OCP 7
Nice, 30/01/2010
Template structure in Berber: restrictions on derivations Mohamed Lahrouchi CNRS – University Paris 8 UMR 7023
[email protected]
I. Templates (1) McCarthy 1979, 1981 Many templates:
Classical Arabic verb derivation
√ktb ‘write’ Forme I
a. CVCVC
b. katab
Forme II
CVCCVC
kattab
Forme III
CVVCVC
kaatab
Forme IV
CVCCVC
/aktab
Forme VII C C V C V C
nkatab
(2) Guerssel & Lowenstamm 1990 One template: The various forms the verb displays in CA are derived by means of a unique template composed of CV units, some of which play a grammatical role. Head positions
C VAff C V C VDS C V C V The italicized syllables are morphological heads (CVAff = affixal syllable; CVDS = derivational syllable). The boxed syllables constitute the complement of the head. The identification of one of the head positions by some lexical material amounts to derivation. (3) kattab ‘he made write’/ kaatab ‘he corresponded’ / nkasar ‘it broke’ a.
k t | | C V C V C V C
b | V C V
b.
| a c.
n
k t b | | | C V C V C V C V C V | a
n
k s r | | | C V C V C V C V C V | a
1
(4) Only one head position is identified in each form. The identification of both head positions leads to illformedness. e.g.
kasar ‘he broke’ / kassar ‘he shattered’ (intensive) nkasar ‘it broke’ / *nkassar ‘it shattered’
*nkassar is illformed because headed twice (both morphological heads are identified in a single base). (5) Strict CV approach to syllable structure - “[…] the syllable structure of all languages reduces to CV.” (Lowenstamm 1996: 419) - The skeletal level of phonological representations consists of strict alternations of onset and nucleus positions, i.e. C and V positions. - The differences in the surface syllable types lie in the lateral relations that segments share. Proper Governement is one such relation which allows a vocalic position to remain empty when followed by a vowel. (See Scheer 2004 for details and discussion) II. Templatic constraints (6) Purpose - Show that the structure of the template proposed by Guerssel & Lowenstamm 1990 not only offers the tools to accounts for a range of non-concatenative morphological operations but also captures the distributional restrictions these operations obey. - Account for co-occurrence restrictions that a number of nominal and verbal affixes obey in Tashlhiyt Berber. II.1. State, gender, number1 (7)
FS
CS
MS
FM PL
FM
PL
SG
PL
‘cat’
amuSS
imuSSa
tamuSSut timuSSa
umuSS
imuSSa
tmuSSut tmuSSa
‘boy, girl’
afrux
‘pigeon’
ifrxan
tafruxt
tifrxin
ufrux
ifrxan
tfruxt
tfrxin
atbir
‘reed, fishing rod’
itbirn
tatbirt
titbirin
utbir
itbirn
ttbirt
ttbirin
a“alim i“alimn ta“alimt
ti“alimin u“alim i“alimn t“alimt
t“alimin
‘ox, cow’
afunas
tifunasin
tfunasin
ifunasn
SG
MS
SG
tafunast
1
ufunas
ifunasn
SG
tfunast
Abbreviations: FS = free state, CS = construct state, MS= masculine, FM = feminine, SG = singular, PL = plural, Nom. = nominative, Acc. = accusative.
2
PL
(8) Distribution of markers Gender: t prefixed and suffixed to the stem Number: a- (SG) alternates with i- (PL)
-n (PL) suffixed to the stem States:
Free State vs. Construct State (Bound state) FS MS
FM
MS
FM
SG
a-
t-a-
u-
t-
PL
i-
t-i-
i-
t-
(9) State alternation ufrux a. iSSa eat:3MS
CS
a“rum
boy.CS-NOM
bread.FS-ACC
‘The boy ate bread’ b.
afrux
iSSa
a“rum
boy.CS-NOM
eat:3MS
bread.FS-ACC
‘The boy ate bread’ c.
ittSa
u“rum
eat-passive:3MS
bread.CS
‘Bread was eaten’ d.
a“rum
/n
ufrux/ → [uufrux]
bread-FS
of
boy-CS
‘The bread of the boy’ ‘The boy’s bread’ e.
iSSa
tijmi
s
u“rum
eat:3MS SG
sauce-FS
with
bread-CS
‘He ate sauce with bread’ f.
tlla
tammnt
“
u“rum
be:3FM SG
honey-FS
in
bread-CS
‘There is honey in the bread’ g.
tammnt
d
u“rum
honey-FS
and
bread-CS
‘Honey and bread’ h.
sin
ifrxan
two
boy-CS, MS, PL
‘Two boys’
/
snat
tfrxin
two
girl-CS, FM, PL
‘two girls’
3
CS:
FS:
- Subject NPs in VSO sentences - NPs Complements of light prepositions, d ‘and’, s ‘with, by’, “ ‘in’ - NPs complements of quantifiers jan ‘one’, sin ‘two’, etc. Elswhere
Several works on state alternation: see, among others, Achab 2003; Bader & Kenstowicz 1987; Bendjaballah & Haiden 2005; Dell & Jebbour 1991; El Moujahid 1997; Ennaji 2001; Guerssel 1992b, 1995; Ouhalla 1988, 1996. (10) Distribution of state markers: - Why the FM marker t- never co-occurs with the CS marker u-? - Why afrux vs. ufrux, but tafruxt vs. tfruxt (not *tufruxt)? (10) State / Case Guerssel 1992b: 175 “[…] the concept of State has no validity as a theoretical notion, […] what has been labeled a Construct State form is either a Determiner Phrase, or a Kase Phrase where the head K is not realized”.
a- is a kind of porte-manteau morpheme See also Ouhalla 1988. (11) Prosodic deficiency of the CS (Bendjaballah & Haiden 2005) axxam vs w´xxam ‘house’ a. CV
EL : axxam CV CV CV CV CV a
x
a
b. CV
m
EC : w´xxam CV CV CV CV CV w
x
FS a- identifies the first two CV units of the template (11a) CS w- is associated to only one CV unit, resulting in a prosodic deficiency (11b)
4
a
m
(12) When complement to light prepositions, the empty CV in the template of CS serves as a phonological host of light preposition. /g w´xxam/ → ggw´xxam ‘in the house’ P D
P
(13) a.
D
| CV
| CV
g
w
N
CV
CV
x
C V CV a
m
El Moujahid (1997:163) “en théorie, il est difficilement soutenable de conférer à une même catégorie plusieurs fonctions.” a-frux ‘boy’
/
ta-fruxt ‘girl’ ta
u-frux
Case markers Det Gender
b.
/
t-frux-t
Det Gender
Nouns with non-alternating vowels (see Dell & Jebbour 1991) e.g. FS CS ajjis wajjis ‘horse’ adgal
wadgal ‘widower’
akal
wakal
‘earth, ground’
ajdi
wajdi
‘dog’
- Where is the Det? Where is the Case marker? - Where is deficiency? In the CS, in the FS, or in both? c.
- Why is it that u- (or w-) never co-occurs with t-? Why CS tfruxt and not tufruxt?
(14) Proposal:
The co-occurrences restrictions that the CS, Gender and Number markers show follow directly from the structure of the template.
5
t- and w- compete for the same position in the template.2
(15) a.
w
b. f r x | | | C V C V C V C V | u
t
c.
d. f r x | | | C V C V C V C V | | a u
t
t f r x | | | | C V C V C V C V C V | u
f r x t | | | | C V C V C V C V C V | | a u
-
CVDS is omitted since unidentified CS ufrux and tfruxt appear in (15a) and (15b)
-
FS afrux and tafruxt appear in (15c) and (15d)
-
Only one affixal C is available in the initial position. Gender t- takes precedence over CS w- (15b and 15d). When t- is absent, w- appears (15a).
-
Parallel is drawn: in the CS between t- and w-, and in the FS between a- and ta(see also Achab (2000: 8). Given 15c and (15d), the parallel should even be drawn between wa- and ta-, since the empty initial C in (15c) can host w-, leading to unattested *wafrux. In Tashlhiyt Berber, many words still exhibit the prefix wa-: e.g. wabiba ‘mosquito’, wabʒir ‘mallow’, waʃʃ ‘awful’, wafud ‘knee’, wakuz ‘weevil’, wamsa ‘anise’, wazzwit ‘afternoon tea’. Some of them show variants without glide: afud, akuz (Further examples are provided in Brugnatelli 1998).3
The plural marker i- appears in the FS feminine form tifrxin, but not in the corresponding CS feminine form tfrxin, whereas in the masculine, FS and CS are identical. If we proceed as in (15), and assume that t- and i- compete for the same position in the template, leading to the CS form tfrxin, we should explain why these morphemes coexist in the FS form tifrxin. This is figured in (16).
w surfaces as [u], since followed by a consonant. In Tashlhiyt Berber, any word-initial glide (w, j) surfaces as a high vowel ([u], [i]) when followed by a consonant, and remain unchanged when followed by a vowel (e.g. ifta ‘he went’ / jufa ‘he found’). This is also true word-finally, a glide surfaces as a high vowel when preceded by a consonant (e.g. kmi ‘smoke’, xlu ‘destroy’); they remain unchanged when preceded by a vowel (e.g. akmmaj ‘smoker’, amxlaw ‘mad’). 3 Vychil 1957, and Brugnatelli 1997 and 1998 reconstruct wa- as a determiner/demonstrative. Currently, wa- is 2
used as a vocative.
6
(16) a.
I
f
r
b.
x
t
| | | C V C V C V C V |
f
r
x
n
| | | | C V C V C V C V C V | i
a c.
t
f r x | | | C V C V C V C V | | i i
FS and CS ifrxan is represented in (16a), CS tfrxin appears in (16b), FS tifrxin in (16c).
One way to resolve the problem and thus unify the analysis is to assume that CS plurals use the same morpheme as CS singulars: w-. Indeed, there is no need to posit different CS markers in the singular and plural forms, masculine and feminine. To the FS masculine plural ifrxan corresponds the CS /wifrxan/. Surface homophony results from the loss of w- and the association of the following I to the initial C, leading to /jfrxan/ which surfaces as [ifrxan]. (17) illustrates the opposition ifrxan vs. tfrxin. (17) a.
jfrxan
b.
→
f
r
x
n
|
|
|
|
t
C V C V C V C V C V I
(18)
tfrxin f
r
x
n
|
|
|
|
C V C V C V C V C V |
| a
I
i
tafruxt vs. tifrxin (not *tifrxtin)
II.2 Geminated Imperfective As an imperfectivizing mechanism in Tashlhiyt Berber, gemination concerns verbs containing no more than three consonants and no full vowels, as well as verbs with the following shapes: CCU, CCI.
7
(19)
Aorist
Imperfective
a.
lkm
lkkm
‘arrive’
kSm
kSSm
‘enter’
gwmr krz
gwmmr kkrz
‘hunt’ ‘plough’
xrb
xxrb
‘scratch’
xwmʒ
xxwmʒ
‘scrape’
b.
Previous accounts of Geminated Imperfective: Dell & Elmedlaoui (1988, 2002); Bensoukas (2001), Jebbour (1996, 1999), Lahrouchi (2008), among others. (20) Within template morphology, gemination results from the identification of CVDS by means of C-spreading. Medial consonant gemination obtains as figured in (20). (20)
l k | | C V C V C
m | V C V
-
CVAff is omitted since unidentified.
-
Segments are associated with the template from the edges-inward (Yip 1988), resulting in the gemination of medial consonant.
(21) Sonority effect: Any segment is prohibited to geminate in the imperfective if it is the most sonorous segment in the root. (See Lahrouchi 2008: 35 for discussion) Aorist C | k
V
C | r
V
Imperfective C | z
V
a.
C | k
V
C
V
C | r
V
C | z
V
b.
C | k
V
C
V //
C | r
V
C | z
V
c.
C | k
V
C
V
C | r [kkrz]
V
C | z
V
(22) Action nouns A uniform pattern: aCCaC, where the medial consonant is either simple or geminated depending on its sonority. 8
a.
b.
Verb
Action Noun
‘lend’
rd≥l
artt≥al
‘swell up’
bzg
abzzag
‘hide’
ntl
anttal
‘tighten’
frg
afrag
‘be ashamed’ mrg
amrag
‘sort out’
afran
frn
In the first group of verbs the medial consonant is an obstruent preceded or followed by a more sonorous segment, while in the other group it is a sonorant surrounded by less sonorous segments. In the corresponding action nouns, the first group geminates the medial consonant, and not the second group. II.3 Causative Imperfective - Causative verbs are built by means of a monoconsonantal prefix s- attached to the stem. Depending on the properties of the stem, the prefix is realized as a single or geminated segment. It is argued in Lahrouchi (2003) that an initial templatic site is responsible for the size variation of the prefix. - The initial templatic site here coincides with CVAff in (4). - In the Imperfective, Causative forms all use vowel insertion; while their bases use gemination or affixation (see examples in (23).
(23)
Verb
Causative
Aorist
Imperfective
Aorist
Imperfective
lkm
lkkm
sslkm
sslkam
*slkkm
‘stand up’
nkr
nkkr
ssnkr
ssnkar
*ssnkkr
‘hide’
ntl
nttl
ssntl
ssntal
*ssnttl
mun
ttmuna
smun
smuna
*ttsmuna
‘sit down’
gawr
ttgawar
sgawr
sgawar
*ttsgawar
‘change’
badl
ttbadal
sbadl
sbadal
*ttsbadal
a. ‘arrive’
b. ‘pick up’
9
The stared forms are of two types: -
Forms in (23b), where imperfective and causative prefixes compete for the same
position in the template, i.e. CVAff. -
Forms in (23a), which involve the identification of both head positions (CVAff and
CVDS) at the same time, leading to an undesired multi-headed structure. (24) sslkm
a. s
l | | C V C V C V C
sbadal
b. k | V C
s b d l | | | | C V C V C V C V C V | | a a
m | V C V
II.4 Inchoatives Inchoative verbs show a geminated consonant that geminates when preceded by a derivational morpheme, such as causative s-, reciprocal m-, passive tt(u)- and imperfective tt-. (25)
Inchoative
Causative
‘sit down’
ggawr
sgawr
‘be friend’
ddukkl
sdukkl
‘be disgused’
mmuktu
smuktu
‘be rotten’
llugmu
slugmu
‘stand up’
mmatti
smatti
Prefix s- prevents initial consonant form geminating, as shown in (26). See also Guerssel 1992a.
ggawr
(23) a.
g
sgawr w
| | C V C V C V C | a
b.
r | V C V
s
g
w
r
| | | | C V C V C V C V C V | a
III. Conclusion The template captures the distributional restrictions that certain nominal and verbal morphemes undergo. The structure assigned to the template defines the way forms are derived: (i) each form must be headed (i.e. one of the derivational head positions in the template must be identified), (ii) multi-headed forms are prohibited (both derivational heads cannot be identified in a single form).
10
References Achab, Karim. 2003. Alternation of State in Berber. In Research in Afroasiatic Grammar II, Jacqueline Lecarme (ed.), pp. 1-20. Amsterdam: John Benjamins. Bader, Yousef & Michael Kenstowicz. 1987. Syllables and Case in Kabyle Berber. Lingua 73. pp. 279-299. Bendjaballah, Sabrina & Martin Haiden. 2005. The Grammar of Prepositions in Berber (Taqbaylit). Ms. CNRS et Université Lille 3. Bensoukas, Karim. 2001. Stem Forms in the Nontemplatic Morphology of Berber. Université Mohamed 5, Rabat: thèse de doctorat d’Etat. Brugnatelli, Vermondo. 1997. L'état d'annexion en diachronie. In A. Bausi, M. Tosco (éd.), Afroasiatica Neapolitana. Incontro di Linguistica Afroasiatica (Camito-Semitica). Université de Naples – L’orientale. pp. 139-150. Brugnatelli, Vermondo. 1998. La morphologie des noms berbères en w-. Considérations diachroniques. In Mohamed Elmedlaoui, Saâd Gafaiti & Fouad Saa (eds), Actes du Premier Congrès Chamito Sémitique de Fès (12-13 mars 1997). Saïs-Fès : publications de la faculté des lettres et des sciences humaines. pp. 51-67. Dell, François et Mohamed Elmedlaoui. 1985. Syllabic Consonants and Syllabification in Imdlawn Tashlhiyt Berber. Journal of African Languages and Linguistics 7. pp. 105-130. Dell, François & Mohamed Elmedlaoui. 1988. Syllabic Consonants in Berber: Some New Evidence. Journal of African Languages and Linguistics 10. pp. 1-17. Dell, François & Mohamed Elmedlaoui. 2002. Syllables in Tashlhiyt Berber and in Moroccan Arabic. Kluwer Academic Publishers, Netherlands. Dell, François & Abdelkrim Jebbour. 1991. Phonotactique des noms à voyelle initiale en berbère (chleuh de Tiznit, Maroc). Linguistic Analysis 21. pp. 119-147. El Moujahid, El Houssaïn. 1997. Grammaire générative du berbère, morphologie et syntaxe du nom en tachelhit. Rabat : publications de la faculté des lettres et des sciences humaines. Ennaji, Moha. 2001. The Construct State in Berber. Studies in the Linguistic Sciences 31/2. pp. 55-72. Guerssel, Mohand. 1992a. The Phonology of Berber Derivational Morphology by Affixation. Linguistic Analysis 22. pp. 3-60. Guerssel, Mohand. 1992b. On the Case System of Berber. Canadian Journal of Linguistics 37/2. pp. 175-195. Guerssel, Mohand. 1995. Berber Clitic Doubling and Syntactic extraction. Revue québecoise de linguistique 24/1: 111-133. Guerssel, Mohand & Jean Lowenstamm. Classical Arabic Apophony. Ms UQAM et Université Paris 7. Iazzi, Elmehdi. 1991. Morphologie du verbe en tamazight (parler des Aït Attab Haut-Atlas Central) : approche prosodique. Université Mohamed 5, Rabat : thèse de DES. Jebbour, Abdelkrim. 1996. Morphologie et contraintes prosodiques en berbère (tachelhit de Tiznit) : analyse linguistique et traitement automatique. Université Mohamed 5, Rabat : thèse de doctorat d’Etat. Jebbour, Abdelkrim. 1999. Syllable Weight and Syllable Nuclei in Tashlhiyt Berber of Tiznit. Cahiers de Grammaire 24. pp. 95-116. Lahrouchi, Mohamed. 2003. Manifestations gabaritiques dans la morphologie verbale du berbère (parler chleuh d’Agadir). Recherches Linguistiques de Vincennes 32. pp. 61-82. Lahrouchi, Mohamed. 2008. A Templatic Approach to Gemination in the Imperfective Stem of Tashlhiyt Berber. Studies in African Linguistics 37/1. pp. 21-60. Lowenstamm, Jean. 1996. CV as the Only Syllable Type. In Jacques Durand & Bernard Laks (eds), Current Trends in Phonology: Models and Methods, volume 2. Salford: European Studies Research Institute, University of Salford. pp. 419-441. Lowenstamm, Jean. 1999. The Beginning of the Word. In John Rennison & Klaus Kühnhammer (eds), Phonologica. The Hague: Holland Academic Graphics. pp. 153-167. McCarthy, John. 1979. Formal Problems in Semitic Phonology and Morphology. Massachusetts, MIT: PhD. dissertation. McCarthy, John. 1981. A Prosodic Theory of Nonconcatenative Morphology. Linguistic Inquiry 12. pp. 373-418.
11
Ouhalla, Jamal. 1988. The Syntax of Head Movement: a Study of Berber. University College, London: PhD. dissertation. Ouhalla, Jamal. 1996. The Construct State in Berber. In Jacqueline Lecarme, Jean Lowenstamm & Ur Shlonsky (eds.), Studies in Afroasiatic Grammar. The Hague: Holland Academic Graphics. Scheer. Tobias. 2004. A Lateral Theory of Phonology. Vol. 1: What is CVCV, and why should it be? Studies in Generative Grammar. Berlin: Mouton de Gruyter. Vycichl, Werner.1957. L'article défini du berbère. Mémorial André Basset, Paris. pp. 139-146. Yip, Moira. 1988. “Template morphology and the direction of association”. Natural Language and Linguistic Theory 6: 551-577.
12
