphology of Berber. Doctorat
d’Etat Thesis, Mohamed V University, Rabat.
Bensoukas, Karim (2004). Markedness, faithfulness and consonant place in Tashlhiyt roots
and affixes. Langues et Littératures 18: 115-153.

14
Bensoukas, Karim (2014). Featural dissimilation in Tashlhit: Avoiding the repetition of labial
and round. Rabat: Mohamed V University.
Bensoukas, Karim & Abdelaziz Boudlal (2012). An Amazigh substratum in Moroccan
Arabic: The prosody of schwa. Langues et Littératures 22: 179-221.
Boudlal, Abdelaziz (2001). Constraint Interaction in the Phonology and Morphology of
Casablanca Moroccan Arabic. Doctorat d’Etat thesis, Mohammed V University,
Rabat.
Boudlal, Abdelaziz (2006/2007). Sonority-driven schwa epenthesis in Moroccan Arabic.
Languages and Linguistics 18/19: 59-81.
Boudlal, Abdelaziz (2009). The prosody and morphology of a Moroccan Arabic dialect: An
Optimality-theoretic account. Saaebrüchen: VDM Verlag Dr. Müller.
Bucci, Jonathan (2013). Voyelles longues virtuelles et reduction vocalique en coratin.
Canadian Journal of Linguitics/Revue Canadienne de Linguistique 58(3): 397-414.
Chtatou, Mohamed (1997). The influence of the Berber language on Moroccan Arabic.
International Journal of the Sociology of Language 123: 101-118.
Caubet, Dominique (2007). Moroccan Arabic. In Encyclopeidia of Arabic Language and
Linguistics, Volume 7, Lutz Edzard, Rudolf de Jong (eds.). The Netherlands: Brill.
Colon, Georges (1947). Quelques ‘emprunts’ de morphèmes étrangers dans les parlers arabes
occidentaux. Groupe linguistique d’études Chamito-Sémitiques IV.
Dell, François & Mohamed Elmedlaoui (2002). Syllables in Tashlhiyt Berber and in
Moroccan Arabic. Dordrecht: Kluwer.
Echchadli, Malika (1986). Les processus morphologiques et phonologiques des noms en
arabe marocain. MA thesis, Montréal, UQAM.
El Hankari, Abdelhak (2014). The construct state in Tarifit Berber. Lingua 148: 28-52.
Elmedlaoui, Mohamed (1992). Aspects des représentations phonologiques dans certaines
langues chamito-sémitiques. Rabat: University Mohammed V Doctorat d’Etat Thesis.
Elmedlaoui, Mohamed (2000). L’arabe marocain: Un lexique sémitique inséré sur un fond
grammatical berbère. In Études berbères et chamito-sémitiques: Mélanges offerts à
Karl-G. Prasse , S. Chaker & A. Zaborski (eds.), 155–187. Paris: Peeters.
El Moujahid, El Houssain (1997). Grammaire générative du berbère, morphologie et syntaxe
du nom en tachelhit. Publications de la faculté des lettres et des sciences humaines,
Rabat.
Ennaji, Moha (2001). The Construct State in Berber. Studies in the Linguistic Sciences 31/2:
55-72.

15
Guay, F. (1918). La forme féminine berbère à Salé. Archives berbères 3: 31-51.
Guerssel, Mohamed (1987). The status of lexical category preposition in Berber: implication
for the nature of construct state. In. Guerssel, M. & Hale, K. (eds.), Studies in Berber
Syntax. MIT, Lexicon Project Working Papers, Cambridge, MA, pp. 159-190.
Guerssel, Mohamed (1992). On the case system of Berber. Canadian Journal of Linguistics
37/2: 175-195.
Guerssel, Mohamed (1995). Berber Clitic Doubling and Syntactic Extraction. Revue
Québécoise de Linguistique 24/1: 111-133.
Harrel, Richard (1962). A short reference grammar of Moroccan Arabic. Washington, DC:
Georgetown University Press.
Harrell, Richard & Harvey Sobleman (1963). A dictionary of Moroccan Arabic: English.
Washington, DC: Georgetown University.
Harris, John (1994). English sound structure. Oxford: Blackwell.
Harris, Zellig (1942). The phonemes of Moroccan Arabic. Journal of the American Oriental
Society 62/4: 309-318.
Heath, Jeffrey (1997). Moroccan Arabic phonology. In Phonologies of Asia and Africa,
Volume 1, Alan S. Kaye (ed.), 205-217. Indiana: Eisenbrauns.
Honeybone, Patrick (2005). Sharing makes us stronger: process inhibition and segmental
structure. In Headhood, Elements, Specification and Contrastivity: Phonological papers
in honour of John Anderson, Philip Carr, Jacques Durand & Colin J. Ewen (eds).
Current Issues in Linguistic Theory 259: 167–192.

Jebbour, Abdelkrim (1996). Morphologie et contraintes prosodiques en berbère (tachelhit de

Tiznit) : analyse linguistique et traitement automatique. Rabat: Université Mohamed V,
Doctorat d’Etat thesis.

Jebbour, Abdelkrim (1999). Syllable weight and syllable nuclei in Tachelhit Berber of Tiznit.
Cahiers de Grammaire 24: 95-116.
Kabbaj, Ouadia (1990). La structure syllabique de l’arabe marocain. MA thesis, UQAM.
Kaye, Jonathan (1990). Government in Phonology: the case of Moroccan Arabic. The
Linguistic Review 6: 131-159.
Kaye, Jonathan (2000). A user’s guide to Government Phonology. Manuscript, University of
Ulster.
Kaye, Jonathan, Jean Lowenstamm & Jean-Roger Vergnaud (1990). Constituent Structure
and Government in Phonology. Phonology 7/2: 193-231.

16
Kossmann, Maarten (1997). Grammaire du parler berbère de Figuig. Paris/Louvain: Peeters.

Kossmann, Maarten (2017). Berber-Arabic language contact. In Oxford Research

Encyclpedia of Linguistics. Oxford: Oxford University Press.
Lahrouchi, Mohamed (2001). Aspects morpho-phonologiques de la dérivation verbale en
berbère tachelhit. PhD dissertation, University of Paris 7.
Lahrouchi, Mohamed (2003). Manifestations gabaritiques dans la dérivation verbale en
berbère tachelhit. Recherches Linguistiques de Vincennes 32 : 61-82.
Lahrouchi, Mohamed (2013). Templates, markers and syntactic structure in Tashlhiyt Berber.
Lingua 133: 53-72.
Lahrouchi, Mohamed (2018a). Syllable structure and vowel-zero alternations in Moroccan
Arabic and Berber. In The Routledge Handbook of African Linguistics, Augustine
Agwuele & Adams Bodomo (eds.), 168-180. Oxon, UK: Routledge.
Lahrouchi, Mohamed (2018b). The left edge of the word in the Berber derivational
morphology. Glossa: a Journal of General Linguistics 3/1: 1-25.
Lahrouchi, Mohamed & Rachid Ridouane (2016). On diminutives and plurals in Moroccan
Arabic. Morphology 26/2: 453-475.
Lahrouchi, Mohamed & Philippe Ségéral (2010). Peripheral vowels in Tashlhiyt Berber are
phonologically long: Evidence from Tagnawt. Brill’s Annual of Afroasiatic Languages
and Linguistics 2: 202-212.
Lahrouchi, Mohamed & Shanti Ulfsbjorninn (2018). Formalising Headedness-Driven
Phonotactics: English, French, Tarifit and Tashlhiyt Berber. PC, Colloque du Réseau
Français de Phonologie, Paris.
Lowenstamm, Jean (1991). Vocalic length and centralization in two branches of Semitic. In
Semitic Studies in Honor of Wolf Leslau on the Occasion of his 85th Birthday, vol. 2,
Alan S. Kaye (ed.), 949-65. Wiesbaden: Harrasowitz.
Lowenstamm, Jean (1996). CV as the Only Syllable Type. In Current Trends in Phonology
Models and Methods, Jacques Durand & Bernard Laks (eds.), 419-442. University of
Salford: European Studies Research Institute.
Ouhalla, Jamal (1988). The syntax of head movement: A study of Berber. PhD dissertation,
University College, London.
Ouhalla, Jamal (1996). The construct state in Berber. In Studies in Afroasiatic Grammar,
Jacqueline Lecarme, Jean Lowenstamm & Ur Shlonsky (eds.), 278-301. The Hague:
Holland Academic Graphics.

17
 Scheer, Tobias (1996). Une théorie de l’interaction directe entre consonnes. PhD dissertation,
Université Paris 7.
Scheer, Tobias (2004). A Lateral Theory of Phonology: What is CVCV, and why should it
be?, Studies in Generative Grammar 68/1. Berlin: Mouton de Gruyter.
Taïfi, Miloud (1991). Dictionnaire Français-Tamazight. Paris: L’Harmattan.
Tilmatine, Mohand (1999). Substrat et convergences: le berbère et l’arabe nord-african.
estudios de dialectogía nortafricana y andalusí 4: 99-119.
Watson, Janet (2011). The phonology and morphology of Arabic. Oxford: Oxford University
Press.
Zellou, Georgia (2010). Moroccan Arabic consonant harmony: A multiple causation
hypothesis. Toronto Working Papers in Linguistics 33: 1-15.
Zellou, Georgia (2011). Moroccan Arabic borrowed circumfix from Berber: Inverstigating
morphological categories in a language contact situation. In Les frontières internes et
externes de la morphologie, Gregor Perko (ed.), 231-244. Ljubljana: Znanstvena
založba Filozofske fakultete.

18

View publication stats