:v

 

 

F.

3

:

whu

L ma. .w
5,7,.
.1? If”.

. art;

1‘1. ‘ fer. ;7_.b‘xw:l.

. A . v; ,
Ew.n,u...~..lzii.32.
' ax. VItffi)-.ll
2:1: .
xfltt1

x »31 a

6v

Iv

O4. ‘u:2.t1?..1y7 ,
zrrur ;

¢.r‘w

.11..

.é;

5
.6

.353 '

33.5:
:31 fl?
' §

c

‘ -
3:32
v.10?!
‘
1 ,
Q 9
s51
L: «2%
fir 5:
1g: "t’
‘ .
:3.
3‘

I
V?
.'3
:1

y

.
”=1“
l,:
5,:
:ie
3‘ '

‘
“t-

. ,7 .
"N ~
n’fihh‘a'fih
he. 9,1,!J, ,
v v 71 “r3“;
.1,1.x,v , '2‘.
’v' I.’ - r! 5
'7’ '73! z‘u‘:
“‘7’ ‘0‘ '
c595?” »
‘ f ., -
139%
‘ ,
, .
i v. w
r15, #1:.
7 Traps
V y“? .
‘ '
‘ F V. I i
If‘tl;1"1q.
9"4'3

, 1
gy,:
.p
0?:
'
u 2
17.7'

‘R

a

,1
'7

:3 . _ 2:? ,

s . xw‘mvfimfim
:«u wwquherhrmmnfiwr
twili 7a....
afi
31..

‘w

'7
9s 37 w?
r:

«3’.
32.321111

"$1 _ ,
. 4d: .1
fir». $5 .
«wave 3.} _
4%.“;
2;»...
:1, a.»

am:

a
v
la-
“9::
- '
w‘flé

L‘L'LI,

rs

+§LU
r.

’3 4‘ fig
m5 3-... n
;1' 'ugagagj‘s

L2: II t L

1 b . .

agitaaa‘
4 q .

313

3;";

S

.-

_‘. ;, 3

W11
(-2

,. .
. wag...‘

.
:.?..2..:;
{5:2 :53. a:
:11.-

7:

.a v

u

r." a; 1:“
“z: 2 ‘
:c: NJ

3

.

.75... u; q in y
a. 1: 3 .31.?! .551 {if
fitmi“ffil‘_flaif ,WIA‘? .
1‘15? ”damn? Ruhr... ...
Eza 1.. i}?
:g.:.a
- 915.)?

v;

R .
ti
,.

‘. “ 7”,
LIN-g:
w“?’= y
51,:
v i
g .

t...

.3113}:
5. 177193.. ‘ than 4.... v I
11:» . I , L
.1..1.\..v..r. muslin.

. w»
y 1
r

:5:
‘v

32

s

. . . 1;?
,(,l., I~37¢
HRH: “I";

3.:
an 33.3? t7}
..u.:.é_:.n.1.;h .vizf .
svaaxv4ni1ni.rel-13:12.5...1.
31.53.; .21

a“:

flirt-lira _
w.v\..:,\an:§>¢ :0: .v‘

7.1.4.2... >yalv'vv’1wiayxa :2 7'2):

rrvaiz... 3151;133:717 l>13>1aޤw3

3,. ..zzcatfl..!...1u¢'cir:_;v7a‘15; ..

:21}

£12.. .1, 3.: V._. :1). :
3.17.353.3.5.3.1117 .\ r a: at
. i,“

:1): 22.1:
I: ,4. $235

 

.ry>v¢.... pik'? 2-x . ‘1

 

 

 

L

 

1

V

«

Hz

.1: 41.3...
31:2? ‘13.
.y’:in...1pé;ar;..:11

.r

$.11 4 . I b
4.91.? .

«iv:

w

i?
.k
s 30.....vézu
$5? . , 3
:fiiizisi

r .Hafi..t.,.!w
r. (yin-L

”3‘

5.9,.2}...
3.22.1.3:
.3:

.m ‘
2‘5.

‘1
l2 s
‘5
r”.
{I

‘
E.
A
#9

A?

3*“

v" A; ‘ 3:
a‘i‘g‘a'
Evy-é

E
‘l ,

3.7
fi

.
I
L!
{'3

A...

E A
3391
szxc.
.'. a
2%.: ~
”Wt

.

1;.
‘-

a:

as

:;

4’

A:

n. {35.4 Ft».
3.." , I.

 

 

K'é '1".
L 369% an
.‘5‘ |§€ :, A
”a";

1:

.3 .xabnefigianl
attalhl.
15.: 5.215.!
at
Lotnfilxbik! l
4.1:... c...>)at.)«|s.
51%))...3
nuittf-hax.
Fin.
.Atiiltiit 4

 

ll“llllllllllllllll »

LIBRARY

Michigan State
University

 

This is to certify that the ‘ l

thesis entitled

BROKEN PLURALS IN MODERN IRAQI ARABIC \

presented by

Theresa McLaughlin Al—Azzawi

has been accepted towards fulfillment
of the requirements for

Ph.D. degree in Linguistics

 

Maijfo
“39

Datew

0-7 639

 

 

 

MK“; I
04'8‘73
0 A 124

1

 

 

 

 

llOKEl

l‘ne prl
lemma
my p11
unever ‘

file. l

ABSTRACT
BROKEN PLURALS IN MODERN IRAQI ARABIC
By
Theresa McLaughlin Al-Azzawi

The problem of broken plurals (i.e.. internal or
interdigitated plurals, as opposed to the so~ca11ed
'sound' plurals. realized as inflectional suffixes)
has never previously been solved satisfactorily for
Arabic. Mary M. Levy (The Plural of the Noun in Modern
Standard Aggbic, Unpublished Ph.D. Dissertation, Univer-
sity of Michigan. 1971). while predicting the correct
Plural(s) for any given singular item via phonological
rules and devices to handle exceptions, does not offer
explanation for the peculiarities of the system (e.g.,
multiple plural association).

In this work'the hypothesis posited by Talmy Givén
("Some Historical Changes in the Noun Class System or
Bantu: Their Possible Causes and Wider Implications” in
W, Chun Wu Kim and Herbert
Stahlke, eds., Urbana, 1971) to account for the noun
class system of Bantu is used as a basis for the research
into the apparently chaotic pluralization system of Arabic.
The study is based on one subdialect of Arabic, namely
Moslem Baghdadi Iraqi, which is a Bedouin dialect.

 

ill: at
minted 1
author of
fall on
malaria
in, non-
:a-to-one
Plural 015
ill! of 1
morale]

lie the

tampon

Elle 'g

llie

W— 7

Theresa McLaughlin Al-Azzawi

This study finds that classes of Arabic nouns are
associated with particular broken plurals on two bases.
A number of plural classes are associated with broken
plurals on the basis of culturally perceived semantic
characteristics. This phenomenon reflects an older
n-ary. non-anthropocentric. non-hierarchical system of
one-to-one semanto-morphological correspondence. A few
plural classes. are associated with broken plurals; on the
basis of phonological shape, or canonical form. This
phenomenon reflects a partial rearrangement of the system
since the time of the hypothesized change to a binary,
anthropocentric, hierarchical structure. which gave rise
to the 'sound' grammatical masculine/feminine dichotomy.

The theory on which the study is based is that of
stratificational grammar. One of the requirements of the
theory. that ”units" on one level (or stratum) are re-
lated only to “units" on immediately contiguous level(s)
is found to be too stringent. A less rigid model is
adopted in order to account for the direct relationships
which seem to exist between the gnostemic and the morphemic

strata of Iraqi Arabic.

 

 

 

BROKEN PLURALS IN MODERN IRAQI ARABIC

By

Theresa McLaughlin Al-Azzawi

A DISSERTATION

Submitted to
Michigan State University
in.partia1 fulfillment of the requirements
for the degree of

DOCTOR OF PHILOSOPHY

Department of Linguistics and Oriental
and African Languages

1973

 

 

 

 

Committee Members:

Dr. David G. Lockwood. Chairman

Dr. Ruth M. Brend

Dr. Julia S. Falk
Dr. Ernest N. McCarus
Dr. Irvine Richardson

 

 

Copyright by
THERESA MCLAUGHLIN AL—AZZAWI

1973

 

llth

uh:

[ht

DEDICATED
especially to

my husband
ASIM HASHIM AL-AZZAHI
without whose continual encouragement
and mature understanding I would never
have accomplished this end

and to
JULIA SABLESKI FALK
who unknowingly caused me to grow in
maturity and understanding
and to
WILLARD AND MARGUERITE BEECHEB
whose book, Beyond Success and Failure,
inspired me to accept academia for
what it is and myself for what I am
and to
the chairman of my dissertation committee
DAVID GEORGE LOCKWOOD
a learned teacher and a friend
and ultimately to
my mother and father

KELEM MAKER AND ROBERT DORSEY MOLAUGHLIN

11

So oft in theologic wars.
The disputants. I ween,
Rail on in utter ignorance
or what each other mean
And prate about an Elephant

Not one of them has seen:

from the poem
The Blind Men and the Elephant
by John Godfrey Saxe

111

ACKNOWLEDGMENTS

I wish to thank the members of my thesis
committee Dr. Ruth Brend. Dr. Julia Falk.

Dr. David Lockwood. and Dr. Irvine Richardson
of Michigan State University and Dr. Ernest
N. McCarus of the Department of Near Eastern
Languages and Literature of the University

of Michigan for their spirit of cOOperation
in the prompt return of comments and sug—

gestions.

iv

 

:st of lab?
ll of Fig
‘1:th Syi'

his:
I. llll

TABLE OF CONTENTS

List of Tables .
List of Figures

0 e e e e e O O '
ist of Symbols and Abbreviations
Chapter

I. INTRODUCTION

1.1 The Topic

1.2 Previous Scholarship . e
1-3 Modern Iraqi Arabic . .
1.“ Data 0 e e o o e c o e 0

II. PRELIMINARIES .

2-1 Thioietical Approach to the Topic

0 0 Background . e e

2.1.2 Axioms and Primitives .
2-1.3 Types of Relationships
2.1.4 The Simplicity Measure

2.1.5 Major Goal 0 e e
2.2 Semology . .

2.2.2

2.2.1 The Gnostemic and Sememic
Componential Analysis . .

2020201 Definition 0 e 0
2.2.2.2 Universal Status

2.2.2.3 Related Phenomena
2.2.2.h Cognitive Validity

Markedness . . .

2.2.2 O O C .
2.2. The Gnostotactics . . . e . .
202.5

.0...

Strata

The Semotactics . . . . e - .

2-3 Strata Intervening Between the

Phonemic . . . . . . o .
2.3.1 Lexemic .

O

20302 Morphemic e e a e e
2.3.3 The Relationship of

System to the Morphemic
2.3.h The Relationship Of the

the

Sememic

Gnostemic
System

.000.-

Morphemic
System to the Phonemic System . .

tic

2.u Ph°n°108¥ e o e e e e e e O
2.h.1 The Phonemic and Phone
2.h.2 Componential Analysis .
2.“.3 Markedness . . . o . .
2.“.h The Phonotactics . . -
2.u.5 The Phonetic Tactics .

V

Strata

xiii

xix

\OCDOHd‘ H

A GENERAL DESCRIPTION OF IRAQI BROKEN PLURALS

301 G3
302 Th

neral Remarks -- Language in Flux . o '
e Old Semantic Classification

3.2.1 Semantic Taxonomy (Arransed by

Phonological Shape)

3.2.1.1

3.2.1.2

3.2.1.3

3.2.1.14“

3.2.1.5

3.2.1.6

3.2.1.7

3.2.1.8
3.2.1.9

vi

FMaaL Plural Class
3.2.1.101 FVVL o
3.2.1.1.2 FvML(a)
1.3 FvaL .

O O O O C O

0201.

02.1.1.“ FVMVVL

MaaLa Plural Class . . .
02. 02.1 FVVL c o o o I
02. 02.2 FVML(&) o o o a
020 02.3 FVLV o o o o O
020 92.4 FVMVL . o o o I

3.1 FVMLV o o o a
3.2 FVVMLa . . .
3.3 Fvai(yya) .

MaaMiL/FuMaaMiL Plural

i Plural Class .

2.1.u.1 F(i)hvaa . . .
2.1.4.2 FUMMVL

tot-4P
SD
co
m

M knuJo*d \nanoEg unaxau>w on»
HHHh HHHH

iMaaMiiL Plural Class . .
3.2.1.5.1 FvMMaaL(a)

Round 0 o o o o
3.2.1.5.2 FvMMaaL(a) Male
FuMaaMiiL Plural Class . .
3.2.1.601 FVMMVVL Round 0
3.2.1. .2 FvMMva Square

1

3

(“POBJNHVFQAJNTOAJg

Hz HHHHHHHHHHz

Plural Class . . .
l FWWWMa)Rmmd
2 FvavL(a) Round
3 FMva(a) Round
h FvaLa Round .
5 FvaML(a) Round
6 FvavL(a) Round
7 FvMMvL(a) Square
8 FvavL(a) Square
9 FMva(a) Square
.10 FvavL(a)

I o o o. o so 0 o
no .0 o oo o. o
o o o. o o. o o.

b ‘TQ‘Q‘TQ‘Q\TQ‘QN3h O\
o o. 0 on o 0.

Square

u Plural C1883 o o o
0801 FUMLV(V)L o o o

a)?
9

Plural Class . . . .
1.9.1 F(v)Maaya . . -

h3W M)M MMJ\»UMJ\JHHJ\DM3w

 

 

 

 

 

 

 

3.2.1.10

3.2.1.11
3.2.1.12

30201.13
30201.1“
30201015

30201016
3.2.1.17

30201018

3.2.1.19
302.1020

3.201021

vii

FiMi Plural Class . . . . .
3.2.1.10.l FaML(aaya)

FaMLaat Plural Class

FuMuuL Plural Class . . .
FvaL Strong

3.2.1.12.1

iMaaL Plural Class

1.1h.1
2H1 1h. 2

FvaL Bulk
FvML Bulk

FvML Strong . .
0

FvaL Abstract
FvML Abstract .

F:

H
'21 E.
<3.
E...-
Arm
4 a
you
.0.

FVMVL o

FVMIIVOOO

FiMaaLa Plural Class
FaMaaMuL Plural Class

FaMaaMiL Plural Class

3. 2 .1.17. 1

M17 2
.2 1.1

uuwuwm.
NNNN

O O O O O
HHHH

C

HHHH
vvvvq
\10\U\1='Cn

3.2.1.17.8
3.2.1.17.9

?aFMiLaa? Plural Class
FBMi o o o

3.2.1.18.1
3.2.1.18.2

FvMMvL(a)
FvavL(a)
FanvL(a)
FvMMvL(a)
FvavL(a)
FanvL(a)
FvMMvL(a)

Instrument

FvavL(a)

O 0 0
Strong
Strong
Strong
Place
Place
Place

Instrument . .

FanvL(a)

Instrument

F3M11L o

FiMMaL Plural Class .

FuMMaL Plural Class

3.2.1.20.1
3.2.1.20.2

i i

F M L

3.2 1.21.1
3. 2 1. 21. 2
3. 2 1.

FvMiiL

F88M1L o o
Plural Class .

FMva Weak

?aFMaL Weak .
21. 3 ?aFMaL Color

79

80
80
83
83

83
8h
8h
8h
85
85 .
85
86
86
86

86
87

87

30201022

30201023

3.2.1.2“

3.201025
3.201026
3.2.1.27

30201028

30201.29

30201.30

30201.31

V111

FuMuL Plural Class . . .
3. 2. 1. 22. 1 ?aFMaL Weak
2.1.22.2 ?aFMaL Color

Plural Class . .
Fva Weak .
Fva(v) Weak
FvMLv Weak .
FVMva Weak
Fva Close .
Fva(v) Close
FvavL Close

1MLaa

O O.
NNNNNNN:
O

HHHHHHH

NHUDONHVDOA>

no a on
no
'QCBU1CNQRJF'

3-
F
3.
3
3
3
3
3
3
Fa

'U Ufidkdhnflkflh)

5-;

La lural Class . .
3. 2. 1. 24. 1 FaaM1L . . .

FaM11L Plural Class . .
FMaaMLa Plural Class 0 o

FMUUL Plural Class 0 o 0
3020102701 FVMVL BOdy
Part 0 o o 0
3020102702 FVML(V) Body
Part 0 o o o
302 1 2703 FVMVL craCks
3 ? 1 27 u FvML(v)
Cracks 0 o o
3.2.1.27.5 FvaL Death
3 2 1.27 6 FvML(v) Death
3o2.1.27.7 FVMVL
Elo ated .
3020102708 FVML V)
Elongated .
3.2.1.27o9 FVMVL Plant

FMuuLa Plural Class .
3.2.1.2801 FVMVL o 0
3020102802 FVML o o 0

FvMLaan Plural Class
3. 2. 1. 29. 1 FvavL .
3. 2. 1. 29. 2 Fva(v)L . .

FVIMaaLaat Plural Class

3.2.1.30.1 FvaL Paired
Body Part .
3020103002 FVMLB Plant

FMaaLaat Plural Class .

96

99
99

99

30201.32

3.2.1.33
30201.34
30201.35

3.2.1.36

30201.37

3.2.1.38

3.2.1.39

3'2010u0

3.2 01 on].

3.2.1.142

ix

99

?aFMaaL Plural Class . .
3020103201 FVMVL ' ' . 100
3.2.1.32.2 FvML . . . . 101
3020103203 FVMV o o 0 ' 102
302010 32.“ FVVL o o O 0 102
3.2.1. 32.5 FvavL . . . 102
FaMaaLL Plural Class . . 102
FaMaLaat Plural Class . 102
F1MMaaL Plural Class . . 103
3020103501 FaaMlL o o o 103
FuMMaaL Plural Class . . 103
3020103601 F88M1L o O O 10“
FuMaLaa? Plural Class . 104
3020103701 F8M11L Agent 105
3920103702 FBM11L StronglOS
3020103703 F3M11L Weak 106
FuMaat Plural Class . . 106
3020103801 FaaMi o o o 107
FaMaaMiLa Plural Class . 107
3.2.1.39.l FvMMva(v)

Agent 0 o o 107
3020103902 FVMMVLV

Agent 0 o o 10?
3020103903 FVMMVVL(V)

Strong . . . 107
3.2.1.39.“ FvMMva

Strong . . . 108
FaMaaLi Plural Class . . 108
BozolouOol FVML(V)

Place 0 o o 109
3020104002 FVML(V) Time 109
302010u003 FVMVL Time 0 109
3.2.1.h0.4 FvML(v) Weak 109
?aFM1LarV?aF1MLa Plural
Class 0 o o o o o o 110
Bogolohlol F(V)MVVL A

Cover 0 o o 110
3.2.1.u1.2 F(v)Mva

Statement . 110
3.2.1.hl.3 F(V)MVVL

Instrument . 110

Questionable Plural Class-

88 o

O O O O O O O O O 11

1. MW

 

 

 

 

3. h. 2. 1 Multiple Plural
Association . .

The Change in Semantic Classification
The Newer Model -- A Conflation . .
3. h. l Diachronic Hypothesis .
3. h. 2 Synchronic Evidence . .

3.0.2.2 Borrowed Items . . . .

“Ll A.Classification of the Data . . . .
17.2 The GnOBtOtactics o o o o o o o o 0
“0:3 The “orphOtacticB o o o o o o o

h. 3.1 Semantically-based Plurals .

4. 3.1.1

0.3.1.2

“030103

0.3.1.0

usBoloS

“030106

“030107

ATHEOBETICAL DESCRIPTION . . . o . . . . . .

Round Plurals o o

0. 3.1.1.1 FMaaL-type

4. 3.1.1.2 FvMaaM1(i)L-
type 0 o o 0

”03010103 FVML-type o 0

Square Plurals . . . .
u.3. 1. 2.1 FvMaaMi(i)L-
“03010202 FMaleln-type
n.3.1.2.3 FiMaL-type .

Paired Body Part Plurals

M 3.1.3.1 FvMaaLaat-type

Abstract Plurals . . . .
M 3.1.M 1 FaMaaLL-type
h.3. .2 FaMaLaat-type
.3 ?aFMaaL-type
.lt FVMuuL-type .

a?
u)
Lu4;4:
O O O

o
Mnflkflp \0
o

818 o o . o o o
FaMaLa-type .
FvML-type . .
FquLaa7-type
FaMaaLi-type
FVMMaL-type .
FMaaMLa-type
FaMiiL-type o

.E
o oo 01- o
01.0 o o
HHHHHHHH

o
oo o co 0<oo fij

OVU UuflkhUfifianflfiPJ fMP43:
oo o... 0:00

‘<w\JOWA$flJDOFH1

813 o o o 0 0
FVML-type - .

pa

0
O
mcumwp Hg

° 0
\J\»UMJ\0H \flh' umwkoun»

o

FJH

O

urals . . . . -
FquLaa7-tYP0

FvHaaL-type °

P

d‘
O

O
HHHHH:

esteem to estates:

0

o o

o co ocvm

~QKTQ‘Q\J
o

N

FaMaaMilL-type

7aFM1Laa7-type
FVMuuL-type . '
FaMaaMVL~tYP3 -

O O O Q

111
113
113
115

115
120

12“

125
130
103
150
15h
15h

167
172

176

176
177
178

179
180

183
183
18h
186
187

189
189
190
193
19“

198

199
200

201
201

202
202
203
205
206
207

 

0.3.2

“030109

0.3.1.10

0.3.1.11

0.3.1.12

0.3.1.13

9.3.1.10

Agent Plurals . . . . . . 209
0.3.1.8.1 FVMMaaL-type o 209
4.3.1.8.2 FuMaat-type . 210
0.3.1.8.3 FuMaLaa7-type 211
0.3.1.8.0 FaMaaMiLa-type 212
”03010805 FMQSMilL-type 21“
Instrument Plurals . . . 21”
0.3.1.9.1 FaMaaMiL-type 215
“03010902 FM‘QMiL-type o 215
“03010903 7‘FM1La-type o 215
Location Plurals . . . . 216
h.3.1.10.1 FaMaaL-type . 216
“.18 Plurflls o o o o o o 217
“03.1.1101 FuML-tpr o e 218
h.3.l.ll.2 FMuuLa-type . 219
”03.1.1103 FlMaaMllL-

type . . . . 220

Female Plurals . . . . . 221
u.3.1.12.1 FMuuL-type 221

Plant Plurals e o e o o o 222

“03.1.1301 FVMaaL‘at“
type . . . . 222

h.3.1.13.2 FMuuL-type . 223

Semantically-based Plurals ~—
Integrated o e e o o o o 225

Phonologically-based Plurals . . . 238

“0302.1

0.3.2.2

0.3.2.3

1"03 Summary

The PhonOtactics e o o e o e e e o
Intervening Realizational Phenomena
The Phonetic Tactics . . . . . . .
Implications for the Metatheory . .

x1

Plurals for the P-shape
FVMMVL e 0
”0302.101

0 o e o e o o 241
FaMaaMiL and
FaMaaMiiL typethl
Plurals for the P-shape

FVML o o o o o o o o e o 2&3

“03020201 FVME ~type o o 2&3
“03020202 ?&FMaaL-type o 24“

Other Phonologically-based
Plur‘ls o o e e o o o o o 2 5

e e e o 0 2H5

zus
2 5
220
261

EEGAPH}

MI (

 

V. RESIDUE o o o e e e o ., e e o e o o o o e e o o 266
5.1 Theory Of Exceptions e e o o e o e e e e 266
5. 2 A Stratificational Treatment 0 o e o e o 267

VI. CONCLUSION e o o o o o o e o e o e e o o e e e e 283
6.1 UnreSOIVOd PrOblenS o o e o e e e o 283
6. 2 Suggestions for Further Research . . . . 285

BIBLIOGRAPHY O 0 0 O O O O O O O O O O O O O O O O O 286

APPENDIX: Questionable Plural Classes . . . . . . . 291

xii

 

1. Print
Strat

1. Coupe
Cons:

II. Comm
he?

Table
I .

II.

III .

LI ST OF TABLES

Primitive Components and Complex Relationships in
Stratificational Notation e e e o o o e o e o e o 23

Componential Analysis of Modern Iraqi Arabic
Consonantal Phonemes . . . . . . . . . . . . . . 1&8

Componential Analysis of Modern Iraqi Arabic
VOWBIPhonemes.o......oo.......’+9

xiii

 

he

1. Non-h
2. flier:
3' lam
‘. 33:11

5. Book
to P

 

 

LIST OF FIGURES

Figure
1. Non-hierarchical Semantic Feature Tree . . . . . .
2. Hierarchical Semantic Feature Tree . . . . . . . .
3. Masculine/Feminine Dichotomy . . . . . . . .'-. . .
I a. Hjelmslev's'Model orLanguage . . . . . . . . . .
5. Hockett's Model of the Relationship of Morphemes
toPhonemes ...................
6. A Stratificational Restatement of Hockett's Model
7. The Present Stratificational Model of Language . .
i 8. Relationships of Composition and Neutralization .
9.NodewithExtraLines ..............
i 10' IkOSml's Hypersememic Diagram . . . . . . . . . .
‘ 11. Integrated Structure of the Gnostemic and Sememic
‘ Strata......................
T 12- Non-hierarchical Semantic Categories in Iraqi
‘ Arabic......................
L} 13' Broken/Sound Morphotactic Dichotomy . . . . . . .
1h. Semantically/Phonologically-based Morphotactic
Dichotcmy
15- Taxonomic Hierarchy of Features Defining
Nominalia ................oo-°°
16oI'thensionofF'igurels............oo
17' Cultural Perceptions in the Gnostotactics of

ModernIrainz-abic. . . . .. . '

xiv

lb

16
18
19
28
29
31+

35

114
131

133

138

' 1&0

1’42

 

. me Sep:

lorphen

1. hefixe
1. 3mm

2. Neutra?

2. item
1% Pump
'13. ML
1% Subeh

18.

19.
20.
21.
22.
23.
2i».
25.

27.
28.
29.
3o.
31.

32.
33.
31+.
35.
36.
37.
38.
39.
no.
Ln.
1.2.

“3.

The Separation of Lexical and Grammatical
Morphemes in the Morphotactics . . . . . . . . .

Prefixes and Suffixes . . . . .3. . . . . . . .
Simulfixes...................
Neutralization......»...........
Alternation ..................
Preemption (Simplified Version) . . . . . . . .1
FNaaL Plurals (Semantically-based) Round . . . .
Subclasses of FMaaL-type Plurals . . . . . . . .

FvMaaMi ( i )L Plurals (Semantically-based) Round

FvML Plurals (Semantically-based) Round . . . .
FvMaaMi(1)L Plurals (Semntically-based) Square
FMaaLiin Plural (Semantically-based) Square . .
FiMaL Plural (Semantically-based) Square . . . .

FMaaLaat and F'VMaaLaat Plurals (Semantically- _
based)P81red BOdy Parts 0 o e e e e e e e e e. 0

Alternate Diagram to Figure 310 e e e e e e e e

FaHaaLL Plural (Semantically-based) Abstract .

FaMaLaat Plural (Semantically-based) Abstract
?aFMaaL Plural (Semantically-based) Abstract - -
FVHuuL Plurals (Semantically-based) Abstract - °
An Example of a Noun in the Derivational system

FBMBLaht) Plurals (Semantically-based) Weak . -

FVML Plurals (Semantically-based) Weak . -

FquIaa? Plurals (Semantically-based) Weak

FaMaaLi Plural (Semantically-based) Weak -

FVMMaL Plurals (Semantically-based) Weak .
. MN
Phonotactic Realization of the Morphon /*/ - '

IV

1148
151
151
152
153
154
158
161+
167
171
176
177
179

181
182
183
185
186
187
188
189
191
193
191+
196
197

 

I. mafia
i. Mill. 1
t. Mum
5. ML P1
1?. Em:
~‘. WHaaL
I‘I. MHIU
7. mm
Saw
3 We;
4; him
if. Sum
35‘ Mae
:7 inn
3 Mae
33' ML
‘1' Finn
‘11 Ma

'1‘ PM

 

45.
as.
#7.
u&
h9.
so.
51.
52.

5M
55.
56.
5%
58.
59.
60.
61.
62 .
63.
6h.
65.
66.
67.
68.
68a.
69.
70.

FMaaMLa Plural (Semantically-based) Weak . .
FaMiiL Plural (Semantically-based) Weak . .
FaMaaMiiL Plural (Semantically-based) Weak .
FvML Plurals (Semantically-based) Color . .
FquLaa? Plurals (Semantically-based) Strong
FVMaaL Plurals (Semantically-based) Strong .

7aFMiLaa? Plurals (semantically-based) Strong

FVMuuL Plurals (Semantically-based) Strong -
FaMaaMVL Plurals (Semantically-based) Strong
FVMMaaL Plurals (Semantically-based) Agent .
FuMaat Plural (Semantically-based) Agent . .
FuMaLaa? Plural (Semantically-based) Agent .
FaMaaMiLa Plural (Semantically-based) Agent

7aFMiLa Plural (Semantically-based) Instrument

FaMaaLi Plurals (Semantically-based) Location

FuML Plurals (Semantically-based) Male . .,.
FMuuLa Plural (Semantically-based) Male . .
FiMaaMiiL plural (Semantically-based) Male .
FMuuL Plural (Semantically-based) Female . .
FVMaaLaat Plural (Semantically-based) Plant
FMuuL Plural (Semantically-based) Plant . .
3-based Plurals Integrated . . . . . . . . -
Prefixes . . . . . . . . . . . . . . . . . -
“Orphotactic Classes of Singulars (a) - . -
MOrPhotactic Classes of Singulars (b) . . 0
Morphotactlc Classes of Singulars (c) - - '
C”responding Plural Classes (8) - . - ' ' ‘
°°rrespond1ng Plural Classes (b) - - . -

xvi

198
199
200
201
203
204
205
206
208
210
211
212
213
215
216
218

220
221
223
224
226
227
228
229
230
231
232

 

1. Mean
2. Mo: 5
:. Sum
3. 6mm

"1. Fellini

 

71.
72.
73.
71..
75.
76.
77.
78.
79.
80.
81.
82.
83.
84.
85.
86.

87.
88.
89.
90.
91.
92.
93.
94.
95.
96.

Adjectives e e e e e 0

Major Sound Plurals .
Suffixes e e e e e e e
Genitive Suffixes . .

Feminine Suffix . . .

FaMaaMiL Plural (Phonologically-based)

FVMMVL

O

FvMaL Plural (Phonologically-based) FvML . . .

7aFMaaL Plural (Phonologically-based) FvML

The Phonotactics . . .

Liquids and Slides . .

Vowels (Preemptive. Long and

Fricatlves e o e e e e e e

StOpSQeeoeeeo

"838.180.000.00

Phonotactic Realization of

and Preemptive Vowels

the Morphon

Alternate Realizations of Glottal Stop

Alternate Realizations of V

Possible Alternate Analysis of

Realization of Glottal Stop 38

Exceptional Plural 7asatida

Exceptional Plural W

Exceptional Plural bduuraat

 

Exceptional Plural w .

Exceptional Plural M};

Exceptional Plural safafii:

 

xvii

V
2.

MN/J.

Alternation and Neutralization of the “Normal"

233
231+
235
236
237
242
243
244
247
248
21.9
250
251
252
254

256
257
258
259
260
269
270
271
272
273
271+

 

'. Excepti
i. Excepti
E. ixcepti
15!. Except
L'l. inept
EL Except

171. Excepi

97. Exceptional Plural giniin . .
98. Exceptional Plural 'aadaa? .
99. Exceptional Plural ’imyiin .
100. Exceptional Plural gissiisiin
101. Exceptional Plural gaXaamra .
102. Exceptional Plural ganawgat .
103. Exceptional Plural niswaan A.

will

 

 

 

 

 

LIS‘

C1

Sp
Ns

Lb
r ..
I Hz
4 Fr

Ph
7 Cl
7 V0
Vd

Ln

closed
spirant
nasal
vibrant
labial
apical
retracted
frontal
pharyngeal
glottal
vocalic
voiced
unspecified

lengthened

xix

LIST OF SYMBOLS AND ABBREVIATIONS

he :3 t3 CD 0

SN
LN
MN

PN

Sem

Phon

 

gnosteme
sememe
lexeme
morpheme
phoneme
gnoston
semon
lexon
morphOn

phonon

Defining Semantic
Characteristic(s)

Phonological (Surface
Canoni cal) Form( 8)

 

 

4111

CHAPTER I

Introduction

1.1 The Topic

Broken plurals have posed a problem for analysts
for centuries. The term 'broken plural' is used in most
books on Arabic grammar to designate an internal plural.
For example, the plural of fundug 'hotel' is W:
the plural of guys; 'moon' is 3111812223 the plural of ra_'§_d_
'thunder' is :u'uud. Such plurals are in contrast with
the so-called 'sound' plurals. realized as inflectional
suffixes. For example, the plural of gattaan 'tattIe-tale'
is gattaaniin; the plural of malika 'queen' is malikaat.

 

A parallel situation exists in English where the
regular plural is realized as a suffix (frequently repre-
sented as -Z). There are instances of internal vowel
alternation, however. such as m. 1111931 16111. $9.211: etc.
It Just happens that this type of pluralization occurs
much more frequently in Arabic.

Traditional descriptions of Arabic contain many
statements such as:

No definite rule can be given for the formation of

the plural. The plurals of nouns can only be

learned individually, a task which is not so for-
midable as would at first appear.1

Structural descriptions do not offer much more insight.

I

 

in we f1]
:1 noun 1
“he are
his. (
shin. .
A no
11th: nou
hum, (
:ihlm
filming

a)?
1

7’4 sy;

”tutu

Eu

2

Again we find statements like "In general the plural of
each noun must be learned with the singular.”2 And,
”There are some thirty possible patterns for broken
plurals. Only a few of them are predictable from the
singular...."3

A recent article by Talmy G'ivénu on the development
of the noun class system of the modern Bantu languages.
however. offers some interesting implications for the
pluralization system of the Semitic languages. The

following is Givc'm's hypothesis:

a) The older classification of (proto) Bantu

was an n-ary. non-hierarchi cal. non-anthropo-
gentric multi-gender system of features;

there was no furthe' elaboration of the feature

[animate] to yield human] . There was a one-

to-one correspondence between the semantic

and morphological noun classification. (1). 41)
This system may be illustrated graphically by means of

a feature-tree:

C noun]

[animate] [inanimate] [abstract] [paired bOdy parts] , etc.

Figure l

Non-Hierarchical Semantic Feature Tree

3

b) A re-analysis of the semantics underlying
the noun universe then took place. The source
of this re-analysis is not clear. This re-
analysis centered around the gubgect specificity
of verbs or, in other words, what noun
perform what acts. Specifically. the position
of humans within this schema was re-analyzed;
the feature [human] was given an added spec-
ification of marking above and beyond anima 2.
Thus a hierarchize , bi-gary. anthropocentric
system of semantic classification of nouns
emerged. in which the category/feature [human]
was placed at the top of the hierarchy of
semantic markedness. (pp. 41-2)

We may illustrate this system graphically as:

[noun ]
[concrete] (abstract)
[animate] (inanimate)
[human] (non-human)
Figure 2

Hierarchical Semantic Feature Tree

The features at the left are the marked of the pair,
While those at the right are the unmarked; with [human]
being the most marked.
0) Largely in order to accommodate this new

emantic classification the Bantu noun class

172 was then created -- de novo -- and

human nouns from class 9/10 were slowly

moved into it. (p. 42)

In Arabic we see a parallel situation. The Arabic

ISO-called 'sound' plurals (i.e.. regular. unbroken, or

 

 

 

 

 

 

flied]. -
“it, r<
m “MEI
none of
82} the
ategory -

irked}.
[h
[MIMI

Smn'

 

l4.
affixed). -_i_i_p_ for the grammatically masculine nouns
and -a_a_t for the grammatically feminine nouns. seem to.
give evidence of 1) a newer plural class created for the
purpose of accommodating a new semantic classification
and 2) the creation of a still more marked grammatical
category -- that of feminine (the masculine being un-

marked). Semantically both are marked.

[h n] [human]
[feminine] [masculine] [feminine] (masculine)

Semantic Gender Grammatical Gender
_ Figure 3
Mascul ine/Feminine Di chotomy

Justification for considering the feminine to be the
marked of the two is found in the incorporation of bor-
rowed words which are already semantically ”marked” as
non-native. Borrowed words which have not been com-
pletely assimilated (i.e., do not receive broken plurals
on the basis of semantic features and/or canonical form)
always receive the marked (or feminine) plural of the
unmarked 'sound' plural category. That is. borrowed items
fall into the unmarked or productive 'sound' plural cat-
esory, but take the marked choice cf the two plurals in

this category -- the feminine.

 

 

 

 

 

 

d] Sin

5

d) Since the old. multi-gender morphological
classification of Bantu nouns had ceased to
reflect the semantic facts. there was no
compelling reason left for nouns to remain in
any specific noun class -- except class 1/2.
That is the change over from one prefix-
agreement class to another would not have
disrupted the semantic classification of
the noun universe as it would have done
prior to the semantic re-analysis. The
moving of a noun stem from one agreement class
to the other. done by slip of tongue, memory
impediments, imperfect learningaof children acquiring
the language, would have presented no im-
pediment to the acquisition or maintenance
of the semantic system.underlying the noun
universe. Thus, non-human Bantu nouns have
begun to slowly 'migrate' all over the noun
class system. giving rise to the present day
chaotic and largely language specific dis-
tribution. (p.42)

If this analysis were hypothesized for the Semitic
languages as well. it would account for the apparently
chaotic situation of the current pluralization system.

Some singular items in Arabic have two or more plurals

in ”free” variation5 -- although one of these plurals

 

usually occurs much more frequently than the other( s).
It is Givén's contention that the morphology of
plural formation in Semitic reveals an underlying M-
aegis; system upon which the more recent morphological/
Bareement genders seem to have been superimposed. Plur-
alization in the older system is not achieved by suffixe-
tion. which is more recent. but involves the interdigita-

tion of plural morphemes with the CCC6 censonantal system.

 

This older system of pluralization-borders on derivation
but is associated with the inflectional phenomenon of
Pluralization.

 

 

 

This di
edits pot!

mm 5‘

6
This dissertation investigates Givén's hypothesis
and its potential as an explanation for the Iraqi plur-

alizati on system.

1.2 Previous Scholapship

The previous scholarship on Arabic broken plurals
is quite limited. There has been only one dissertation
using a modern linguistic theory, namely Levy (1971).
In her transformational-generative treatment of Modern
Standard Arabic plurals Levy divides the singular items
which take broken plurals into taxonomic classes on the

bases of the features rational? masculine/femipine. and

 

of their phonological shape (i.e., canonical form). The
most frequently occurring plural for a given class is
predicted by a major plural rule. The less frequently
occurring are predicted by minor rules, and exceptional
cases are predicted by exception features. The "elsewhere"
plurals (i.e., the sound plurals) are ordered after the
broken plural rules and apply to any item which has failed
to meet the structural description of broken plural rules
(either inherently or via exception features). Levy
admits, however. that though her rules predict the correct
Plura1(s) for any given item, they do not explain the
Peculiarities of the system. She suggests that the theory
of markedncss may ultimately offer this explanation.

A second recent transformational description is

found in Bram (1970). In his treatment of Arabic phonology,

 

he deals &‘
maiden m
tents of
new 01
:c the nu
0the
ml exp]
ieacript‘.
M.) 5‘
the 1m
Weir
Late thq
m on
manic
T?
m
Mint

Elisa;

Fa -:-'t'" '-

7
he deals almost exclusively with verb morphology. He

provides many interesting observations on the phonological
aspects of Arabic. but does not deal with the overriding
concern of this thesis -- the relationship of semantics

to the noun morphology.

Other publications of interest. but which offer no
real explanation are: Erwin (1963) -- a structural
descriptions Fleisch (1961). Kuronwicz (1961), Lekiaschwili
(n.d.) and Murtonen (1961;) -- traditional descriptions.
the last three being diachronic. Despite their diachronic
viewpoint, however. these works offer no real insights
into the semantic history of the pluralization system, but
deal only with aspects of its fluctuations according to
canonical form.

There have been no works on Arabic within the
stratificational framework and no published work using
semantically (or gnostemically) defined morphological
classes. I

The present study attempts to shed some light on
the present-day pluralization system of Iraqi Arabic (and
Perhaps indirectly on the system of Arabic pluralization
for the dialects in general). Specifically. it offers
some explanation for the apparent peculiarities of the
System (over and above predictions). At the same time
work in this area makes some contribution to the meta-
theory, inasmuch as a study of this nature, which indicates

a rather direct relationship between semantics and

 

 

 

nhology‘
hi ha
ha and
sites the
unity t1
motion
aha: ther

wing .

1.3 M

Hod
to the F
the 1
item
hast

Etneous

8

morph010gy, has never been attempted within the stratifica-
tional framework. The direct relationship between morph-
ology and meaning was hinted at in Lamb (1971a) where he
states that "morphotaotic classes which are needed to.
specify the occurrence of morphemes in derivational con-
structions are often semantically defined.... It suggests
that there must somehow be direct links from morphemes to

meaning . . . . "

1.3 Modern Iraqi Arabic

Modern Iraqi Arabic is a term commonly used to refer
to the‘huslim Baghdadi dialect of Iraqi Arabic. Iraqi
Arabic belongs to the group of Arabic dialects called
Eastern Arabic. While members of this group share certain
characteristic features. the group as a whole is not homo-
geneous.

The Muslim Baghdadi dialect is a nomadic or Bedouin
dialect. sometimes referred to as M ('1- said') to
distinguish it from the Christian and Jewish dialects
Of the same city, which are sedentary, and which use the
form gala, ('I said') instead of the form given above.

As the city of Baghdad is) divided. so is the rest of Iraq.
The Eel-WM dialect split of the country is not based
on religious affiliation. however. but rather on geography.
The cities to the north. like Mosul. are in the 921.33;
dialect group, while the cities to the south. like Basra.

as well as the outlying farm districts. share the nomadic

 

 

 

features of
The te
Ms usual

the.

.11 a
The i
he. L1
lieeiheed

e5 singu]
the eeee]

the:

 

 

 

 

 

9
features of the M dialect group.
The term Modern Iraqi Arabic as used herein refers.
as is usual, to the Muslim Baghdadi Iraqi dialect of
Arabic.

1.1; p_a_t_a

The source of data used in the research is threefold.
First. WWW
(Woodhead and Beene: 1967) was used to compile lists
of singulars and plurals and their meanings.8 Secondly.
the acceptability of these plurals was) checked with two
Baghdadi Iraqi informants -- one of whom has been in the
United States for some time and one of whom has recently
returned from Baghdad. As a check on plurals in context,
Spoken Arabic of Baghdad: . An Anthology of Texts (McCarthy
and Baffouli: 1969) was used.

The informants were also consulted as to their in-
tuitions about semantic classifications.

Transliteration of the data was not necessary since
most dialectal studies are written in the Roman alphabet
in order to make the dialectal variations more evident.
It was only necessary to coordinate the various methods
of transcription. (See Section 2.1.9.2 for a chart of

Modern Iraqi Arabic consonantal and vocalic phonemes.)

 

 

.e John Vin
in helleee
‘je hen Cr
,, Ellen G
he one
he of
hypothe

hleolh
henhh 1

old u

deter

 

. ~M“- "

1.
2.

3.

5.

FOOTNOTES

Chapter I

John Van E38 (1938).
Wallace M. Erwin (1963).
Mary Catherine Bateson (1967).

Talmy Givén (1971). Of course. Givén's hypothesis is

but one of several interpretations possible. The out-

come of any research is always dependent upon the initial
hypothesis/hypotheses one chooses to work with. (See
Malcolm Guthrie, "Variation in the Range of Classes in the
Bantu Languages" in La Classification Nominale dans les
Langges Négro—Afrioaines, pp. 3 1-53. Paris. Center atlon-
a1 de la Recherche Scientifique. 1967 for an alternate

hypothesis in which classes 1/2 are posited as being as
old. as some of the non-anthropocentric classes.

No variation is absolutely "free". Context (i.e.. social
situation, style. etc.) Usually plays a crucial role in
determining which variants are to occur.

The triconsonantal pattern is the most common. However.
there are also bi- and quadriconsonantal patterns.

The feature [rational] may be defined as "of or pertaining
to those creatures in the animalia hierarchy able to
reason. i.e.. humans.”

Regarding the question of the relationship of the
collective noun to its corresponding noun of unity. it

was decided that since this relationship was derivational.
collectives would not be considered in the description of
inflectional pluralization -- either sound or broken ..
even though broken pluraliZation borders on being deriva-
tional in nature. For example, ‘

collective tuffaah app1e(8) (in general)
noun of unity tuffaaha an apple
plural tuffaahaat apples (more than two)

while tuffaah in a certain sense may be considered to be

in a type of.plura1 relationship with tuffaaha. this

 

relationship was excluded from discussion in the present
study. Also excluded was the dual.

10

 

 

:‘IilosOphf
79! the c
Etienne.
7‘5: the
liescrij

it it

'\
\m!‘

M

CHAPTER II

Preliminaries

2.1 Theoretical Approach to the Togic
2 a 1 e 1 BECksround

The theory to be used as a framework for the descrip-

 

tion is that of stratificational grammar. It was shown
by William J. Sullivan (1969) in his recent dissertation that
stratificational theory fulfilled all the practical and
philosOphical criteria he set up as necessary prerequisites
for the consistency of a theory with the philosophy of
science. He states that the phiIOSOphy of science ”demands
that the theory be oriented toward seeking in analysis
a description of the system underlying the raw data. and
that it should provide a mechanism for describing that
system." Further. ”since the system sought is the set
of interrelationships of the items of the analysis, the
description must consist of a network of relations.“ (p. 75)
Stratificational theory surely fulfills this last
requirement since the major tenet of stratificational
theory is that language is a network of relationships.
Units. if they can be said toexist within such a
framework, can only be fully defined via a definition of
the entire system of which they are a part. Within a
stratificational description. however. units exist only
outside of language prOper (i.e.. in meaning and sound).

11

 

 

 

 

hazel 1'
hating so
h order he
Math}
L hits“
homes.
mh Lem
My in
th that
5‘13 ranch
756! are
hum:
It
entire E
ipoint
2thch
In

the 3p.
‘53 mu;
{how

I

hhm

MEL

12

Language-1 itself consists solely of the relationships
relating sound to meaning and tautologically vice versa.
In order to discuss certain pertinent points in the network
of relationships it becomes convenient to speak of them
as ”units" and sometimes to give them names. e.g.. morphemes.
phonemes. etc. It must be understood, however. that any
such reference to ”units” in the description which follows is
solely in the sense ofpoints in the network of relationships.
and that the terms morphemes. morphemic signs. etc. are
only mnemonic devices for referring to these points.
They are purely for the convenience of the-analyst in
describing and the reader in comprehending.

It is humanly impossible to actually discuss the
entire system of relationships each time one refers to
a point in such a system. Thus such labels become a
practical necessity.

In establishing stratificational grammar within
the spectrum of twentieth century linguistic theories.
one must look to the two most important precursors to the
theory -- Louis Hjelmslev and Charles Hockett.2

In his Prolegomena to a Theory of Language. HJelmslev
introduced the notion of language as 2.92;“.- intervening
between the content continuum at one end and the expression
continuum at the other. These continuums he labeled
22219213, (i.e.. content purport and expression purport).

The content- and expression-form are independent of. and

 

 

 

 

 

 

{m in art
him: then
hLLtratioL
in). cast
TL patterr
attain ha:
hzht offs
Fine h t

1‘! down

13
stand in arbitrary relation to, the respective purports.
forming them into content- and eXpression-substance. For
illustration, HJelmslev used the example of a net (the
form). casting a shadow on an amorphous mass (the purport).
The pattern outlined by the shadow divides this mass in a
certain way creating the substance. A graphic display
might offer better explanation. (In the top half of
Figure ll. the rays are projecting upward and in the bottom

half downward. )

 

 

 

 

 

L h
uh

hitch
1h

 

the that
{L Saussure
hm.

The n
Lat the r
in the L

Ihe L
h: the n
huge

the com

15

Notice that the dichotomy is the same as that posited by
de Saussure (1959) in his two sides of the linguistic
sign.

The major difference between Hjelmslev and Lamb is
that the relationships Lamb posits are more primitive
than the ones Hjelmslev posits.

The second precursor, Hockett, laid the groundwork
for the notion of intervening levels or strata.3 That is,
language cannot be viewed as a simple dichotomy of form
into content and expression. since there exists overwhelming
Justification for the existence of at least one stratum
relating morphemes to phonemes. Subsequent studies have
shown that similarly morphemes, in general, are not directly
related to meaning, thus establishing form as having
several levels of structure (1.9... strata).

In his article, ”Linguistic Elements and Their Relations".
Hockett considered the following relationship between

morphemes and phonemes:

 

1) pm

he pseu
than}
hose 0‘
)7 is c
”h his
)0 sh

C

he LL

 

16

{} MORPHEMES
R

C ——>pseudo-phonemes P//

 

Pi; pseudo-morphemes
R

PHONEMES //

Figure 5
Hockett's Model of the Relationship

of Morphemes to Phonemes

The pseudo-morphemes may be rte-termed allomorphs and the
PSEUdo-phonemes morphophonemes. The relationships are
those of representation (R) and coupesition (C). Althousfih
1’0 is clear that a level exists between classical morphemes
and classical phonemes, Hockett gave this level no linguis-
tic status.

One familiar with stratificational theory can easily
see the development from this to the morphemic level in
the stratificational framework. "Units" the size of the
traditional morpheme were re-termed lexons (i.e.. the
components of lexemes). "Units" the size of traditional

alllo‘nm‘Phs were re-termed morphemes (1.13.. the t3°t1°al

units of the morphemic stratum). and "units" the size of

the morPhOphoneme were re-termed morphons (i.e.. the

components of morphemes). The phoneme became the stratifica-

tion“- phoneme which was the same size level as the traditional

 

 

hen but
L: L bhni
mar are
tactics L‘
Lama} sy
LL htrod
ih'phEheB.
level hen
Ehhehe,

hit

hm:

than

l7
phoneme but at a higher level of abstraction (i.e.. it was
not a biunique level). The relationships in stratificational
grammar are those of realization and composition. Since
a tactics defines a stratal system, or, in other words. a
stratal system cannot exist without a tactics, a tactics
was introduced to generate the well-formed combinations of
morphemes. The stratum so defined formed an intervening
level between the classical morpheme and the classical
Phoneme, thus solving the anomaly emerging from Bloomfield

that
A morpheme is composed of phonemes.
and at the same time.

A morpheme has allomorphs.

The stratificational restatement of Hockett's model is

Riven below .

FM

 

 

18
‘ MM HOCKEIT ‘3
TERMS
LEKONS
MORiTH EME‘S
E.
\ y
4-- Mer‘vkemes Pseuio‘ — mo emes
e
A90 c
5635/ Pseu (1.0" PM “W5-
50 MORPHONS R,
- Fiuaems PHON Ems-s
‘WKJII;
PURRES.

 

 

Figure 6‘_

A Stratificational Restatement
of Hockett's Model
This was only the beginning, however. Since Hockett's
article appeared, linguists working within the stratifica-
tional framework have found uple evidence to warrant
‘the hypothesis of still further levels of structure, each
level being structured in a similar manner. The present

levels Posited for language are listed below.

 

19

Gnostemic

 

 

Sememic
Lexemic . GRAMMAR

LANGUAGE Morphemic

 

 

 

L Phonemi c PHONOLOGY

 

 

Phonetic

Figure 7
The Present Stratificational Model of Language

This shows a neat division between grammar and. phonology
butcnm_must not forget that the system is an interwoven
whole. The gnostemic and phonetic strata are presently

thought to be, if not outside language entirely, at 133315

at the periphery .

2.1.2 Axioms and Primitives

There are certain.primitives out of which a stratifica-
tional description is constructed and certain axioms upon
Which the theory is based.

The first of these axioms is

A.1 All human knowledge and thus all human activities
including language may be described by means Of a
few logical relationships. It is the ability to
structure his universe by means of these relationships

that is innate in man.

 

ehjehs?

Line
to r
in h
and
eoch
hoe

these 10

hi “

 

 

20

As Hjelmslev (19611 p.127) has stated:

Linguistic theory is led by an inner necessity
to recognize not merely the linguistic system,
in its schema and in its usage, in its totality
and in its individuality, but also man and human
society behind language, and all man's sphere of
knowledge through language.

These logical relationships are two of the primitives

of the linguistic system. ‘ They are "and" and "or."

The second axiom is

A.2 Language, as a mode of communication, functions
in two directions. During encoding it functions

in the direction from meaning toward sound. During
decoding it functions in the opposite direction
(i.e.. from sound to meaning). Therefore, a theory
should not impose a partiality, either intended or

1mplied, for either direction.

Stratificational grammar accepts this axiom by its introduc-
tion of the two primitives of directionality "upward"
and. "dQWnWard."

The third axiom is

A03 Ordering is often crucial within the linguistic

System.

There are many cases of crucial syntagmatic ordering.

In En'Ehlish the subject-verb-object word order is a good

GXamp1e, ' Stratificational grammar accepts this third

 

 

he by
'hordere
LL. the
he nee]
hence L
h’ee pre
in
hotel 1
While
the 1m
threet
h. hhlc
h“ Leer;

Ll

hen

21

axiom by introducing 1) the primitives "ordered” and
I ”unordered” and 2) the notion of stratal precedence,
i.e., the notion that the lower stratal tactics (the
ones nearer sound) take precedence over the higher stratal
tactics during encoding and the higher stratal tactics
take precedence over the lower during decoding.

An additional type of ordering extant in stratifica-
z.

tional theory, termed tactical ordering , characterizes
variable ordering, not dependent on any ordering within
the linguistic system. This type of ordering rather

characterizes the realization of elements in the order
in which they occur, e.g., the order of the realization

of morphons composing any given morpheme.

Lamb (1972: pp.675-6) gives the example from Monachi:

Gloss: to haul water bucket

a) Morphonic: pa noo 78' na P8 noo 'nu hE
Phonemi c : panoo 2a ' napanoo ' noho
Gloss: gold specifically gold

‘0) Morphonic: ?oono ?E ?oono 7E 'su
Phonemic: ?oono?o 70011070 ' 8°

Given the following two realization formulae:

R1 E/VlK ] V1
32 u/oK J 0

Where K =2 any consonant or consonant cluster. "9 can account

for the apparent need for reversed ordering ‘70 handle examples

Far

 

LL and b)
h emp‘
hLthch

hreelie
h exam}

Lie reef

2.1.3 h
The
hLel he
hated L
heneee
:heee '
[of one
)TPBSL

5’1! or

-‘-—| ”PM " fl -

22
a) and b) above with stratificational tactical ordering.
For example, in a) the morphons are realized in the order
in which they occur in the morpheme. In 'nu hE, first 3;
is realized as p, then 3; is realized as p yielding 29113.
In example b) (?E 'su) first I}; is realized as p, then

2 is realized as p yielding ?o'so.

2.1.3 Types of Relationships
The Justification for establishing an intervening

level between two linguistic "units" such as that illus-
trated above between the classical morpheme and the classical
phoneme lies with the fact that the relationship between
these "units” is not a simple relationship (i.e., it is
not one-to-one). Relationships may be divided into two
types: simple (or one-to-one) and complex (i.e., many-to-
one or one-to-many) .

The types of complex relationships existing within
the linguistic system may be shown to be composed of-the
primitives of the system of stratificational grammar (i.e.,
£112. 93;. u ward, 19.11.439.21: ordered, unordered). Following
18 a table of complex relationships and their primitive

Comp onents .

 

hhhtiw

lhersifiL

Seetreli:

hpeer

hits

23

TABLE I

Primitive Components and Complex Relationships
in Stratificational Notation

COMPLEX RELATIONSHIP PRIMITIVE COMPONENTS STRATIFICATIONAL
_N_________OTATION

Diversification pp, downward, ordered
22: We p___________ncrdered
Neutralization 9_r_, u ward, ordered

pp, u ward, unordered

and, downward, uno dered
Portmanteau and, upwapd, ordered
and, u ward, unord_______e____red

Zero Simple or pp, downward,
ordered or unordered

T‘T
7T
LL
Composite Id downward o dered ‘yl
29" ”’ fi
7%
9
¥
)1

Empty Simple or tactically specified
and, u ward, opdezed or unordered

:25
l

 

21+
The following is an explanation, with examples, of
how to read the stratificational notation outlined in
Table I. _
1) a is realized as b or c, but b is
TT chosen in preference whenever'possible.
b C.
LN/SooJ/

For example, the stratificational morpheme M/b:T/
(traditional allom'orph) is chosen only in the environment
of comparative or superlative. Elsewhere (the unmarked meT/ “96:44]
or rightmost branch) the stratificational morpheme M/gud/

is realized. a
2) 1* | a is realized as b or c (or a is a class
b consisting of the members b and c).
C.

degedfi veg

- For example ,

Pratt). <3 ream.

0. b

3) I l Either a or b is realized as c but a
takes precedence. This occurs rather
C.
u 1 e

a_ b infreq ent y

h) ac Either a or b is realized as c.
. S -

L/bes/
For examp1e, eitherb big (in the sense of elder) or

~18 (1n the sense of large) can be realized as the lexeme

L/big/, causing the ambiguity in a Phrase like 'my big sister '

 

for em

For ex

For L

‘—_ III-m” ‘ -

25
L 0 ’
5) E a is composed of b followed by c.
b. C,
Ms/matm/
For example,
EZn/'/QE/ /$w/'
Cl ,

6) a is composed of b and c simul-
taneously (or without regard to
order).

5 (L
For exam 1e LN * LN
P . /¢meye the lexcn [men/ (actually
man + plural) is com-
posed of the morpheme
M/m3£'.n/ simultaneous
A1 .+ with the plural pre-
/maen/ /E / emptive morpheme M /5+ /.
7) (2 k) a and b occurring simultaneously

are realized as c.

eo" V/S/L/ s
z-example, horse ena.e the sememes /horse/ and

S/male/ occurring simul-
taneously are realized as the

“em” SN/stalli 011/.

sW/SLLu/Izon/
8) C: b a followed by b is realized as c.

M/a/ “/Le/

C. I‘d/zh/ ‘

Ithrench, the morphemes "/3/ followed by M/1e/ are realized

3‘3 the morphon MN/o/.

 

her eh}
hem 5
held!

heath

 

 

26

9) X x is realized as zero. For
I example, one of the realizations
of the plural morpheme in English
ID 13 as

“pl the morphotactics generates a
determined element (”empty morph”)
z in the environment of child and
plupal. This 1 has no connections

M, to higher levels.
[zaylcl/ V‘

10) (9 fl is realized as x. For example,
X

Lamb's Outlipe of Stpapizicapiopg; Grma; and Lockwood's

I roduct on to St at icational L s 10 contain fur-

 

ther elaboration on the notational system of stratifica-
tional grammar. Reich's ”Symbols, Relations, and Structural
Complexity” includes an explanation of their 1081081 1m-

Dlications and dynamic interpretations.

2.1.14 The Simplicity Measupe

One of the best features of stratificational grammar
is its simplicity measure. The same measure can be used
to measure a grammar as a whole or to measure any part of
it: Thus. stratificational grammar can decide if a sim-
P11fication in one part of the grammar has led to an over-
all complicauon or is really a simplification in terms
°f the grammar as a whole.

The need for a simplicity measure stems from the fact

that 81mFelicity and generality go hand in hand, and

 

meefioatic
1e keeping K5
meme to a
We to 0
equally desc
ee.e.. the n

Thus a1

ehe empirice

A desc
(self-
as pos
centre

of ex}
exhaue
the re
Lee (1966‘
“0 aeooun
55011 (1 .e.
We 1y
[Wire WE
ee lines
Elements
Wally

‘3 subs t!

 

2?

stratificational theory wishes to make use of this fact

in keeping with the philosophy of science. Any theory

wishing to achieve explanatory adequacy must

be able to offer a process for deciding which of two

equally descriptively adequate grammars is the simplest

(i.e., the most general) (of. Chomsky, 1965: pp. 30-?)
Thus an explanatorily adequate theory must include

the empirical principle (Hjelmslev: 1961, p. 11 )

A description shall be free of contradiction

(self-consistent), exhaustive. and as simple

as possible. The requirement of freedom from

contradiction takes precedence over the requirement

of exhaustive description. The requirement of

exhaustive description takes precedence over

the requirement of simplicity.
Lamb (1966b) has suggested a measure for deciding between
two accounts which are equivalent in effective informa-
tion (i.e., equally exhaustive) but which differ in
surface information (i.e., differ in simplicity). This
measure was later refined by Reich (1968). It amounts
to what has been termed "counting lines and nodes.” But
one must remember that nodes simply define relationships
and lines the ”elements” taking part in the relationship.
(Elements exist only in substance. Counting 111193 13
“Wally counting the number of times ultimate connections
t30 substantive elements are involved in a particular

relationship.) Thus counting lines and nodes is equivalent

to cduntins +'s and -'s and letters in algebraic descrip-

tions.

 

For exam

ettement the

t) (a
11) a n

he method c
“he member e
member of ti
are stated.
the and s
the state
it) is sim;
Mum
Hhile
We toll
item des
fitted the
he into e

math

men a ‘

Mean

28
For example, it is obvious that ii) is a simpler

statement than 1). yet it imparts the same information.

1) (a-b)+(a-c)+(a-d)
ii)a.(b+c+d)

One method of measuring the simplicity would be to count
the number of times a relationship is stated and the
number of times elements entering into those relationships
are stated. Thus 1) is shown to have five stated relation-
ships and six stated elements. while ii) is shown to have
three stated relationships and four stated elements. So
ii) is simpler or. in linguistic terms, it captures more
general izat i one .
While this example does not mirror exactly the tech-
nique followed in determining the simplicity of a stratifica-
tional description. it serves to eXplain the rationale
behind the simplicity measure. Actually. not every element enter-
ing into each relationship is counted -- only those in
excess of three. For example,
X
ii 5
(a) (b)

Figure 8
Relationships of Composition and Neutralization

given a common relationship such as composition (a) or

neutralization (b), the normal situation is to have

 

 

 

 

hee cone
etelloeeee‘
a. tech
meet me
me et le
mate the
ease (th
is the

extra on

29
three connecting items. That is, in (a) x is composed of
y followed by 2, while in (b) either q or r is realized as
s. A node is defined in terms of its connections upward
toward meaning and downward toward sound -- one incoming
and at least two outgoing. Less than two outgoing would
make the relationship vacuous, i.e., simple. The limiting
case (i.e., the case where three such connections exist)
is thus taken as defining the relationship or ppgg. Only

extra outgoing lines are counted. For example,-

C
e 3 h
Figure 9

Node with Extra Lines

Figure, 9 would be counted as one node (relationship),
defined by e and any two of f. g. h. and one extra line
(one extra "item" entering into the relationship).

Since a total stratificational description is given
in terms of nodes’and lines (i.e., in terms of relationships
and their connections) the same simplicity measure may

be used throughout .

2.1.5 Majo; Goal

The major goal of stratificational theory is a

description of man's "competence to W his language"
(Lamb: 1971a, p.103). That is, the competence and

 

 

 

 

hemmee
mm.
the mot
‘ete eppmoe
:ement mod
teeelete 1
:e all of “
empe than
he been me
teetell g
meet pote
The e
title are
the team
teetehc
‘5 it me:
It I
)3 E mode

heme

30
performance models proposed by the theory are closely
linked. '
- The model may be called cognitive since its aim

. is to approach a working model which is compatible with
current models of the brain and its functions. It attempts
to relate language to other forms of human activity and
to all of "knowledge." Thus the model has an even broader
scope than the description of language. It can be and
has been used to describe other human activities such as
baseball games-Sand dance gesturesé. It obviously possesses
great potential.

The way in which the competence and performance
models are related is simply that the performance model is
the dynamic (or innervated) counterpart of the static
competence model -- orthe competence model ”plugged in“
as it were.

It must be pointed out that this performance model
is a model of ideal performance. That is, it does not
account for the variables of actual performance such as
memory limits. physiological malfunctions, etc.

Stratificational grammar aims at achieving an isomorphic
model. A graphic illustration might be helpful. The
goal of stratificational grammar might be compared to the goal
01‘ setting up a model of the inner workings of the computer
(often called the hardware). Even though one cannot see
the inside of a functioning brain as one can the inside

Of a functioning computer so that the actual isomorphism

 

 

 

the model may be c
[stemmed about ho“

:hmi model will an

2.? 8pm
2.2.1 The Gnostemic
The semology of
imteteo strata -- t
imie system is our]
teem and to contaf
ledge will my free
hetemeehich impose
heart to form dif
imethemeeords. dii
I‘m'r‘ntle. or one
Iieee the world.
The tactics 0
70? the taxonomic
)1“) W161. etc.
Meets" of
.feleI. These: gt
eeheepondence Wi
1.5.. the “lam
the reasons f0!
heed oi Enostom

l
51mph r ela t1 0)

31
of the model may be compared, it is hOped that when more
is learned about how the brain functions, the stratifica-

tional model will approach this isomorphism.

2 .2 Semologx

2.2.1 The Gnostemic and Sememic Strata

The semology of stratificational grammar is divided
into two strata -- the gnostemic and sememic. The gnos-
temic system is currently thought to be outside of language
mower and to contain all of human knowledge. This know-
ledge will vary from culture to culture as it is this
system which imposes differing divisions on the content
purport to form differing divisions of content substance.
In other words. different cultures tend to view the world
differently. or one's culture places a bias on how one
views the world.

The tactics of the gnostemic system, besides providing
for the taxonomic hierarchies of the animal world, the
plant world, etc. provides for well-formed combinations
of the ”units” of meaning called gposgemes, e.g., G/go/,
G/male/. These gnostemes are usually in one-to-one
correspondence with gnostons, the components of gnostemes
(i.e., the relationship‘is usually a simple one). Some-
maJor reasons for positing a structure of gnostemes com...

' Posed of gnostons when the "composition” appears to be

a simple relationship are:

 

 

 

1) There are c331
hpa,e.g.. certain i
hemlly pessivized

manta at this level

la) Muskie
1b) Hi§ ha‘

ha above example is
following which cenh

Whittle interpret;

2!) The Cl
2b) The h

33) The c
3b) Us

1"” 13 also the I

\
reheen “0mm“ !

2) This Struc
lsanalogous t 0 th
W” Pmper.

This second,
llh-llhgmstic :
in“

ell), , mm,
139031

mture , B‘

“d at thi
’il

32

1) There are cases of complex compositional relation-
ships,e.g., certain idiomatic expressions which can be
internally passivized need to be broken down into com-

ponents at this level. An example follows.

1a) Muskie threw his hat in the ring.
lb) His hat was thrown in the ring.

The above example is in contrast to such examples as their
following which cannot be passivized and still retain the
idiomatic interpretations

2a) The cow kicked the bucket. (ambiguous)
2b) The bucket was kicked by the cow. (unambiguous)

3a) The child threw a fit.
3b) A {it was thrown by the child. (unacceptable)

There is also the fact that pormanteau realizations exist

between gnostemes and gnostons.

2) This structure bf gnostemes composed of gnostons
is analogous to that posited for a stratal system of
language proper.

This second reason is not particularly convincing since
extra-linguistic structure need not mirror exactly linguis-
tic structure. But since there exist examples such as that
in 1b), a relationship 01‘ composition between ~emes and -ons
is posited at this level, even though most of the relation...
Ships will indeed be simple.

 

 

 

 

l‘ne aememic strat
this within imam
tlewel are realized
teamemic level. I
w? to separate strat
t necessary to show
ll that complex
mtaememes or. in c
we not one-to-one.
2) that the (101

ll tlcstemic level
llferent from, tha
It is commonl:

he is the propos
“m“ is the tex
Mount for tel

mm“. howevew

hr.

To 1unattat
Elm between the
mtctntlng “I p!

l

 

 

33

The sememic stratum is the first level of the semology
which is within language prOper. The gnostons of the gnostem-
ic level are realized as sememes. the tactic ”units” of
the sememic level. In order to Justify the existence
of two separate strata -- the gnostemic and sememic -- it
is necessary to show

1) that complex relationships exist between gnostemes
end sememes or. in other words, that the relationships
are not one-to-one, and

'2) that the domain of the gnostotactics, which defines
the gnostemic level, is greater than, or, in some sense,
different from, that of the semotactics.

It is commonly held that the domain of the semotac-
tics is the proposition, while the domain of the gnosto-
tactics is the text, or discourse. (Relations necessary
to account for texts are presumed to include all of human
knowledge, however.) Thus part 2) [above is accounted
for.

To illustrate the fact that complex relationships
OXist between the gnostemic and sememic strata. thereby

accounting for part 1) above. an example from Ikegami

(1970. p.6Le) will be used.

 

Hg/jounfi/

Ike 3am!

we Wreememes

mstemes cf the

Mitt “Vera“. it

all and that th!

1th “an‘ (hum
l

31+

H‘S/molle/ HS/kuman/

   
 
     

S/humah /

O
S/Man/

e

§£u31r/I

Figure 10
Ikegami's Hypersememic Diagram

The hypersememes (HS) of Ikegami's diagram are the
gnostemes of the present discussion. Note that there
exists diversification. in the realizations of young and
male and that the sememe S/boy/ can be interpreted as
_'y0un8 man' (human 4- male + young) or as a 'male child'

(“man + young + male).

 

 

 

 

 

hillustration of

mic and sememic s

C
C

MORSE 0E
7‘47 ’

 

tow

Inte
the GnOat:

35

An illustration of the integrated structure of the

gnostemic and sememic strata is given below.

 

 
     

 
   

CONTENT PURWRV

aim :
DIS COURSE 0
Tax";- .> Q’

 

*_ GNOSTEMES

Vt

‘I‘ACTIC. PLAN ES ‘

bfiMi K :
Rowan-cog

 

z
2
E
all)
3
5%:

Figure 11

Integrated Structure of
the Gnostemic and Sememic Strata

 

 

 

 

 

22.2 W
2.2.2.1 293220.121.

Lyons (1963' P
‘eteehnique for th
eenntic relations
sentences containir
eminentlal analye
19%) consisted of
hints of amine

Proportions 3
Mill “Omani c

222 the featureS ‘
thieease). (male
were extract“1 as

Katz and W1
ewtheabove. 1301
features. termed
temple being th'
emetic markers
lltitan) and (meal
titers in the f
it [young heist
Elwht] .

The stratii

lot. It rec 081

36

2.2.2 Componential Analysis

2.2.2.1 Definition
Lyons (1968: p.1e76) defines componential analysis as

"a technique for the economical statement of certain
semantic relations between lexical items and between
sentences containing them”. Previous work in semantic
componential analysis (of. Goodenough, 1956 and Lounsbury,
1956) consisted of identifying distinctive features on
the basis of analogy , or proportional relatedness.

Proportions such as the following were set up
man: woman: child we stallions mare: foal

and the features (human) vs. (non-human, or equine in
this case). (male) vs. (female), (adult) vs. (non-adult)
were extracted as components.

Katz and Fodor (196M distinguish components such
as the above, termed 'semantic markers' from item-specific
features, termed 'semantic distinguishers', the classic
example being the lexical item bachelor with two of its
semantic markers (i.e., features or components) being
(human) and (male) and two distinguishers under these
markers in the feature hierarchy [who has never married]
and [young knight serving under the standard of another
knight] .

The stratificational approach makes yet another distinc-

tion. It recognizes the fact that "the meaning of some

 

 

 

 

 

  
   
    
  
  
  
  
   

enie can be analyz
‘ hielso that
some investiga
the mistake of
can be analyze
for example. t

components in
plus a coupon

(Ismbe 1969 ,

lecrucisl state
intent is one whi
anemic signs. T
lithe taxonomic
lactic distingu
lithe structure w
hie they are pat
hose semantic fe
ll.e., which occu

llms) are consié

3.2.2.2 Univerg

Although it
lollies are lens
Mine given E

(knowledge on

 

37

words can be analyzed into components" (Lamb: 1969, pp.(+6-7)
but also that

some investigators of semantic structure have made
the mistake of supposing that taxonomic structure
can be analyzed in terms of sememic caponents:

for example, that m should be analyzed into
components including all the components of evergreen
plus a component which distinguishes pines from
other evergreens. But such a component would occur
only as a component of pine. A true sememic component
occurs in two or more different sememic signs. For
example, femalg occurs not only in mare, but also
in 33;, ow, vixen, h_e_n_. sisteg, Queen, and so forth.

In addit on, true semantic components are sememes

and as such they articipate in the sememic syntax.

(Lamb: 1969. pp.leg-7)
The crucial statement here is that a true sememic com-
ponent is one which occurs in two or more different
sememic signs. This in effect says that those features
of the taxonomic hierarchy which are item specific (the
semantic distinguishers of Katz and Fodor) are not part
of the structure of language ( 1.e.. in the stratificational
sense they are part of the gnostemic stratum). Only
those semantic features which function in the semotactics

‘ (i.e., which occur as components of two or more sememic

signs) are considered components in the linguistic sense.

2 .2 .2 .2 Univergal Status

Although it seems obvious that the taxonomic hier-
archies are language specific (and probably even speaker
8Decific given each speaker's greater or lesser degree

01‘ knowledge on a particular subject). whether or not

 

 

 

 

 

  
  
  
  
  
  
  
  
   
  
  
  
  
 

meet the cmponen
lot so obvious.
one. Lyons (1968s 1:
hot empirical avid
oeeeent time would 1
the hypothesis . '
osmotic components

on Isle indefinite

 

insole aware that
she new distincti
taxonomic hierarc
there exists a oer
[whoops a subset
levels of the war
eeeos that every
olmlflcant disti
hose distinction:
lie is only hypo
into which, if a

lo the arguments

3.2.2.3 Related
in interest
healer has the
Matter how 1‘!
liferentiate 1‘

‘39 speaker 1180‘

 

‘-—.--1I|.-|'" IIIII

38

some of the components of these hierarchies are universal
is not so obvious. Katz and Postal (1961+) suggest that they
are. Lyons (1968: p.1e73). on the other hand, says that
”such empirical evidence as there is available at the
present time would tend to refute, rather than confirm.
this hypothesis.” lntuitively it seems that the set of
semantic components cannot be a closed set since a speaker
can make indefinitely7finer distinctions once he or she

is made aware that a distinction exists, and in order to
make new distinctions one must add new features to the
taxonomic hierarchy. But intuitively also, it seems that
there exists a certain functional subset of these components
(perhaps a subset of those which function on the sememic
levels of the world's languages) which is universal. It
seems that every human language would make a linguistically
significant distinction between (male) and (female) since
these distinctions are overtly manifested in the universe.
This is only hypothetical, however. Actually the decision

as to which, if any, components are universal is not crucial

to the arguments herein presented.

2.2.2.3 Related Phenomena

An interesting fact about semantics is that each
speaker has the capacity to describe any item or phenomenon
no matter how fine or gross a distinction is needed to
differentiate it from another and no matter what language

the speaker uses .

 

 

 

 

  
  
  
  
  
  
 
 
  
  
   
  

heoontext to incl
nominee appears
hinolined to ask
ill-defined struct
hing able to comp
should not prohibi

oitheir structur

2.2.2.1) Co itiv

is far as th
«ponents of m
ninetificaticnali

Possess correlat

2.2.3 linkages:

The notion o
of simplicity wi
The is, the nor
he required to
Inked a portion

Verbal pai‘
line As useful
Mrs are simi]

the" towards

 

39

Another phenomenon which makes describing the entire
gnostemic and sememic strata for any given language utter-
ly futile is the fact that given any utterance interpreted
as nonsense under normal circumstances. one can extend
the context to include, say. the surreal, and then the
utterance appears semantically well--formed.8 One might
be inclined to ask "Why bother to try to study such an
ill-defined structure?" The answer is that fear of not
being able to completely describe these stratal systems
should not prohibit onefrom trying to describe parts

of their structure .

2.2.2.4 'Cosnitive Validity
As far as the cognitive validity (or reality) of

components of meaning, it is quite safe to say that
stratificationalists presume that such components actually

possess correlates in the mind of the speaker.

2.2.3 Markedness
The notion of markedness has to do with the notion
of simplicity with respect to the relationship of composition.
That is, the more marked a unit is, the more components
are required to define it. And conversely, the less
marked a particular item is, the fewer its components.
Verbal pairs such as gaging and 3933, 99m; and 52
serve as useful illustrations. It is obvious that these
Pairs are similarly related by the direction they imply

(i.e., towards or away from the speaker). It is also

 

 

 

 

   
  
  
   
 

hlons that they i
the more marked
considering the fac
tone, which cover

olonientation' (

oonteininoe the cam
otlll more sdequat
this composed
hoeen to be the
hrkedness i
tional descriptio
termed 'antoms‘
him like 93mg 8
enlyzed componen
lower linguistic
lid girl are clear
heather. The C
lMair for Eng:
one 'liow gig is
“she?“ unless
With of the per
his m can 1
“he lexemes 5

5m 'ant onyms

 

‘—_' r—rr..-c’l-OI '

no

obvious that they involve motion. The decision as to which
is the more marked member of each pair is arrived at by
considering the fact that 52 and take ”are more general
terms. which cover movement without a specific direction-
al orientation" (Lamb: 1969. p. 1m ). Thus 993;; may be
analyzed as a sememic sign composed of the sememes 59

and direction toward. while lying may be analyzed as
containing the components tag plus direction toward. A
still more adequate analysis might be found in describing
Elie. as composed of go + carrying (something). so 59 can
be seen to be the least marked of the four.

Markedness is relevant to each level in a stratifica-
tional description. Pairs which have been traditionally
termed 'antonyms' (and which were grouped together with
pairs like 29g; and 32) such as mg and old may be
analyzed componentially in stratificational terms on a
lower linguistic stratum, namely the lexemic. 193mg
and pig are clearly related by the fact that one is 3:12;
the other. The question as to which is the unmarked of
the pair for English may be found in questions of the
sort “How 9_l__d is she?” where one does not ask "How m5
is she?" unless the speaker is focusing attention on the
Youth of the person about whom he is asking the question.
Thus mpg can be analyzed as a lexemic sign composed
of the lexemes 9;}; and 319;. The rationale for considering

these 'antonyms' on a lower level than the verbal 'antonyms'

 

 

  
   
  
    
  
 
  
  
  
  
  
  
  
  
  
 

 

that

The native sp
opposite of bi
the opposite
This knovledg
for according
serve equally
he have thus
lithe units of th
hvenentionerl not
truth.
It might be
hononio hierarc
mused of fever
he set to which
mud daisies. p
Thus the ter
than the term fo
has supportive
he child genera]
learns (whether 5
«the general t1

Maps even to

m.

3M The Ghost
No real pii
liven. lamb ha;

Sch as the tax

 

1+1

is that

The native speaker of English knows that the
opposite of m is little and not small; that
the opposite of large is small. not little.

This knowledge clearly does not involve meaning,
for according to meaning. small and little could

 

serve equally well as opposites of b 5. (Lamb: 1969, p.145 )

We have thus far discussed markedness with respect
to the units of the sememic and lexemic strata. but we
have mentioned nothing about markedness on the gnostemic
stratum.

It might be that the unit flower in the gnostotactic
taxonomic hierarchy of vegetation might be said to be
composed of fewer components than any of the members of

the set to which it is related by extension.9

namely
annual daisies. perennial chrysanthemums, etc.

Thus the term flower might be said to be less marked
than the term for any specific flower. It might be mention-
ed as supportive evidence that in language acquisition,
the child generalizes the word for the first flower he
learns (whether it be a particular flower. say, daisy.
or the general term ;:lower) to include all flowers and

Perhaps even to include the still less marked category

plant I

2.2.14 The Gnostotactics

No real picture of the gnostotactics has ever been
given. Lamb has alluded to the structures it contains

such as the taxonomic hierarchies mentioned before. But

 

 

  
  
  
  
  
  
  
  
  
  
  
  
  

hue the gnostotac
munchiedge, it
hlspoiut in the :1
lies. and probably
test my overall
heme in partic
lush of the gnos
but and the struc
ledge oi referents
The present
lulled subportio

puiions which ar

hecrphotactio 0

hi provide for
sashes (i.e., 11
he semotactics
Stratum and the
enemies may b!
Prism the dos}
he at this 1
936 described.
his of ghost!
”Mal and top

Wesponding

 

 

1+2

since the gnostotactics is thought to contain all of
human knowledge, it is rather presumptuous. at least at
this point in the development of the science of linguis-
tics, and probably also for a good while to come. to at-
tempt any overall precise formalization even for one
language in particular. Topics which come under the
domain of the gnostotactics are such things as the con-
text and the structure of discourse and mutual foreknow—
ledge of referents.

The present study will concern itself with very
limited subportions of the gnostotactics. only those
portions which are relevant to the ultimate definition

of morphotactic classes.

2.1.5 The Semotactics

 

The semotactics, whose domain is the proposition,
must provide for possible well-formed combinations of
sememes (i.e., linguistically functional units of meaning).
The semotactics is the structure which defines the sememic
stratum and the syntagmatic configurations which it
Renerates may be roughly compared to the sentences com-
prising the deep structure10 of transformational grammar.
It is at this level where such notions as deep level case11
are described. Semotactic realizations of the different
types of gnostemic focus (predicate. propositional. func...
tional and topic) ultimately insure their

corresponding surface structure (i.e., lexemic) realizations,

 

 

 

 

   
  
  
  
  
  
  
  
   
   
  
  
  
 

The semotactic
hheen certain pro
Maine. This re
inersnnple. the di
sentences is whethe
in to one analysis
in non-topicelize.
been generated by

thin; generated

 

2.) Strata Intern

2.}.1 Lexeuic

The stratum

 

train of the la:
tactics nust prov
hpsndent clauses
hostotectic prop
he senotsctics T
Marked as int
fluoride the v:
TIlnslizstion.
The lexotet
intimate levn
‘3' transformsti
TMinute-obj“
We realized

“1.7818, if n

 

 

43

The semotactics also accounts for relationships

 

between certain propositions which share a common semantic
structure. This relationship has been termed agnation.12
Fer example. the difference between active and passive
sentences is whether or not the goal is in "focus" (accord-
ing to one analysis) and the difference between topicalized
and non-topicalized sentences is whether "topic“ has

been generated by the gnostotactics and is thus capable

of being generated by the semotactics.

2.3 Strata Intervening Between the Sememic and Phonemic
2.3.1 Lexemic

The stratum below the sememic is the lexemic. The
domain of the lexotactics is the clause and thus this
tactics must provide for various realizations of dependency.
Dependent clauses are the ultimate realizations of the
gnostotactic propositions unmarked for "propositional" focus.
The semotactics has generated as dependent any predication
not marked as independent. It is left to the lexotactics
to provide the various permissable realizations such as
nominalization, relativization and subordination.

The lexotactics. in generating structures at the
approximate level of abstraction of the surface structures
01’ transformational grammar. accounts for the subject-
Predicate-object surface functions. Subject is the lore.
tactic realization of functional focus. According to one

analysis, if no other deep (or gnostemic) level function

 

 

 

 

 

his focus, then is

slim of agent as

W @9212
The domain of
precisely the morp
nlngicrd word). I
sisters and affix
hm that the tee
hi of the morph

elenents but also
lexenic:
norphemic u

this study we
M and its

1.3.3 The Relgti
Hogphgnic

is suggests:
tin cases which
tortotactic oon
till they defir
he more usua‘.
hlized as uni
the being re

[Mb (1977

Wm on of '

 

 

ML

 

is in focus, then the lexotactics provides for the reali-

zation of agent as subject if there is an agent.

2 .3 .2 Morphgmic

The domain of the morphotactics is the word (more
precisely the morphological word, in contrast to the phon-
ological word). It provides for well-formed combinations
of stems and affixes. An illustration from Lamb (l9?la.p.ll9)
shows that the tactics of the lexemic stratal system and
that of the morphemic account not only for different

elements but also for different ordering of these elements.

lexemic: 3rd sg. erfect assive take
-z + be -en + take -

morphemi c s have en

In this study we will be concerned with the lexemic element
113;; and its various morphological realizations.
2.3.3 ‘lhe Relationship 9: the Gnostemic System to the
Mor h m c S tem

As suggested by Lamb (1971:!) there appear to be cer-
tain cases which call for a direct link between certain
ghastotactic configurations and the morphological classes
which they define. This is not the usual case, however.
It is more usual for units on the gnostemic level to be
realized as units on the sememic and then lexemic levels
before being realized as morphemes.

Lamb (1971;) posits a new device to handle the direct

connection of morphemes to meaning while guaranteeing

 

 

 

 

  
  
  
  
  
  
  
  
  
  
  
  
 

heir participation
ullethie device 1:
ludiancnd node 0
cede into the tact
more locations
leer level are (11
lllterent connecti
their different

level.

Wt The Relati
Phonemic S

The morphemi
end thus more clc
lurphemic. The d
l the syllable.
include the phon
1lllchl clause, e

lle have seer
lithe morpholog:
Marlee. for
tile boundaries
lit the tactic:
"not. (or. Se
3 the syllable

The units

dPhonemes.

M 2'1“]- e)

45
their participation in higher level configurations. He
calls this device the 'representative'. The representative
is a diamond node on a higher level which connects up-
wards into the tactics of that level and downward to one
or more locations at the level below. These -emes on the
lower level are distinguished from one another by their
different connections to the gnostemic system and sometimes
by their different connections to the tactics of the lower

level.

2.3.4 The Relationship of the Morphemic System to the
Phonemic System

The morphemic and phonemic systems are contiguous
and thus more closely related than are the gnostemic and
morphemic. The domain of the phonotactics is most common-
ly the syllable. However, the domain may be expanded to
include the phonological word, phonological phrase, phono-
logical clause, etc.

We have seen that the domain of the morphotactics
is the morphological word. It is quite clear that morpheme
boundaries, for the most part, do not coincide with syl-
lable boundaries in any given language. Thus. it is clear
that the tactics of the two levels are separate and dis-
tinct. (Cf. Section 2.4.Le for a more detailed discussion
01‘ the syllable.)

The units defined by the phonotactics are stratification-
81 phonemes. (For more on the S—phoneme, of. Sections 2.1.1

and 2.4.1,)

 

:_ i

The norphemic

r uetel to expressi
thought to be t
then the morph

 
     
   

The elements
telephones of
lee-phone. liorpho
Inlte or the phon

it heals:
Md The Phone
The phonolo
eltlo strata --
stratum is a par
stratum is thong
eiratel systems
The phonemic sys

environments of

 

hencrphonic 31
tried stratific
lute ambiguity
tional phoneme ‘
mi.

The tactic
that of stratf
let there is '

llent in level

 

1+6

The morphemic system is the closest grammatical
system to expression substance. The phonological system
is thought to be the only linguistic stratum intervening
between the morphemic system and expression substance.

The elements on the two levels are related as follows.
The morphemes of the morphemic stratum are composed of
morphons. Morphons are in turn realized as phonemes (the

units of the phonotactics). which are composed of phonons.

2.1+ Phonology
2.“.1 The Phonemic and Phonetic Strata

 

The phonology of stratificational grammar consists
of two strata -- the phonemic and phonetic. The phonemic
stratum is a part of language proper, while the phonetic
stratum is thought to be outside of language. These two
stratal systems handle essentially different phenomena.
The phonemic system handles the realizations and tactic
environments of the "segmental" alternations extant in
the morphonic alternation pattern. The realizations are
termed stratificational phonemes, a term coined to elim-
inate ambiguity with classical phonemes. The stratifica-
tional phoneme is more abstract than its classical counter-
part.

The tactics of stratificational phonology is stated in

terms of stratificational phonemes. This is not to say

that there is no place in the theory for a biunique level equiv-

alent in level of abstraction to the classical phonemic.

fl

the contrary. the
re oi the points
hcglv adhered t
here to be a link
«lame and th
have event be act
1 eetitiee and the p
r The ctratif
level -- the phcn
viretiriceticml
tie the limguis
language and exp
every phonemic s
llee voicing ever
the uorphon m/d
vlth [e] and vic

 

try dietributic
llroment.

1 Currently “
he phonetic 10
Interface. Th!
umterpart. h!
“phenomena If
"inflation 1|
Manage. and

internaticallz

 

it?
On the contrary, the maintenance of a biunique level is
one of the points which most stratificationalists have
strongly adhered to. They would argue that in order for
there to be a link between the phonic correlates outside
of language and the phonological units inside language.
there must be mutual predictability between the phonetic
entities and the phonemic.

Thus stratificational theory provides for such a
level -- the phonemic sign level -- which is below the
stratificational phonemic level in the realizational plane.
It is the linguistic level closest to the interface of
language and expression. It is here that. for example.
every phonemic sign P8/t/ in a language with non-distinc-
tive voicing word finally, whether it was related to, say.
the morphon MN/d/ or the morphon MN/t/. would be associated
with [t] and vice versa. except in the case of complemen-
tary distribution where the variant is predictable by en-
vironment.

Currently the connections between the phonemic and
the phonetic levels are thought to be across the language
interface. The phonetic stratum. like its gnostemic
counterpart, has not been studied a great deal. The types
0f phenomena which it is thought to handle are phonetic
assimilation within the domain relevant for the particular
language, and specification of phonetic components not

8S’stematically distinctive.

 

 

it is general

  
  
   
  
  
   
 
 
   

tullv analyzed.
nlmticn.

In a componen
uehpheeic is n
elpninte of arti
lieelf. That is.
reticular coupon
in. In a strati
ertnally play a
mi only these 1

in example
in llcdern Iraqi
(See the List of

Componentialc

Lb

 

b8
2Jh2 ngpogential Analysig
It is generally agreed that phonemes may be componen-
tnally analyzed, and that these components play a role in
explanation.
In a component theory such as that used in S-grammar
the emphasis is not on the contrast along the dimension
<m points of articulation. but rather on the component
itself. That is. the emphasis is on whether or not a
particular component plays a role in the phonological sys-
tem. In a stratificational analysis only a few components
actually play a role in the phonology of a given language
and only these few are generated by the phonotactics.
An example of a stratificational componential analysis
for Modern Iraqi Arabic is given in the following tables.
(See the List of Symbols and Abbreviations.)

TABLE II

Componential Analysis of Modern Iraqi Arabic
Consonantal Phonemes

Lb Ap Rz Fr _ Ph Gl
- p t c‘S k 7
01 Ph t q
Vd b d 3 g
__ r c s h x n h
3P Ph d s
Vd a z s '
- w y
N3 m n
~ 1
Ph 1
Vb .3

 

 

1 Ant
Componentia ‘

Fr -
1 a

e-
L111: a

The componem
universal status.
elnce all languag
Ihieh function as
eonants. Other c
hmever. And if
Wires differel
filletlfiable by t!

for the given 1a

W Harkedneg
In a recent

Inkedness must

”19 1n slmplic

“T markedness 1

deflmtion of a
Med wither
Wm“ am:
Manama}
related to 81m

“when, 3

 

1&9

 

TABLE III

Componential Analysis of Modern Iraqi Arabic
Vowel Phonemes

Fr Lb

 

The components used are not automatically accorded
universal status. ‘ The component vocalic is universal
since all languages distinguish between those phonemes
which function as vowels and those which function as con-
sonants. Other components are not as easily Justified.
however. And if the description of any given language
requires different. or more. components. their use is
Justifiable by the £213 (i.e., the system of relationships)

for the given language.

 

2.14.3 Markednegs

In a recent article. Lockwood (1969) implied that
markedness must first be considered with respect to its
role in simplicity. The universal status of the concept
of markedness itself is not denied. but the premature
definition of a certain vowel or consonant as universally
unmarked without sufficient empirical Justification is
cautioned against.

Markedness. on the phonological level. is directly
related to simplicity with respect to the relationship of

composition. just as it is on the semological level. That

 

 

 

 

 

nnne fewer compo
n‘nnnene. the less n
:m that defining
ne needed to defln
lsneld to be comp
being its only can

nnked vowel for a

2M W
The most 11m
“Hello, There
lefinltlon of sue
definitions have
llllnble 18 not (

nee nnnnlfestmol

nnnstnntively do

cf linguistic f0

mflltlon in, and

51° ”stem.

Neither ca,

slums. excep'

”hm These 0
the
no

lens. an

mam‘Ies of t

Men funct]

n Slllable min

nn
8

 

 

50

is. the fewer components needed to uniquely define a

 

phoneme. the less marked it is. If no components other
than that defining a class as a whole. e.g. Vo(calic).

are needed to define a particular phoneme, then the phoneme
is said to be completely unmarked. In the case of V0

being its only ccmponent. it would be said to be the un-

marked vowel for a given language.

2AA The Phonotactics
The most limited domain of the phonotactics is the

syllable. There has been much argument as to the empirical
definition of such a linguistic structure. None of the
definitions have proven satisfactory. however, since the
syllable is not open to empirical definition. While it

has manifestations in expression substance. it is not
substantively definable. The syllable. as a configuration
of linguistic form, is definable only in terms of its
function in, and relationship to. the rest of the linguis-
tic system.

Neither can we define the component parts of the
syllable, except by their function with respect to each
other. These components are usually termed the onset.
the nucleus. and the coda. We will accept them here as
Primitives of the system which achieve their definition
by their function in that system. Vowels usually function
as syllable nuclei. while consonants usually function as

onsets and sodas.

 

 

 

 

 

 

The phonotacti
syllable. however.
inning exist for

he phonological l

2.3.5 The Phonetj

 

The phonetic
hon. allophonio ‘
phenomena. The d
“Wee specific

innalio.

 

51
The phonotactics is not limited in domain to the

syllable. however. since larger relevant phonological

domains exist for some languages. These may be termed

the phonological word and phonological phrase.

2.’+.5 The Phonetic Tactics

The phonetic tactics provides for phonetic assimila-

 

 

tion. allophonic variation. and other such sub-phonemic

phenomena. The domain of the phonetic tactics is usually

language specific maximal clusters, both consonantal and

vocalic .

 

 

 

 

 

 

 

n hneiaee as her
inning between mean

in expression sub:

2. This is not 130
hence on the d"
isteneehre. irube
he figured promi
1? or indirectly) .
in greatest infln
he Swedish 1111811

3. Lint lndepend
intervening at
that the first

into which he

i See [AW 8 "
Sydney in. La
‘.. Yoshihiko In
t "Indefinitel

”Minion f

I' mtinted 1
“item (1s

 

 

If

FOOTNOTES

Chapter II

1. Language as here used means that structure inter-
vening between meaning (or content substance) and sound

(or expression substance).

2. This is not to say that other linguists had no in-
fluence on the development of the theory. Undoubtedly

de Saussure, Trubetzkoy, Bloomfield. Chomsky, and others
have figured prominently in its development (either direct-
1y or indirectly). But Hjelmslev and Hockett Were probably
the greatest influences. Lamb also frequently mentions

the Swedish linguist Noreen as being influential.

3. Lamb independently realized the necessity for positing
an intervening stratum while working on his dissertation.
He was the first to assign linguistic significance to this

stratum which he at that time termed morphophonemic.

Ln. See Lamb's ”Some Types of Ordering” (1972).

5. Sydney M. Lamb (1971b).

6. Yoshihiko Ikegami (1971).

7. "Indefinitely" is in fact limited by, for example,

physiological factors. but we do not know what these limits are.

8. Suggested by David G. Lockwood in conversation. Also see
Adam Makkai (1972)-

52

 

 

 

 

:, W here ‘
nerelntionshlp of
nzntificetioml te
fininedas a set of

hindered o_r_s_ with!

h. the terminolog
hit odds. Trans
The sound to near
inward direction
the S. Folk has
in theories is n
ihehenvens"), s
in sound to meg
level is higher ‘

it surface Striic‘
n
.1. See Hum”

53

9. Extension here is a philosophical term used to denote

 

the relationship of a set (or class) to its members. In
stratificational terms the extension of a "unit" may be
defined as a set of ”elements" related to it by downward

unordered ors within the gnostotactics.

10. The terminology chosen by the two theories seems to

be at odds. Transformationalists refer to the progression
from sound to meaning as surface to deep. implying a

downward direction as with the earth and the sea (although
Julia S. Falk has suggested that the terminology of the

two theories is not at odds. using the phrase "depths of

the heavens”). Stratificationalists view the progression
from sound to meaning as upward. That is. the sememic

level is higher than the lexemic (which is roughly equivalent

to surface structure in transformational terms).
11. See Fillmore's "The Case for Case" (1968).

12. This term originally appeared in Gleason (1965).

 

 

(General Desert

:.1 muss:
It 18 lamb' 8

 

defined system. 1
the parable of tht
591m "At any 31‘
lathe to be con
lithered must a (1
More?" (p.101)
Tit-off point. 11
ihmmly contriw

Beckett ack1
shebilitiea or m
Misting for In

“new to the

Hell-definition.
is to say no phs
he only reason
“What we have
Was a photo

a:
“aching the

 

CHAPTER I I I

A General Description of Iraqi Broken Plurals

3.1 General Remarks -- Language in F135;
It is Lamb's contention that language is not a well-

defined system. In a recent article (Lamb: 1971a) he uses
the parable of the tree to exemplify his conclusion. He
asks: "At any given point in time how far formed must a
leaf be to be considered a leaf; and vice versa. how far
withered must a dying leaf be not to be considered a leaf
anymore?" (p.101). Even if one were to impose a certain
cut-off point, it would be just that -- an imposition ~-
a humanly contrived division of a natural continuum.
Hockett acknowledges the fact that "there are many

stabilities of many orders" (p.810 in languages. some
persisting for milennia. He attributes Chomsky's central
'fallacy' to the fact that Chomsky saw in this stability
well-definition. Hockett, on the other hand, goes so far
as to say no physical system is well-defined. He claims
the only reason we get a stable picture is "because that
is what we have sought.” (p.83) To use his simile --
much as a photOgrapher who takes a picture of the horses

approaching the finish line is not surprised. on having

51+

 

 

he picture develol
he.

In a review a
pehte against Hoe
Mention of laugh
thence/performs
perhaps the centr
111-definition of
he he equates t
the well definit!
ehether rules of
‘f they do exist
require that sue
hustled sent

Given the 1

W380 as an 5

(h)! have for th
The Arabic
henchmen. of
““51th fret:
he ”responds
hlhel classes
{he my morph
“he the blur
“the. The c

55
his picture developed. to see that the horses are not run-
ning.

In a review article Lakoff (1969) makes many good
points against Hockett's arguments for analogy as an ex-
planation of language acquisition and his lack of a com-
petence/performance distinction. but he misses what is
perhaps the central issue of Hockett's treatise —- the
ill-definition of language. Lakoff misinterprets Hockett
when he equates the existence of grammatical rules with
the well definition of language. Hockett is not arguing
whether rules of grammar do or do not exist but whether,
if they do exist, it is a legitimate goal of theory to
require that such rules account for a well-defined set of
grammatical sentences.

Given the fact that stratificationalists accept
language as an ill-defined system, what relevance does
this have for the present study?

The Arabic pluralization system. if one accepts the
implications of Givén's hypothesis. is now in a state of
transition from an older system in which there was one-to-
one correspondence between semantic and morphological
Plural classes to a currently evolving system in which
the only morphologically marked category will be the
IEflllells/fieminine grammatical gender. As one views Arabic
today the pluralization system appears an idiosyncratic non-

8Yetem. The only generalizations which have been made are

 

 

those concerning t
the singular tc
see in this static
theithere, it be
hatch is e conflg
eeptiohe' that are
Ween the two .
classification of
“unit part in th
totem. These it
the ‘hrolten' plm

hips.
3.2 M!

It appears
Ihith are Charge
the few broke
their, that the
WWW it

The follow

56
those concerning the relationship of the canonical form
of the singular to that of the plural. If one tries to
see in this stationary portrait the mobile continuum cap-
tured there. it becomes evident that, indeed. the non-
system is a conflation of two systems and that the 'ex-
ceptions' that are present appear to be caught midway
between the two -- not conforming to the old semantic
classification of the particular plural class and yet not
taking part in the regular infle ctional plural of the newer
system. These items seem to have been shuffled around
the 'broken' plural classes on the basis of phonological

shape .

3.2 The Old Semantic Classification

It appears that there are certain semantic features
which are characteristic of large numbers of items in more
than a few broken plural classes. To actually try to
specify what the old semantic classification must have been
in toto would be impossible. But a few hypotheses may
be put forward regarding the general skeletal framework.

The following semantic features may be extracted:

 

abstract

round -- incl
0

15

square -- in
strong -- in
weak -- incl
male genital

female genit

other body 1

plants -- 1
colors

atent
location

1nStrlunent

These feet
litter go be me

other 3138.1
htnor “asses ‘
teacherries 815
the 01d “Men
“all the “330:
thereotenzmg
iterate from
homes no“ 8

the “miles:

57

abstract

round -- including convex. concave, circular. spherical,
cylindrical and rod

square -- including cubic

strong -- including high, respected and warring

weak -- including defective. low and childlike

male genitalia -- including males and phallic symbols

female genitalia -- including cracks. fissures and
holes

other body parts -- with a possible distinction between
paired and non-paired. including
body functions

plants -- including plant parts

colors

agent

location

instrument

These features (above) are characteristic of what
appear to be major semantic classes.

Other smaller minor classes often appear. These
minor classes will be mentioned in passing but will not
In accorded significant status in the discussion of
the old system. Perhaps each of the subcategories included
under the major categories above might be thought of as
Mmracterizing a separate semantic class, e.g.. convex as
separate from circular. etc. While the present analysis
asmmws more general categories. a more particular extension

M‘the analysis could easily be made.}

that of sifl

bulk

close

cover
statement
meal

female relat
death
elongated
time

Examplea of

liter and minor

3.2.1 W
The forms I
“We reallzat
to not Wrespo]
established by
"om or the Flu
metas %
:f the first. m
‘m exa”191e, at
We surface‘
‘.'.t tategmes

m
mm.
In Wu

0011301181)“

L

 

A

58

A list of significant minor categories follows:

bulk

close

cover

statement

meal

female relative peer or younger
death

elongated

time

Examples of semantic classes characterized by these

mayn‘and minor categories are found on the following pages.

3.2.1 Semantic Taxonomy

The forms presented under each heading (below) are
surface realizations. not underlying forms. Starred items
do not correspond to the most prevalent canonical form
established by the other members of the group. The canonical
form of the plural heads the example. It is in a form

mmh as FMuuLa where F, M and L designate the position

 

M‘the first. middle and last radicals of the items in

mm example, and the vowels are those actually occurring

Mlthe surface. Under each plural heading comes the seman-

No categories which distinguish it and the canonical forms
M‘singular items taking part in the particular plural

Mass. In quadriconsonantal entries, LL doublets signify tWO f1_
nal consonants which are the same, while MM doublets signify

any two medial consonants.

 

The lists in
monlcal form 01‘
pcht 1n the analy
plural classes inc'
semantic features
hnonlcal forms 0:
mnonlcal forms 8
lbll reflects a a
slmllmr to the ct
hhahlll there

all Plural such :

kltabu
kltu
hltl

m(the
mulll

lmorlng semant
hghauy' for
llth/kl/ has a
that Mike 8111

llth /Ul/ “nee

ml the Singuj
hm“ begin
h Plum Dre
hm prefix

!
ll! Choices

59
The lists in this section (3.2) are organized by
canonical form of the plural because this was the starting
point in the analysis for investigating whether certain
plural classes indeed were characterized by one or more
semantic features. The predictability of certain plural
canonical forms on the basis of the corresponding singular
canonical forms such as was described in Levy (1971) prob-
ably reflects a situation in an earlier stage of Semitic
similar to the current situation in Bantu. For example,
in Swahili there are pairs of classes designating singular

and plural such as

kitabu p1. vitabu 'book(s)'

kitu p1. vitu 'thing(s)'

kiti p1. viti 'chair(s)'. etc.
and

mkono p1. mikono 'arm(s)'

mwili p1. miwili 'body/bodies', etc.

Ignoring semantic characteristics one could predict phone...
losically, for example, that (almost) every noun beginning
With /k1/ has a plural form beginning with /vi/. One

cannot make similar predictions about singulars beginning

With /m/ since there are several plural prefixes associated

with the singular prefix I_I;-. A statement such as: any
singular beginning with /m/ will be associated either with
the plural prefix 33-. or the plural prefix m -. or the
Plural prefix m— may be made. That is. there are only

a few choices possible. But such an analysis is lacking in

 

heriptive adequat
h predictions on
higher level ass
this, certain s
Ihh certain gem
olinhili is not
W913. one cam
hthout including
theses. That is
“mar/Plural a
linemen”.

In Arabic, ,
shun/plural ‘
erller Pairing
tend to “Sign
singular “Mule

This Sectic
{gunption of t
hm °anonical
M5“ and m
framework) is 1):

liter chec‘
“Water with
ha“tie charac
haof items ‘
Mined u
when cl

Jill Glass) l

60
descriptive adequacy, since it soon becomes evident that
the predictions on the basis of canonical form follow from
a higher level association of semantics and morphology.
That is. certain singular/plural prefixes are associated
with certain semantic characteristics. While the system
of Swahili is not as simplistic as these examples tend
to imply, one cannot adequately describe the language
without including a reference to semanto-morphological
classes. That is. one cannot adequately describe the
singular/plural associations as a purely phonological
phenomenon.

In Arabic. what correspondence there is between
singular/plural canonical forms probably reflects a) an
earlier pairing similar to Bantu and b) a present-day
trend to assign plurals analogically, on the basis of
singular canonical form.

This section (3.2) is intended, then. as a general
description of the semantic correlates found for all the
plural canonical forms. A theoretical description (i.e.,
the data and findings described via a modern theoretical
framework) is presented in Chapter 4.

After checking through lists of singular items
associated with a particular plural canonical form. the
semantic characteristic(s). by which relatively large num-
bers of items were related, were noted. Singular items
which formed the core (i.e., which were associable with
the semantic characteristic(s) extracted for the given

plural class) were then presented to two native speakers

{totem Iraqi Are
they felt that i
mtlf so, what sen
loaorelete the gf
tally the infon
presented here.

The following
hhnltyl “Jill
hemswhich are as
lhrticular C8110!
m“ M. u:
hates” mum 1
[W] s'cless.
““1 llassh &
hlchpmess a c
)5 Was FvliL.

Items not 11

Emits are 011111

43103 do not 00,
h
(tension in t]

ly'lchalpear to

his or P‘Bhape

61

of Modern Iraqi Arabic. The informants were then asked

if they felt that these items were semantically related
and. if so, what semantic characteristic(s) they would use
to so relate the given singular forms. The classification
given by the informants corresponds to the classification
presented here.

The following terminology is used to eliminate
ambiguity: Plural olas , used to refer to the set of
items which are associated with a particular plural of
a particular canonical form. e.g.. the FMaaLi plural

classy S-class. used to refer to the set of items which

 

possess a common semantic characteristic, e.g.. the
[round] S-class, (one or more S-classes may characterize a

plural class): P-class, used to refer to the set of items

 

which possess a common phonological singular shape, e.g..
the P-class FvML.

Items not listed in this and the following lists of
examples are omitted because their semantic character-
istics do not correspond to the semantic components under
discussion for the particular plural class. Such items.
which appear to be associated with a plural class on the

basis of P-shape. are discussed in Section 4.3.2.

Mllll FMI‘L P1

5;!

round

In order tc
between the 31"!
Wing plural .
fiven here. Th‘
Plural of Q
I°n6f113' is I.

Item 1m
mum, WI}
10% °f the its
{M the eleph;
1nd mm the ‘

i
"t' thick. b1

62

3.201e1 FM‘L Plurll €13.88

 

Se Phon (singulars)

round Fva
FvML(a)
FvaL
FvavL

In order to indicate more clearly the relationship
between the singular canonical forms and their corres-
ponding plural canonical form. a few examples will be

given here. The plural of kuub 'cup' is kwaabs the

 

 

plural of tall 'hill' is tlaalz the plural of £1113

 

 

'one fils' is glagg.

Items listed below are convex, concave. circular.
spherical. cylindrical and rod. They comprise approximately
70% of the items taking part in this plural class. Notice
that the elephant is one of the animals of this class
and that the adjectives are those descriptive of rotunditya

fat, thick, big. Note that one of the characteristic

 

nnnnicsl forms 01
nd the following
ran {on related
holes a predict
Honk middle rac‘
hing singular in
1mm, shore ‘
includes FwL. F

lives is thiiL.

MILL]. m

hub u
kill): '1
Juno n.
inub .
kuuz o
buuq s
*QUUII I
Em .
2113 n
fill
biir
Diip
hii

€11!

3|2I1.1.2

tall
“11::
nun
Zlbb
gill
iabb

 

63

canonical forms of the singular for the items in this

amithe following plural class is Fva which is the sur—

face form related to the underlying form FVwflL, where V

denotes a predictable vowel alternation inzu. and ! denotes

a weak middle radical. i.e.. y or w. Actually the under-

lying singular form for all the nouns in this plural class

is FvaL. where the second vowel may equal zero (which

includes Fva, FvML(a) and FvaL). and that of the adjec-

tives is FvMiiL.

30201e101

kuub
kuux
Juud
euub
kuuz
buuq
*quuri
siif
ziig
fiil
biir
piip
E11!

siix

302010102

tall
muxz
hibb
éibb
sill
dabb

Fva

'cup'

'hut'

'animal skin bag'

'tube'

'clay urn for storing water'
'bugle'

'teapot'

'slice' ‘37

'shirt-front opening' ‘\./’
'elephant'

'oil well'

'barrel'

'skewer'

'skewer'

F ML a

'hill'

'brain'

'pottery vessel for storing water'
‘penis'

'small snake'

'large lizard'

 

nngl.oont'd

has?
tang

sahba
sarss

'be

MAJ.) FLM‘

rsgiX
nuns
{islam
Euful
libal
Jams
Salas
31m
fills
islar
Elihu!
luhur
0ng1
Nam
I‘umuh

mm

6h

FvMLfa}. cont'd

hawd

tawg

tawl

tawb

halls
salla
gifts
gubba
kalla
dalla
lassa
xayma
gahba
garsa

302010103

ragig
tubag
balam
guful
Jibal
Jaras
Safes
nits;
filis
hajar
muhur
juhur
Ougul
Qalam
rumuh
fuxud

'basin'
'hoop'

'bolt (of cloth) '
'cannon'
'pigeon nest'
'basket'
'11P.

'dome'

'head'
'coffee pot'
'carcass'
'tent'
'whore'
'narghile'

FvaL

'pOckmark'
'basket'
'rowboat'

'lock'
'mountain'
'bell'

' cage' @
'shoulder'

'one file (coin)'
'stone'
'personal stamp'
'anus'

'weight'

'pen'

'lance'

'thigh'

CD

 

”Jill“ M

310110
Oixiin
kabiir
ksOiii

3.2.1.2 FliaaLa
§_en

[round]

The items
takins part in
that since the:
which are char
hips there was
hen ”181ml
i cilindrica]
Silently gene
“is f mm 1
The surf.
acthilly FVHL
hy°1°fabst
““1. Mime
ienensl “ride:

65
3.2.1.1.!» mm
310110 'fat'
Oixiin 'thick'
kabiir 'big'
kaOiif 'thlck'

3.2.1.2 FMaaLa Plural Class

 

§gm Phon (singulars)
[r ound] Fva

FvML(a)

FvaL

Fva

The items listed here comprise 50% of the items
taking part in this plural class. It might be posited
that since there are a good number of plural classes
which are characterized by the feature [round] that per-
haps there was a different class for each type of round
item Originally (that is. a class of circular, a class
0f Cylindrical. etc.), and that these classes were sub—
Sequently generalized to [round] causing the general
mixing found in each [round] plural.

The surface P—class Fva is at an underlying level,
actually FvML. i.e.. gal! 'bucket' is gal! at a deeper
level of abstraction. Again. all these singular P-forms
(FWL. FvML(a). FvaL and Fva) may be included under the
general underlying P-shape FvaL. where the second v may

equal zero.

 

3.21.2.1 mi

nuug ‘ in
hm: 'hn
fuun 'f!
puul 'cl
hmz ' c‘
hub '15
niil ' I
piip 'i
12.1.2.2 w
hiss '1
hibla '
ham '

”11.2.3 171;:

islu

32.L2|n EX!

$111111.
ourui
Rumor
iifir
(lien

 

66
3.2.1.2.1 LL
muug 'inner corner of the eye‘
kuux 'hut'
fuum 'foam'
puul 'checker piece'
kuuz 'clay urn for storing water'
cuub 'tube'
miil 'Indian club'<-/_:::::;:::J
piip 'barrel'

3.2.1.2.2 FVML‘a)

kuss 'vulva'
hibla 'prcgnant'
hayya 'snake'

302010203 FVLV

dalu 'bucket'

30201.20“ FVMVL

sufut 'basket'

quful 'lock'

Eumar 'moon'

sifir 'zero'

Xifan 'winding sheet'
tuxum 'cigarettc holder'
surum 'rectum'

3.2.1.3 FMaaLi Plural Clasg

§£fl Phon (singulars)
[round] FvMLv
vaLa
Fvai(yya)

 

 

The items here com;
insinp of the class. i
then. Twila, mum
innhhll plus a suffi:

12.1.3.1 Lam

kitli 'kettle'
‘um 'closecl1
W1 'hose 1’:

31211.3.2 FWHLS

“”333 'irriga
Bsagya 'irriga

3'2'1'30 FWLigua}

mun I bugle

3.2.1.1;
W

m Phc
W] N:
Fix]

The P-me Fii)‘
innFiinhLa Where
the underlying fem
Thitens here comp;

his or the Plural ‘
1.2.
kilnsa

Mara

67
The items here comprise one-third of the total mem-
twrship of the class. All the singular surface forms
(FvMLv, FvaLa. Fvai(yn)) correspond to one underlying
form thflL plus a suffix (either -i or -(y)a).

30201-301 FVMLV

kitli 'kettle'
'urwa 'closed. rounded handle'
qamé1 'hose for a narghile'

3.2.1.3.2 FWMLa

8&83y8 'irrigation channel’

to I.
saagya 'irrigation channcl' E free variants

33.1.3.3 FwLigna)
buuri 'bugle'
tquIyya ° night sti ck '

3.2.1.4 FiMaaMiLAJFuMaaMiL Plural Class

 

§£2 Phon (singulars)
[round] F ( 1 )Mvaa
FuMMvL

The P-form F(i)Mvaa corresponds to an underlying
fOrm F(1)MVMLa where the second M is a glottal stop. e.g..
the underlying form of kiliisa 'engraving'.is gillZߤ_:_2o
The items here comprise five—sevenths of the total member-

ship of the plural class.

3-2.1.h.1 F(i}Mvaa

kiliisa 'engraving'
sidaara 'common Iraqi headgear'

 

1‘ ”The cont' d

1131118 Wall 1
81mm '61”

312111113 w

mnsra' 'topU

1.2.1.5 F "1”“ P1‘

'33! Ph...2.n
[round] FY“
[mile]

This plural 0183‘
the nnsannn plural ‘
1nd hisahiiL plural
hisaiit plural claSS

All the items 1!
his that [male] slat
in exist plural alas

teristio feature.

12.1.5.1 remnants

*miizaan

' as
sindaana 'fi
smears ‘ s

T2.1.5.2 smears
Titiasl ‘m
‘rsiil '}

68
ngzhvaa, cont'd
fisiila 'palm shoot'
sidaana 'clay container'
ssiiba 'pigtail'
3.2.1.”.2 F'uMMVL
musra' 'top(toy)'

§gm Phon (singulars)
[round] FvMMaaL(a)

gale}
This plural class is a morphological alternant to
the FuMaaMiiL plural class below. Together the FiMaaMiiL
and FuMaaMiiL plural classes are alternants of the FiMaaMil
FuMaaMiL plural class above as length is (usually) predictable.
All the items in this class are accounted for here.
Note that [male] also characterizes this plural class. There
do exist plural classes for which {male} is the sole charac-

teristic feature.

3.2.1.5.1 FvMMaaLga) Round

*miizaan 'scales' (more abstractly m1?zasn)
sindaana 'flower pot'
gurraafa 'scoop'

3020105e2 FVMMaaL 9. M8 8

riJJaal 'man'
*rajil 'husband'

 

3.2.1.6 W
Sen 2222

[round] Fviiih
[square]

ill or the items 1
into that [square] a1 s
to [roundjh There do

lathe sole characteri

3.2.1.5.

3.2.1.6.1 FVHMWL Bot

naa‘umr 'uat
“Theses ' sci

1.2.1. .
62 thliva So

811an ' ate
lllltaan s 00‘
d1 1188b ' "as

3.2..
17 Win

3LT Pt
round] F1
[Square] F‘

F]

This Plural :1“
theiul. plural clag
nth the FMaaMuL pi

1
110115151“; t0 corret

liner tn. items !

69

3.2.1.6 FuMaaMiiL Plural Class

 

Sen Phon (singulars)
bound] FvMMva
@quare

All of the items of this plural class are listed here.
Note that [square] also characterizes this plural (in addition
to flound]). There do exist plural classes for which [square]
is the sole characteristic feature. See the note under

3.2.1.5.

3a201e6e1 WWW]... Bound

 

naa'uur 'water wheel'
*mugass 'scissors'
O

3.2.1.6.2 FvMMva Sguare

Subbaan 'steelyard'
liiwaan 'covered, paved area'
diilaab 'wardrobe'

3.2.1.7 FMaaMiL Plural Class

 

§£fl Phon (singulars)

EF°undj FvMMvLEa; FvavL(a)
FvavL a

Ewan] FMva a)
Fva (a)

This Plural class (FMaaMiL) is an alternant to the
FMaaMuL plural class below. While the items associated
With the FMaaMuL plural class provide the necessary en-
virOmnent to correctly predict the final vowel as E- 3° do
many or the items associated with the FMaaMiL plural.

Therefore the two plural classes are presented separately.

 

 

Tons of the items 1

:tetno semantic classe
nu half (7/12) of the
this -- [round] , 1/3

The singular P-shg

respond to an underlyii

lean to a glottal std

corresponds to an uncle

“at be a glottal etc

to an underlying Fviivi

tlottsl stop.

3.2.1.7.1 FVWVLR) 1

Shutter
matter
mlhbas
manna
Zunguta
uni.
*giffa

'larg
'muff
'rinp
'mosq
‘pim;
‘akui
'11p

1
112sls7e2 Wm

{Wit
gasbir

eeye

‘ tem

Mild.) ISM)-

8bill . cl:
bzlim . b“!
31mg 'ne
'Taama ' tu

1.2.1.7.n mea
\J
maTina

'
“Ilka la

'1”

70

None of the items here really needs explanation.

'flw two semantic classes which are given here account for
overrmlf (7/12) of the total number of items in the plural
class -- [round] , 1/3 and [square] . l/h.

The singular P-shapes FvMMvL(a) and FvavL(a) cor-
respond to an underlying form FVMQMVL {ta}, where the first
M may be a glottal stop. The singular P-shape FMva(a)
corresponds to an underlying FQMVMQL §+a2, where the second
M may be a glottal stop. FvaML(a) and FvavL(a) correspond
to an underlying FVMVMQL {ta}, where the second M may be a

glottal stop.

3.2.1.7.1 FvMMvaa) Round

Sawsar 'large Z-handled basket'
masmar 'muffler'

mihbas 'ring'

masjid 'mosque'

zunguta 'pimple'
jimjima 'skull'

*é’irra 'lip'
302010702 FWMVL‘S.) Round
haajib 'eyebrow'
saabir 'temple'
302010703 FMWLSaz Round
sbiil 'clay pipe’
bziim 'buckle'
glaada 'necklace'
'maama 'turban'
3020107014 FWMLB. Rom—g
mayjna '1arge wooden PGStle'

dawlka 'pitcher'

 

 

302.1O705 w

sinisla ' c

3.2.1.7.6 mm

fitifla 'l

3|2I1|707 FVHMVLSE

nswgad 'ff
mag'al 'ff
dalgag 'm
hams '1‘

32.1.7.8 Fvvnng

Qualab 'fc
888d“ ' te
kangada ' s}

3I2|1.7,9 FMWL ‘1

Sfllfa s.-

33:10.10 FVMW

311‘111‘ o
tiriiga I

‘.2.1,
’ 8 M
Les

[round]
All the Mm.

the note under 3

71
302010705 FVMVML‘a) Round.
sinisla 'chain'

3.2.1.7.6 FYMVVL‘B) Round

fitiila 'wick'

3.2Ju7.7 FvHMvL§a2 Sguare

mawgad ' fireplace '
mas'al 'fireplace'
daws’ag 'mattress '
bayrag ' flag'

3.2.1.7.8 Fvangaz Sguare

qaalab ' form '
daadir ' tent '
kaagada ' sheet '

3.2010709 FMWL‘a) Sguare

sfiifa 'protective bordering'

3.2.1.7.10 Fvangaz Square

siriir ' bed'
tiriiga 'slat'

3.2.1.8 FMaaMuL Plural Class

 

 

m Phon (singulars)
[r ound] FuMLv ( v )L

All the members of this class are listed here. See

the note under 3.2.1.7-

 

 

}.2.1u8u1 Flu—m

[MSW ' 1e
nutbsg 'd‘
nuglsab '81

3.2.1.9 M“!
figs;

[round]

These are the
plural class is er
below. The Mn 1
pironnent in this

separately.

3.2.1.9.1 EMMA

'asaaya
mraaya

72
3.2.1.8.1 FuMMvgsz

gumgum 'large copper coffee pot'
mutbag 'double-tubed flute'
muglaab 'garbage dump'

3.2.1.9 FuMi Plural Class

 

Sem Phon (singulars)
[round] F ( v ) Maaya

These are the only two members of this class. This
plural class is an alternant of the FiMi plural class
below. The invu alternation is not predictable by en-
vironment in this case, thus the classes are presented

separately.

3.2.1.9.1 FszMaaza

'asaaya 'stick'
mraaya 'mirror'

30201.10 FiM; Plural Clasg

Sem Phon (singulars)
[round] ?

The underlying P-shape for the members of this class
and the one above (3.2.1.9) is FvaL plus an optional suffix
H.0r-ya), Either v can be t and L may be a glottal stOp.
There are only three members in this class. two Of which

are listed here. The third is ZQLQl 'blind'.

33.1.10.1 FaMLfaaya)
*‘aba 'aba' (more abstractly 'ab?)
rahhaaya 'grinder'

 

 

3.2.1.11 E;
'83.! .
[round
There:
happen:
dalag

3.2.1.12 g
pep

[stror
[absti
[bulk

This 1
listed hen
strong. 1/
[on],

The [
Notice the
To War. E
Waning
Writing;
included I
We a:

the 190153:
The
chapter

sum“

 

73
3.2.1.11 FaMLaat Plural Class

 

Sem Phon (singulars)
[round] '2
There are only two members in this class:

bamyaaya ' okra '
dalag ' column '

3.2.1.12 FuMuuL Plgal Class
Sem Phon (singulars)

[strong, war] FvaL
[abstract] FVML
[bulk] -- minor class

This is quite an interesting class. The items
listed here account for about 7/9 of the total membership --
strong, 1/33 bulk. 1/33 abstract. 1/9. One of the classes.
[bulk] , is a subclass.

The [strong, war] class is the most interesting.
Notice that the items in this class all relate in some way
to war. Even the pawn from the game of chess is included.
L1Shtening and thunder -- natural phenomena which are
oftentimes associated with the wrath of the gods -- are
included here. Another interesting thing is that the
animals are the predatory "king of beasts" type " the tiger.
the leopard and the lion.

The items in the subclass [bulk] are all somehow
Characterized by large size. large numbers, long time or

collectivity.

 

 

The singul
numhrlying 1
th vovvel in o:

This plur;
ins(FihuuL)
vovel alternat

To arrive
[abstract] , it

1)items
Tflt.

ZTitems
“Sin Sectio:
That the v;
she.

3) items
Tract. while

”Te Sven.

3'2'1'1201 I

{‘8st
mslik
fariq
'aris
fatal
rahan
I‘a ' 3d
“sum
{hub
Tiriz
{We
Win

7a

The singular P-shapes FvaL and FVML correspond to
an underlying FvM¢L, where fl may be realized as an anaptyc-
tic vowel in case the final two consonants are dissimilar.

This plural class is an alternant of the following
class (FiMuuL). They are presented separately because the
vowel alternation inJu is not always predictable.

To arrive at a working definition of the feature
[abstract] , it was decided to include

1) items which did not refer to concrete things. e.g.,
samba.

2) items which were relative. e.g.. 1/9 is abstract
(as in Section 3.2.1.321) since its Value is dependent

on what the value of the whole is. while million is con-

 

crete.

3) items which were not specific. e.g.. time is ab-

 

stract. while day is not, since day refers to a specific

time span.

3.2.1.12.l FvaL Strong

habis 'imprisonment'
malik 'king'

fariq 'difference'
'aris 'throne'

tabul 'drum'

rahan 'pawn (chess piece)’
ra'ad ‘thunder'
xasum 'opponent'
harub 'war'

hiriz 'amulet'

barig 'burn'

husin 'fort'

A

 

 

 

hhd Strcpy
vahid
nimir
fehad
Tasad

TZJ.12.2
hadd

lanz
kals
simn
sadd
barq
xstt
ssh

”$12.3

'urud
dam}
lobar
baliv
Tami
lali
Isms
Gall
'asi
Tali
Girl

 

75
FvaL Strong, cont'd
wands 'wild animal'
nimir 'tiger'
fahad 'leopard'
?asad 'lion'

3.2.1.12.2 FvML Strong

hadd 'border'
hanz 'treasure'
ka?s 'trophy cup'
simm 'poison'
sadd 'obstacle'
barq 'lightening'
xatt 'line'

bur3 'tower'

 

3.2.1.12.3 FvaL Bulk

'urud 'width'
dahar 'long time'
saham 'fat'

haiim 'bulk'
Jami' 'plural'
7311f 'thousand'
xamur 'wine'
03113 'snow'
'asir 'era'

qarin 'century'
Oirib 'roe'

Zari' 'crop'
sa'ab 'people'
sahal 'plain'
Wahal 'mlre'
rafi§ 'variety of large turtle'
di'id 'barge'

 

hhlbulk. cc
'ltir

same
laham

hld.12.h

I"?

hand
hebb
raTs

”$12.5

I’I‘!

behiO
Tasil
salad
XaTin
ltl.12.6 1

hell

have

 

 

76
FvaL Bulk. cont'd
'itir 'perfume'
samug 'mucilage'
laham 'meat'
30201.12.“ FVML Bu1k
band 'bundle of 100 sheets of paper'
habb 'seed(s)'
ra?s 'cape (1and)'

3.2.1.12.5 FvaL Abstract

 

bahiO 'research'

?asil 'origin'

wa:ad 'promise'

ga?in 'matter, affair'

30201.12o6 FVML Abstract

 

hall 'untying'
haqq 'truth'

sank ' doubt '

hamm 'anxiety'
dann 'assumption'

3.2.1.13 FiMuuL Plural Class

 

§£E Phon (singulars)

[strong . war] FvaL
FvML(v)

Here again is a [strong. war] class and the character—
istic P-shapes are the same as those for the previous such
plural class. The P-shapes of these two plurals (this one ‘-
quuLsand the previous one -- FuMuuL) indicates that they
mm alternants. See the note under 3.2.1.12. however. Only

“use out of the eight members of this class share the seman-

tuscharacteristic [strong. war] °

 

phl.13.1
sijln

)h1.13.2

v
jays
Tundi

3.2.1.1v p
sap

htro

The i
sacral re
cwvethey
Thong. p
Fhmu (3'

hdape so
ms 01“!
MdIlh']

sahap

 

77
3.2.1.13.1 FvaL

sijin 'prison'

3.2.1.13.2 FvML v

Jays 'army'
Jundi 'soldier'

3.2.1.1h FiMaaL Plural Class

§gm Phon (singulars)
[8 tr orig] FvaL
FvMLv

The items here fall under the category [strong] for
several reasons: 1) because they pertain to war, 2) be-
cause they are respected, 3) because they are actually
strong. For a good example of [strong] via war. of.
FiMuul (3.2.1.13) and FuMuuL (3.2.1.12). For a related

P-shape see FiMaaLa (3.2.1.15). Six of the ten items in
this class are listed here.
3.2.1.1h.l FvaL

 

saham 'arrow'

3020101402 FVMLV

xisla 'good quality'

harba 'bayonet'

qal'a 'castle'

*diib 'wolf' (more abstractly diyb)
*nasiim 'wind'

3.2.1-15 FiMaaLa Plural Class

 

Sem Phon (singulars)
[strong] FvMLv

There a

niyys
nasal

3.2.1.16 E
M
[stror
This p

{shoving p

his class

tics of th.

 

”TV items
Tlural cla
“'0 0f the
1st by [s

dunbv
Tubs

 

 

78
There are only two items in this class:
miyya 'hundred'
nasmi ’helpful‘

3.2.1.16 FaMaaMuL Plural Class

 

gem flag; (singulars)

[strong. war] FvMMvL(a)

This plural class (FaMaaMuL) is an alternant of the
following plural class (FaMaaMiL). While the forms in
this class provide the right environment for the predic-
tion of the final vowel of the plural pattern as 1;. so do
many items associated with the plural FiMaaMiL. Thus these
plural classes are presented separately. There are only
two of the eight items in this class which are character-
ized by [strong. war] I

dumbuk 'tapering drum with skin head'
Qumbula 'bomb'

3-2.1.17 FaMaaMiL Plural Clags

Ell. Phon (singulars)
[8 tr ongl FvMMvL ( a )

FanvL ( a )
[place] FvavL ( a)
[instrument

All the S-classes combined ([strong], [place] . and
[instrument] ) account for a little more than 50% (8/15)
0f the total number of items in this plural class.

Here again strength can imply respect. numbers. high
9°81tion or actual strength. The strong] S—class accounts

1‘0? 1/10 of the total number of items.

 

 

host c
the prefix
vhstever is
formed. Fe
he. the j
is this 8-
lnto this
Thdri-oor
lsllaalliL,
The place
vevbership

Notie
items are
Other qua.
This mu
ll the pi
3T hderl
hop,
3.2.1.17.

The
'asl

 

79
Most of the items in the [place] S-class begin with

the prefix ma— which has the meaning 'the place at which'
whatever is indicated by the consonantal radicals is per-
formed. For example. from 359 'write' we get maktab 'office'.
i.e., the place where things are written. Some of the items
in this S-class do not begin with the prefix maq,but fall
into this class regardless because of their meaning and
quadri-consonantal pattern. (Cf. FaMaaLi. a subclass of
FaMaaMiL, for[place with a triconsonantal surface pattern.)
The [place S-class comprises 1/3 of the total plural class
membership.l

Notice that the instrumental prefix is m;- . Not all
items are so prefixed in the [instrumental] S-class. however.
Other quadri-consonantal instruments also fall into this class.
This group accounts for 1/10 of the total number of items
in the Plural class. FvMMvL(a) and FvavL(a) correspond to
an underlying FngMvL (+a), where the first M may be a 81°tt81

stop.

3.2.1.17.1 FvMMvLSa) Strong

'ungur 'race (nation)'

'askar 'army'

Jahfal 'group'

qunsul 'consul'

mixlif 'young male'

tantal 'giant'

Enfxara 'someone to be proud of'
Jandala 'giant'
*daktawr 'doctor'

*faylasuuf 'philosopher'

 

 

3I2|1|17.2

fasri
xaani
‘aali
Taani
taaii
qaafi
naabf
|leads

”Jul-Vs?

Tasi
'aru
habi
hadi
his
Tabi
'ali
hand

3:2,1|l7'

1181‘!
Map]
Taxi

3.2.1.17.2

3.2.1.17.3

3.2.1.17.“

faaris
xaanim
'aa?ila
?aanisa
taa?ifa
qaafila
naabiga
*?aadmi

fasiil

'aruus

habiiba
hadiira
katiiba
qabiila
'asiira
hamuula

marfa'
majlis
maxfar
mafrag
maqta'
maqla'
makbas
maq'ad
maktab
mal'ab
manbar
manjam
manza'

80

Fvanga) Strong

'knight'

0mm.

'family'

'young lady'
'religious sect'
'caravan'
'distinguished man'
'human'

FanvL(a) Strong

'platoon'

'bride'

'beloved (feminine)'
'patrol'

'battalion'

'tribe'

'family'

'stock (family)'

FvMMvLSa) Place

'stand supporting a clay water Jug'

'conference‘
'guard post'
'crossroads'
'cross-section'
'stone quarry'
'packing house'
'seat'

'office'
'athletic field'
'pulpit'

'mine‘
'dressing room'

 

 

 

 

Evtivta P1!

ranzil
nanfad
ravqi'
ranqal
ransab
tarsal
ragssl
na'mal
na'had
vs'rad
mat'ah
malted
ra'bar
rather
new
malral
dalla:
Taipei
helm.
matte?
masks
tasra
pasta
Tasha
Maths
Table
helm;
little
“at!
“561:
“Elbe

mark
Tarq

 

81

FvMMnga) Place. cont'd

manzil
manfad
mawqi'
manqal
mansab
mafsal
magsal
ma'mal
ma'had
ma'rad
mat'am
ma:bad
ma'bar
matbax
masna'
maérab
maslax
masqat
mashed
mastal
maskan
masrah
masbaé
masbah
mathaf
msbésl
maima'
mahlaj
maxraj
madxal
maxbaz
maxzan
markaz
margag

'inn'

'opening'
'place'
'brazier'
'position'
'Joint'

'sink'

'factory'
'institute'
'showroom'
'restaurant'
'temple'
'crossing place'
'kitchen'
'factory'

'camp site'
'slaughterhouse'
'place where something falls'
'scene'

'plant nursery'
'residence'
‘stage'

'place where date molasses is made'
'swimming pool'
'museum'
'drainage ditch'
'meeting place'
'cotton ginning plant'
'exit'
'entrance'
'bakery'
'storeroom'
'station'

'dance hall'

 

 

narqae
rahpai
parse
septa
dafta
Tatla
funds
'anbe
rand:
qultv
Jadvlv
math
rain
ready
magt
ralv

marqad
mahjar
marsad
maytam
daftar
?atlas
funduq
'anbar
xandag
qultug
Jadwal
mathana
magnaqa
masyaxa
magbara
malzama
mahfada
mamlaha
mamlaha
mantiqa
makhala
majsara
madrasa
madbaga
mazra'a
mahkama
mahbara
matfala
mazbala
daskara
sarnaqa

*karxaana

 

82

FvMMnga). cont'd

’mausoleum'

'stone quarry'
'observatory'
'orphanage'
'notebook'

'atlas'

'hotel'

'warehouse'

'moat'

'overstuffed chair'
'canal'

'grist mill'
'gallows'
'sheikhdom‘
'oemetery'
'section'

'folder'

'place where salt is obtained'
'kingdom'

'area'

'long-necked Jar for kohl'
'slaughterhouse'
'school'

'tannery'

'farm'

'court'

'inkwell'
‘ouspidor'

'trash heap'

'small piece of paper'
'cocoon'

'factory'

*tallaxanna 'casino'
*xastaxaana 'hospital'

 

 

 

3.2.1.17-5
saahi
tashi
ism
teari
qaa'i
*nllnr
hhl.l7.t

rake
rail
rasi
null
mall
dari
vans
sahi
use
ziri

mu

 

 

83
3.2.1.17.5 Fvanga) Place
saahil 'seashore'
taabiq 'floor'
3aami' 'mosque'
seari' 'street'
qaa'ida 'pedestal'
*miinaa? 'port'
3.2.1.17.6 FanvLSa) Place
makaan 'place'
madiiq 'narrow passage'
masiif 'summer resort'
muéiif 'hostel'
mabiid 'ovary'
dariih 'tomb'
manaara 'minaret'
sahiifa 'page'
kaniisa ‘church'
ziriiba 'pen, corral'
Jaziira 'island'
siriifa 'hut'
tahaara 'toilet'

hadiiqa 'garden'

3.2.1.l?.? FvMMnga) Instrument

misjab 'gun rack'
minJal 'sickle'
mizraf 'awl'
mijhar 'microscope'
mihwar 'pivot'
mirJal 'boiler'
mixlab ‘claw'
mugzal 'spindle'
madfa' 'gun'
pardam 'rivet'
XanJar 'daESer'

 

84
FvMMnga) Instrument. cont'd

 

sarbas 'reel'

sullam 'scale. ladder'
munxul 'sifter'

warwar 'pistol'

randaj ’carpenter's plane'
mijrafa 'kind of shovel'
mitfa?a ‘fire extinguisher'
mihraqa 'incinerator'
mizwala 'sundial'

manfa'a 'use'

silsila 'chain'

sunbula 'spike‘

matba' a 'printing press'
majrasa 'grinder'
*bunduqiyya 'rifle'

3.2.1.17.8 Fvanga) Instgument

haajiz 'divider'

faariza 'comma'

*mizlaag 'sliding bolt'
*mihgaan 'funnel'
makiina 'machine'
hadiija 'camel saddle'
bariina 'drill'
“'arabaana 'cart'

 

3.2.1.18 ?aFMiL§§? Plural Cl§§§
§e_m Phon (singulars)
[strong. FaMi

respected] FaMiiL
Three-fourths of the items in this class belong to

the [strong. respected] S-class.

 

p.2.l.18.l it
real
hall
gani
Tart
daki
pawl
vafi

31.1%.? E

sadiiq
tabiib

33.1.19 Fee

—-

Sfl

beak]

Both its
Threaten”
ll alternant
they bOth do
it 13 inter-e
bless are Ch

Thu. the me

ti [leak] Vi

Plural clear

have;
‘atng

 

85

3.2.1.18.l FaM;

wasi 'guardian'
nabi 'prophet'
gani 'prosperous’
Cari 'wealthy'
daki 'intelligent'
qawi 'story'

wafi 'loyal'

3020101892 FEM1;L

sadiiq 'friend'
tabiib 'doctor'

3.2.1.19 FlMMaL Plural Class

 

§gg Phon (singulars)
[weak] FvMiiL

Both items in this class have a common semantic and
characteristic P-shape. This plural 01353 (FiMMaL) is
an alternant to the following plural class (FuMMaL). While
they both contain items characterized by the feature [weak:l,
it is interesting to note that the members of this plural
class are characterized as [weak] via old and worthless,
“file the members of the FuMMaL plural class are characterized
as [weak] via young and childlike. The two members of this
plural class are:

kidiig 'worthless horse'
'atilg 'old'

 

 

Mlluzo F__u
§_eg

beak,
child]

This p]
above (3.2.1

characteris1

3.2.1.20ll

ragil'
Sa§11m
futilr
waiiih

3:2.1203

qaafir
saaéij
’7a‘zal

 

86

3.2.1.20 FuMMaL Plural Class

 

Sem Phon (singulars)

Bleak, FvMiiL
childlike] FaaMiL

This plural class is an alternant of the FiMMaL class
above (3.2.1.19). All the members of this class are

characteristic of children. (Cf. also the noteunder 3.2.1.19).

3.2.1.20.l FvMiiL

radii' 'suckling'
Basiim 'inexperienced'
futiir 'foolish'

waki 11.1 ' brash '

3.2.1.20.2 FaaMiL

qaasir 'legally minor'
saadij 'simple'
*?a'zal 'defenseless'
*rahhaal 'migratory'

30201.21 FlMlL Plural C183§
§gm Phon (Singulars)

[fieak. low] FMva
[paler] ?aFMaL

This Plural class (FiMiL) is related to the following
plural class FuMuL (3.2.1.22). However. since the 1/Uu
alternation is not entirely phonologically predictable.
these classes are presented separately.

Five out of the ten items in this plural class are
color-derect adjectives. The defect adjectives in this

“£33 have been combined with slipper and sandals because

 

 

the shoe, espel
contempt. To
race by crossi

of insults.

3.2.1.21.1 [M

mdaas
m'aal

3.2.1.212 g

7am!
?ahdab

312.131.} ?!
hbyad
?amlah
Tadlam

”.132 NM

‘

Se

\

[heals
defem

[color]

This is
fiche Arab 9
‘fran 1/2 of t
ifmi‘les of
M the se
“Maw
:tnnological.

\30 the notl

 

8?

the shoe. especially the bottom of the shoe, is held in
contempt. To show the bottom of one's shoes to another's
face by crossing one's legs on a desk is one of the highest

of insults.

3.2.1.21.l FMWL Weak

 

mdaas ' slipper '
m'aal 'pair of sandals'

3.2 .l.21.2 ?aFMaL Weak

?a'ma§ 'myOpic'
?ahdab ' hunchback '

 

3020102133 ?8FM&L CO;OI

?abyad ' white '
7amlah ' grey '
?adlam ' dark '

3.2.1.22 FuMuL Plural Class

 

.3331. Phon (singulars)

Beak. 7aFMaL
defective]

[color]

This is a plural class which was early recognized
by the Arab grammarians to be semantically defined. More
than 1/2 of the items in this class (7/12) are 80-0811“
81Meeting of color (l/U) or defect (1/3). and they are
all 0f the same canonical form. There does not appear to
be any Predominant characteristic, either semantic or
ph"10105516111. for the remaining members of the class. (Cf.

a‘180 the note under 3.2.1.21.)

 

Mnllzz 01 ?

7am:
'famlat
Wm
We?
h‘waj
7a'dab
7a'ta1
hm;
?a;1a'
him
hmad
?arga§
hms
hhual
him
“3an
Mm

32.1.22 ‘2

'fafimm
738%]
We:
?5§ha‘
“he?
M83
him
7&3!»
fine
M}.
We

 

88

3.2.1.22.1 7aFMaL Weak

 

?anwaJ 'foolish'

?amlat 'hairless'

?a'war 'one-eyed'

?a'was 'squinting'

?a'wa3 'bent'

?a'dab 'paralyzed in one hand or arm'
?a';a1 'defenseless'
?atra§ 'deaf'

?as1a' 'bald'

7agram 'having a harelip'
?armad 'sore-eyed'

?arga§ 'pockmarked'
?axras ‘mute'

?ahwal 'cross-eyed'
?ahmaq 'dumb'

?acqa1 'cross-eyed'
?a01al 'confused'

3020102202 ?aFMaL COlor

 

?ahmar 'red'
7agbar 'dusty'
?asfar 'yellow'
?a§hab ' 115m grey'
?asrah 'inflamed‘
?a§gar 'blond'
?aswad 'black'
?asmar 'brown'
7axdar 'green'
?abras 'albino'
?abga: 'speckled'

 

The P
to an unde
nlottal st

The n
eluded in
e.n., La};
e.g., 31.1.?
included ‘
fun the w

The
CbStrless
account 1

flats ..

33.1.23
Xfirn
gab
X18

 

89
3.2.1.23 FiMLaan Plural Class
§gm ghgg (singulars)
[weak. low] Fva
[close] gzggév)
FvavL

The P-classes Fva. Fva(v) and FvMLv all correspond
to an underlying FngL §+suffix). where M or L may equal
glottal stop.

The members of the [fieak, low] class have been in-
cluded in Section 3.2.1.23 for various reasons: position.
e.g., valle , ground; defective or not worthy of respect,

e.g., blind, peasant; common. e.g., crow. Other animals

 

included here are rather weak and usually serve as prey
for the warring predatory animals, e.g.. sheen. gazelle.

The other S-class. [close:). is a minor class designating
dbseness of personal relationship. The items in both S-classes
account for about 5/8 of the total items in this plural

class -- weak, 1/2; and close. 1/8-

3.2.1.23.l Fva Weak

 

xara 'feces'
sabi 'youth'
xisi 'eunuch'
tili 'lamb'

3.2.1.23.2 Fvav Weak

faar 'mouse'
fiaa' ’ground'
saaq 'shank'
Zaag 'crow'

 

 

nnnnn Weak, <
raa'i
naadi
*kraa'

*graab

n2.1.23.3 B

ta‘mi
xusna
*n'hydi
”mas

3.2.1.23.n n

W 111‘
xaruuf
Gazaal
*filis

)I2.1.23.5
hams

)I2.1.23'6

Jaar

3.24%.?

fifii:
fiadiic
111811]

3.2.1.21. .1
§g§
"881;

 

 

 

9O

Fva‘v) Weak. cont'd

raa'i 'shepherd'

waadi 'valley'

*kraa' 'lower leg and foot of cow or sheep'

*graab 'crow'
3.2.1.23.3 FvMLv Weak

7a'mi 'blind'

xuswa 'testiCIe'

*m'aydi 'peasant'

*7axras 'mute'
302010230“ FVMWL Weak

bi'lir 'camel'

xaruuf 'sheep'

gazaal 'gazelle'

*filis 'one file (coin)'
3020102305 FVMV Close

hama 'father-in-law'

 

3.2.1.23.6 Fvagv) Close

Jaar 'neighbor'

3.2.1.23.7 FvavL Close

 

fifiij 'friend'
sadiiq 'friend'
nisiib 'in-law'

3.2.1.2# FaMaLa Plural Clasg
§gm Phon (singulars)

[weak . low] FaaMiL
Seven out of the ten total items in this plural class

are characterized by [weak. low] . They are all people held

 

n Ion esteem.

n2.1.2n.1 as

Win
seam
faajir
qaatil
ham
Jaahil
*‘ajuuz

3.2.1.25 1311.1
SE
beak, 101
This is a
””31 (or re,
uslane to the

Chmcterintnc

 

91
in low esteem.
30201.2“.1 FaaMiL
xaa?1n 'traitor'
saafil 'low'
faajir 'libertine'
qaatil 'killer'
kaafir 'infidel'
jaahil 'ignorant'
*‘ajuuz 'old'

3.2.1.25 FaMiiL Plural Clagg
gem Phon (singulars)

[weak. low] ?

This is a notable plural class because there is even
a local (or regional) song in Iraq equating the dancing of
a slave to the donkey's kicking. There is no common P-
characteristic. The two members of this class are:

hmaar 'donkey'
'abid 'slave'

3.2JJ26 FMaaMLa Plural Class
§gm Phon (singulars)

[weak, low] FMaML {i i
at

 

These items account for 2/5 of the total membership of
this plural class:

mdarrat 'low'
mgaddi 'beggar'

 

 

3.2.1.27 .FJ
§§l
[body
[crack
[death

[elong
film

The 1‘
hornet. ho
for about
that 1/2 c
associated
Tented p:
feTale-nos:
on tracks
have “1181‘
leuntlc

The
than is
italla.
out. .o
“9 all
addition
hog 5~o
maths)
T1ling ]

Tho

F" —‘ w 'v

92

3.2.1.27 FMuuL Plural Class

Sem Phon (singulars)

[body part] FvaL
[crack, hole]FVHL(V)
heath]
Elongated]
Elant]

gminor S-classes

The items in all of these S-classes ([hody part] .
[crack, hole], [death], [elongated]. and [plant]) account
for about 3/4 of the items in this plural class. Notice
that 1/2 of the body parts are either directly female-
associated, e.g., belly. which is always one of the exag—
gerated parts of a fertility goddess. or indirectly
female-associated -- characterized by holes (e.g.. page)
or cracks (e.g.. double chin). The other body parts may
have migrated into this plural class on the basis of the

semantic characteristic [body part] and/or P-shape.

 

The second semantic feature characteristic of this
class is rather interestingly associable with female gen-

italia. Note that there are three words for hole. two for

 

nggk. one for £22 and one for §2ll_° The other items
are all associable with this general semantic notion. An
additional very interesting fact is that the animals in
this Seclass characteristically appear as Jagged lines
(cracks) in the sky to the naked eye. They are all high
flying birds.

There are two minor S-classes. [death] and [elongated] ,

cmnained in this plural class. The items under [death]

 

seem to have
ts tnteresti:
cracks, ho

to our angel
In good or t
Most 01

as [round] t
One in‘

that the pe
m be clas
This is not
It may be 1

alternate I

See a?

class (3,2

33,137.].

bath
Juhtn
“'11
Mi]
{1311;
:3ng
$3M
fart
ltlgu
nah}
Hr!
kart

93

seem to have to do with a concern for the afterlife. It
is interesting to note that geeze may be grouped with
[cracks, holes:lor with [death] . The gee;e here is akin
to our angels and demons (depending on whether the genie
is good or bad).

Most of the items under elongated] could be classed
as [round:].

One interesting thing about the [plant] class is
that the people from the rural areas of Southern Iraq
may be classed here. These people are mainly farmers.
This is not the only plural class this item fits. however.
It may be Just an accident of P-shape in this case. Its
alternate plural is discussed under the section Synchgoeic

Evidence.

 

See also the note under the related FMuuLa plural

class (3.2.1.28).

3.2.1.27.1 FvaL Body Part

batin 'belly'

juhur 'anus'

Ji'ib 'hips'

Jifin 'eyelid'
halig 'mouth'

xasim 'nose'

dari' 'udder'

farij 'vulva'

lugud 'double chin'
nahid 'female breast'
wirik 'hip'

karis 'potbelly'

 

 

1
htvL Body Par

dani'
girif
fang
hm!
311m
Jimi'
zanid
sadir
3111'
that
(the
'unig
galub
matln
girin
larig

3'2'1'27I2 E

C1(Its
5am
(arr
raw
‘ayn
dan
xadd

3I2|l.27.3

“3811b
matheld
8131:
3am
tint;
81hr

94
FvaL Body Part, cont'd

dami‘ ‘tear(s)'

girif 'hoof of a butchered animal'
farig 'part in one's hair'
hini! 'chin'

jilid ' skin'

Jimi' 'fist'

zanid 'upper arm'

sadir 'chest'

éili' 'rib'

dahar 'back'

'ikis 'elbow'

'unig 'neck'

galub 'heart'

matin 'shoulder'

girin ’animal horn'

darig 'bird dropping'

3.2.1.27.2 FvMLgv) Bod: Part

 

days 'breast'

samm 'handful'

caff 'palm of the hand'
raéd 'jaw'

'ayn 'eye'

dayl 'tail'

xadd 'cheek'

3020102703 FVMVL Czacke

nagub 'perforation'

mahad 'eradle'

sisir 'peel'

garid 'hole'

gidir 'small cooking pot'

gihif 'piece of broken pottery'

 

MWLCracks.

gabur
tam
(aruf
turuf
dtri'
xurum
tall:
tarih
banid
fali'
fatir
(arig
saguy
*naasuur

”MM I

M
daff
tazz
gibb
Eagg
§att
ta};
ba§§a

batta

”.1376

Sabur
“3'13
nllir
ianib
‘V

85181hit

 

95
FvaL Cracks. cont'd
gabur 'grave'
tasit 'wash basin'
daruf 'small leather bag'
suruf 'hole'
diri' 'shield'
xurum 'hole‘
jalix 'scrape'
jarih 'wound'
banid 'bonnet'
fali' 'crack'
fatir 'crack'
sarig 'split'
sagur 'falcon'

*naasuur 'eagle'

302o1-27.4 FvML§v2 CrackgI HolesI ate.

 

'15; 'nest'

daff 'tambourine'

hazz 'nick'

éibb 'large porous pottery vessel'
fiass 'r1p'

satt 'river'

E8§T 'bird'

bassa 'duck' «’V“~

batta 'duck'

3.2.1.27.5 FvaL Death

 

gabur 'grave' (given under holes. crackS)
na'is 'bier'

nidir 'sacrificial offering'

danib 'offense'

‘gaahid 'witness'

 

3.2.1.27.6

rabb
qiss
tats
Jtdd
'ayb
tinni

3'2|1027l7
satir
Earid

3'2|1|27.E

saff
tayl
batt

3.2.1.?“

Em
hiri
tidi
tau
fart

96

3.2.1.27.6 FvML‘v) Death

rabb 'lord'

qiss 'clergyman'

hadd 'fate. lot'

Jidd 'grandfather'

'ayb 'fault'

Jinni 'genie'

3.2.1.27.7 FvaL Elongated
satir 'line'
carid 'cigar'

3.2.1.27.8 FvML v Elon ated

saff 'row'
tayl 'wire'
batt 'thin bracelet'

3.2.1.2799 FVMVL Plant

 

sitil 'young plant'

hiris 'plant'

Jidi' 'tree trunk'

jadir 'root'

fari' 'twig'

'irig 'root'
*sruugi 'person from rural areas of Southern Iraq'

3.2.1.28 FMuuLa Plural Claee

Sem Phon (singularS)
[malel FvaL
FvML

This is another very interesting class for several
reasons. (1) the characteristic P-shape for the plural
°f this class is similar to that of the previous plural

Class:

 

Fhu --
FMuula --

(2) too out of
very clearly I
or conceivablt

The sing
an underlying

tyotlo vowel

33.1%.]. 1

rajll
nasal
butll

”.1283
'am
zibb
‘ayr
dayl

am
out
*hilz

H.129 1
Sam

[male

The 1

0T adage“

(hm is E

 

97

FMuuL -- female genitalia
FMuuLa -- male genitalia:

(2) ten out of the 1h items in this plural class are
very clearly either finale] . male-associated body parts.
or conceivable phallic symbols. e.g.. Leil.

The singular P-shapes FvaL and FvML correspond to
an underlying FvM¢L. where fl may be realized as an anap-

tyctic vowel in case the final consonants are dissimilar.

3020102801 FVMVL

rajil 'husband'
nagal 'bastard'
butil 'bottle'

3.2.1.28.2 FvML

'amm 'paternal uncle'

zibb 'penis'

'ayr ‘penis'

dayl 'tail'

sayr 'leather strap'

*diid ‘rooster' (more abstractly diyc)

*hiiz 'epithet for a man who submits to sodomy

for a price' (more abstractly hiyz)

30201.29 FuMLaan Plural Class

 

§2§ Phen (singulars)
[male] FvavL
Fva(v)L
The items here are either male. male body parts.
or adjectives descriptive of masculinity. 9080' 22229-
3.2.1.28)

There 18 another [male] plural (Of. FMUVLB'

 

but its it
the aemant

or the tot

3.2.1.29J
(qu
qadi:
(an
“urb
*‘adu

3'2DL29.

fan
V
saat
*raat

3|2o1o30

or
o
a

I

’U'U
W

[Fla

The
some 0
semantic
(”Anne

The
Datum

mteIn.

teens t1

 

98
but its items have a different set of canonical forms.

The semantic characteristic [maleJaccounts for about 1/2

of the total number of items in this class.

30201029.]. FVMWL

 

sujaa' 'brave'
qadiib 'penis'
sariik 'partner'
*‘urbi 'tribesman'
* 0 ad“ 0 enemy o

3.2.1.29.2 Fva‘sz
faaris 'knight'
saabb 'young man'
*raahiib 'monk'

3.2.1.30 FVIMaaLaat Plural Class

 

Sem Phon (singulars)
[paired body FvaL

parts] FvMLa

[plants]

The items here have no other plural and comprise the
entire class. Note the closely related P-shape and
semantic characteristic of the following plural class
(FMaaLaat, 3.2.1.31).

The symbol V indicates the first vowel in the plural

Pattern is the same as the first vowel of the singular
Pattern.

An interesting fact is that the S-class [plant] always
seems to be associated with the same plural clasies) as the

S-c1ass [body part] .

 

 

3.2.1.30.1

mm
Tubut

t2d.30.2

nabta

3.2.1.31
Sen

[pan
dreo

Them

hm:
ridlt

See.

above.

33.1.32

(0
o

I

a

r——1
[D
2’

The
T in th
The
compo

H'e'l Ti

 

99
3.2.1.30.l FvaL Paired Body Part

?idin 'ear'
?ubut 'armpit'

3.2.1.30.2 FvMLa Plant

nabta 'plant'

3.2.1.31 FMaaLaat Plural Clase

§ee Phon (singulars)

[paired bod / FiMiL
dress part3

These are the only two items in this class:

Xitif 'shoulder'
ridin ’sleeve'

See also the related plural class FVIMaaLaat (3.2.1.30)

above.

3.2.1.32 ?aFMaaL Plural Claes

§££ Phon (singulars)
[abstract] FvaL

FvML

Fva

Fva

FvavL

The items here comprise about 1/3 of the items enter-
Ed in this plural class.

The singular P-shapes FvaL, FvML. FVMVo and FVVL
correspond to an underlying FvMflL. where M or L may be weak
(i.e., W. y or 7) and fl may be realized as an anaptyctic
VOWGI- See also section 3.2.1.12 for a definition 0f the

feature [abstract] .

 

t2d.)2.1 T

naqlt
wazin
uuja'
nadar
nadll
nuluk
tofu?
qadar
fahim
gored
‘umuq
'umur
'tris
'atab
'ugur
'adao
taraf
Tam
olnlf
huh
Tabah
sudu;
subu'
{W
Rabat)
m(loin
Tiliq
I‘llbu!
mm
mm}
xThus
Xllluq
ISitar
“To?

3.2.1.32.l

waqit
wazin
qua'
nadar
nadil
muluk
lafud
qadar
fahim
garad
'umud
'umur
'iris
'atab
'ugur
'adad
taraf
darar
sinif
sibih
sabah
sudus
subu'
di'if
sabab
raqam
riziq
rubu'
ribih
dihin
xumus
xuluq
xatar
rats?

100

FvaL

'time'
'weight'
'pain'
'consideratiOn'
'low'
'property'
'kindness'
'fate'
'understanding'
'intention'
'depth'
'life'
'marriage'
'damage'
'1/10°
'number'
'extremity'
'harm'
'kind, sort'
'resemblance'
'apparition'
I1/6o

01/70
'double'
'reason'
'number'
'livelihood'
'quarter'
'profit'
'mind'

01/5!
'character'
'danger'
'mistake'

 

 

w. cont

hillm
hmr
hukum
hasab
haram
hajlm
Jlnia
Juzul

Tao

FvaL, cont'd

hilim
hilif
hukum
hasab
haram
hajim
sinis
Juzu?
Oumun
Jurum
Osman
quul
tusu'
01110
ta‘ab
bu'ud
?alam
?ama1
bahiG
waham

101

'dream'
'swearing'
'judgment'
'esteem'
'forbidden’
'bulk'
'kind'
'part'
01/80
'crime'
‘price'
'weight'
01/99
01/30
'trouble'
'distance'
'pain‘
'hope'
'research'
'delusion'

3.2.1.3202 FVML

naw'
lawn
tayf
gawq
didd
hurr
Jilf

'kind'
'color'
'vision'
'longing'
'opposite'
'pure'
'boorish'

 

 

3|2|1l32 I
rah

3.2.1.32.

me]
not
om'
haai
3112

3.2.1.32

 

102

3.2.1.32.3 Fva
ra?i 'opinion'

3.2.1.32.4 Fva

maal 'property'
ruuh 'soul'

diin 'religion'
haal 'condition’
3111 'generation'

3020103295 FVMWL

sariif 'distinguished'
sabiih 'like'

habiib 'beloved'
barii? 'innocent'
*sirriir 'very bad'

3.2.1.33 FaMaaLL Plural Class

 

§£E Phon (singulars)

 

[abstract] Fva(L)v
All the items in this class are listed here:

xaassa 'property'
haasa 'sensation'
'aami 'common man‘
*madarra 'harm'

(Cf. also the section above 3.2.1.32 and 3°°t1°n
3.201012 .)

3.2.1.34 FaMaLaat Plural C1883

 

Sem Phon (singulars)
[abstract] FvMLa
There are only two members in this class both Of

which are abstracts

 

 

xldma
Jabha

3.2.1.35 }
.Sfl
[agen
Two-t
to the sen
hooters p1
11km.
character!
for that a
this plum
tton 1N“
alternant

Thins.
T“!
one
say
my
has
“at;

3'2'166
Sam
[at
All

Mantle

PThalia

 

103

xidma 'a service'
Jabha 'front'

3.2.1.35 FiMMaaL Plugal Claee

Sem Phon (singulars)
[agent] FaaMiL

Two-thirds of the members of this plural class belong
to the semantic class [agent] . The additional three
members probably came into the class on the analogy of

pilgrim. They are barefooted. fastin . and ele. The

 

characteristic P-shape of the singular is the same as

for that of the following plural c1ass(3.2.1.36). to which
th18 Plural class is related. Since the vowel alterna-
tion inJu is not completely predictable. however. these
alternant classes will be presented separately.

3.2.1.35.1 FaaMiL

saayig 'goldsmith'
saani' 'apprentice'
saayis 'stableman'
saayih 'traveler'
haayik 'weaver'
*hajji 'pilgrim'

3.2.1.36 FuMMaaL Plugal Class

 

Sem Phon (singulars)
[agent] FaaMiL

All of the 21 members of this class fall into the
semantic category [égent] . This plural class has the
P‘Bhape characterizing the singular forms. FaaMiL. 38

does the plural class FiMMaaL (above). which is also an

 

{agent} clai

3.2.1.36.1

qealid
{oath
heatit
team
naaqit
naali
toaa' 1
tall
daabt
an
oaay
may
real
haal

[agent] class.

104

3.2.1.36.1 FaaMiL

 

Qaa?id
faajir
kaatib
kaafir
naaqid
naa?ib
waa'id
taslib
daabut
saayih
saayis
zaayir
raakib
haakim
taajir
Oaayir
haaris
haafié
*hakam.
*?a'zab

'commander'
'libertine'

'writer'

'infidel'

'critic'
'representative'
'preacher'

'student'

'officer'

'traveler'

'driver'

'visitor'

'passenger'

'governor'

'merchant'

'rebel'

'watchman' ‘
'man who knows the Koran by heart'
'umpire'

'unmarried'

3-2-l.37 FuMaLaa? Plural Class

M

Sem

[agent]
[strong]

[weak:]

Here again we have a plural cl

mainly by the semantic feature [agent] .

Phoe (singulars)
FaMilL

ass characterized

Its character-

istic P-Shape is FaMiiL. This class also has two other

major categories. [strong] and [Yank] '

Together these

 

 

account 1

agent. It

3.2.1.37

*-
MW

105

account for the entire membership of the plural class --

agent, 1/2; strong, 1/6; weak. 1/3.

3.2.1.37.1 FaMil; Agent

waziir 'minister'
naqiib 'union leader'
kafiil 'co—signer'
faqiih ’theologian'
'amiil 'agent'
'amiid 'dean'
'aqiid 'colonel'
'adiid 'supporter'
'ariif 'sergeant'
'adiil 'the husband of one's wife's sister'
sahiid 'martyr'
safii' 'mediator'
*saa'ir 'poet'
sariik 'partner'
raqiib 'censor’
safiir 'ambassador'
?adiib 'writer'
?amiir 'emir'
7amiin 'trustee'
xabiir 'expert
mudiir 'director'
ra?iis 'leader'
za'iim 'leader’
wakiil 'deputy'
*xaliifa 'caliph'
*‘aalim 'scholar'

3020103702 F8M11L Stzolm

zamiil 'friend'
sariif 'distinguished'
fariiq 'team'

 

tahiiL Sta
'eeif
tori:
nail
aaii

Mil-.370

bexi
begi
raft
sax!
half
ta"
Jab
xab
set
93:
qac'
fac
la‘
lac
"at

“.13

TKiln 1
Ti ch:
“Till:

 

106
FaMiiL Stroeg. cont'd

'adiim 'great'
kariim 'generous'
najiib 'noble'
wajiih 'notable'

3020103703 F8M11L Weak

 

baxiil 'stingy person'
basiit 'simple'
radii' 'suckling'
saxiif 'fool'

baliid 'stupid'
ta'iis 'miserable'
Jabaan 'coward'
xabiiO 'troublesome'
safiih 'foolish'
eariib 'strange‘
qadiim 'old'

faqiir 'poor'

la?iim 'evil'

wadii‘ 'lowly'
*?ahmaq 'dumb'

3.2.1.38 FuMaet Plgga; Clase

§£E ghee (singulars)
[agent] FaaMi

The examples comprise the entire plural class.
Again there are other plural classes characterized by [agent] ,
but this one has a unique characteristic P-shape for the

31n8ulare FaaMi.2

 

3.2.1.38.
teat

teas
oaac

than

a:

e

.1
at
s

T
i

 

 

 

 

 

107
3.2.1.38.l FaaMi

Jaani 'culprit'
Jaabi 'collector'
taagi 'tyrant'
saaqi 'waiter'
raa'i 'shepherd'
haami 'protector'
waasi 'informer'
haawi 'amateur'

*daahiya 'genius'

3.2.1.39 FaMaaMiLa Plural Class

 

 

 

Sem Phon (singulars)
[agent] FvMMva(v)
[s tr on 8] FvMMva

The examples listed here comprise the entire class.

3.2.1.39.1 FvMMvajv) Agent
qursaan 'pirate'
baytaar 'veterinarian'
bayruuti 'Beiruti'
*faylasuuf 'philosopher'

3.2.1.3902 FVMMVLV Agent
saydali 'pharmacist'
*qiss 'clergyman'

3.2.1.39.3 FvMMvagv) Strong

‘imlaaq 'gigantic'
jabbaar 'strong'

 

3.2.1.39c

plural
in the
form For
I therel
t c
a Thai
[More
hitch
mom
i cache
The 5.

"Tibet

 

108

3.2.1.39.# FvMMva Stroeg
'abqari 'ingenious'
*malak 'angel'

302 .IJJ-O Fame-Li Plural 0188§

 

§ee Phon (singulars)
[place] FvML(v)
[time]-—minor class

[weak] FvaL

This plural class (FaMaaLi) is a subclass of the
plural class FaMaaMiL(3.2.l.l7). The singular items
in the S-Class [place] .fbr example, have the surface
form FvMLv, which corresponds to an underlying FngMQL,
Where L is a weak radical (i.e., i or ?).

Two-fifths of the total number of items in this
Plural class are accounted for by the semantic classes
[Place] . [time) and [weak] . The items of [place]
which fall into this plural class have a final weak
radical consonant, while those having a final strong
radical consonant fall into the FaMaahiL plural 01888-

The semantic feature [place] accounts for l/h Of the total
number of items in this plural class.

[time] is a small subclass (1/20) 0f the t°t31)°

2&223 'tough person' appears under the heading [weak:l
Since such a person is in general held in low esteem.

This S-class accounts for 1/10 Of the t°t31 number °f

items.

 

 

 

 

hm

109

3.2.1.4001 FVMstz Place

ma?wa
masts
mar'a
mabga
mansa
manfa
mamsa
malha
makwi
maqha
'alwa
masfa
*naadi
*daahya
*sahraa?

'shelter'
'winter resort'
'meadow'
'brothel'

'place of origin'
'place of exile'
'walkway'
'nightclub'
'cleaner's'
'coffee house'
'a high place'
'refinery'
'club'

‘suburb'
'desert'

3.2.1.l¥0.2 FVML‘VE Time

layla
*Oaaniya

'night‘
'second time'

302.1.“003 FVMVL Time

 

duhur

'noon'

3.2.l.(+0.‘+ FVML v Weak

balwa

ma?sa

sakwa
*?aafa
*ma'siya
*daahiya

'affliction'
'tragedy'
'complaint'
'tough person'
'sin'
'disaster'

 

 

[cover]

These Sr
in this

and [in
Plural
3'2 I]. all

110

3.2.l.#1 7aFMiLaA2?aFiMLa Plural Class

 

Sem

[cover]
[statement]
[instrument]

Phoe (singulars)

gminor classes
F(v)Mva

This plural class contains two minor S-classes --
[cover] and [statement] and one major S-class [instrument] .
These S-classes account for more than 1/2 of the 31 items
in this plural class -- [cover]. 1/3; [statement]. 1/8

and [instrument], 1/5. ?aFiMLa occurs as an alternant
plural for those items having same final consonants.

 

3.2.1.hl.l F§v2Mva Cove;
hidaa? 'shoe'
gisaa? 'membrane'
wariid 'vein'
rasiif 'pavement’
qinaa' 'mask'
gilaaf 'covering‘
qmaas 'cloth'
*gata 'cover'
janiin 'embryo'
3.2.1.41.2 Fiv2Mva S§atement
su?aa1 'question'
Jawaab 'answer'
daliil 'proof'
*du'a 'prayer'
3.2.1.b1.3 ngzhva Instrument
sariit 'ribbon'
Jihaa; 'apparatus'
damaag 'brain'
lsaan 'tongue'
slaah ‘weapon'
Jnaah 'wing'

 

3.2.1.1t2
See ‘
defining
either to
relations

carter of

in) 32.
The:
to expla
broken 1)
It
Sound an
and that
Tliral a
White;
broken
Semitic
Ta
Thloh a
Tilr c
hum
The a
sound I
Totem

TTirol

111
3.2.1.42 Queetionable Plural Classee
See the Appendix for a list of plural classes whose
defining semantic characteristic(s) are questionable due
either to the sparsity of items with a strong semantic
relationship comprising the given plural class or to the

number of semantic characteristics needed to define it.

3.3 The Chan e n Semantic Class ficatio

There are actually at least three hypotheses put forth
to explain the origins and development of the sound and
broken plural systems.

It is the consensus among Semiticists that both the
sound and broken systems were present in the Proto language
and that Arabic has made a much greater use of the broken
plural system than have the other Semitic languages. This
hypothesis in effect makes the claim that the sound and
broken systems both have a history dating back to Proto-
Semitic.

Talmy Givén3 takes the position that the broken plurals,
which today appear as infixes. arose from an earlier
affix system. That is. the triconsonantal root (with
infixed vowels) CCC arose from an earlier prefix pattern
*C-CC or suffix pattern *CC—C. Thus he claims that the
sound plural system is the older and that the broken plural
SYStem developed from it. This hypothesis deals with
Phonological assumptions. The present work. however, deals
with semantic assumptions.

Regarding semantics Giv6n goes on to state that if
"human nouns are not in a class by themselves but rather

are dispersed in many

 

 

classes och
holds. th
older cyst
preceding
how
this. To
cream
The nulti
urinals c
fields. a
female
iloates
The
to That
since 11
anthropt
corresp
that th
it me
“Godot
tenant:
tothe
T‘
Tatl
Tllent
Tina

Ti ct

112

classes which are grouped on the basis of other semantic
fields. then {one} may rightfully conclude about the
older system that it was not anthropocentric.“ See the
preceding Semantic Taxonomy and the following Multiple
Plural Associat;on sections for numerous examples of
this. For example, (cf. p.115) knight has two plurals,
one associated with etrepgth and one with masculinity.
The multiplicity of noun classes and the fact that the
animals are distributed on the basis of various semantic
fields, such as [strong. war] a lion. ti e . leoeazd and
[female genitalia, cracks]: eeglfi. saloon, duck. in-

 

 

dicates that the system was n—ary and non-hierarchical.

The hypothesis upon which the present work is based
is that the broken plural system is the older system
since it gives evidence of a non-hierarchical, n-ary. non-
anthropocentric semantic classification in one-to-one
correspondence with morphological plural classes; and
that the sound plural system is the newer system since
it gives evidence of being created for the purpose of
accomodating a newer hierarchical, binary, anthropccentric
semantic classification. This is a parallel hypothesis
to the one Givén (1971) made for Bantu.

This last hypothesis implies a corollary hypothesis:
that languages like Akkadian and Hebrew split off from the
parent language (Proto-Semitic) at a much earlier date
than did Arabic, and thus have had much more time to effect

the change to the newer semantic system. The fact that

 

 

krabic all)?“1
and verbal 8‘
methane a
David G
mherein de
Givén’s pre:
fixed systet
atruction.
Which:
crucial to
the reader
When viewi!

Chapter an

M Thai

‘
3.1m h
The 1
Pluraliza
Perhaps s
hierarchi
form 81V!
Dondence
“Wit:
“as $880

pl‘lrals

U

113
Arabic appears to be conservative in its phonological
and verbal systems, as well. lends support to this last
hypothesis and its corollary.

David G. Lockwood” suggested that perhaps the hypothe-
gm herein developed accounts for the Proto system and that
Givén's prefix-suffix hypothesis for the origin of the in-
fixed system is representative of an even earlier recon-
struction. that of Pre-Semitic.

Whichever hypothesis is ultimately borne out is not
crucial to the outcome of this synchronic analysis. However.
the reader should keep in mind the various hypotheses
when viewing the synchronic evidence presented in this

Chapter and the synchronic analysis presented in Chapter IV.

The News Mode -- A Conflation
3.4.1 Diachgonic Hypothgsig

The diachronic hypothesis for the Iraqi Arabic
pluralization system (with implications for Arabic. and
Perhaps Semitic. in general) is that the old n-ary, non-
hierarchical. non-anthrOpocentric system was of the general
form given in Figure 12. There was a one-to—one corres-
pondence between a given semantic category and its mor-
Dhological plural class. That is. each semantic category
was associated with a particular plural (or with several

plurals in complementary distribution by phonological shape).

 

 

 

 

[lent]
[instrument]
[but

{

Non-1

At fir
[human]
M 1a, 11
Memo“

“9| ea po

 

114

[noun]

[agent]
[instrument] [other body Inn-,3]
[location] [male genitalia]

[plant s] [abstract] [strong] [female genitalia]

[col ors] [round] [weak]

 

Figure 12
Non-Hierarchical Semantic Categories

in Iraqi Arabic
At first it was thought5 that [agent]. [instrument] and
[location] were classes which dealt with deep level case.
That is, it was thought that they were involved with verbal
selectional restrictions. which. according to Givén, mark

them as post semantic change, and that they were of the same

fl

genre a
eubute
peealbl
ly eema
For ere
meg
yet eer
plural
[eontr
(of. s
[locat
of 0111

3.5.2
‘ q
hypotl

those

 

 

115

genre as [human] in that they are associated with the
subcategorization of verbs according to types of arguments

possible. This is not the case. however. These are strict-

ly semantic categories which have nothing to do with function.

For example, a nominal in the category [instrument]. e.g..

grinder. can function as subject because it is [potent].

yet semantically it is also an instrument because its

plural is predictable as such. The terms [potent] and
[controllable] label gnostotactic classes in the nominalia hierarchy
(of. Section h.2), while the terms [instrument]. [agent],

 

[location] are here used as components in the hierarchy

of cultural perceptions.

3.4.2 Sygchronic Evidence

There are two types of synchronic data for which a

 

hypothesis such as Givén's provides explanation: first,
those items associated with more than one plural class and,

second; borrowed items.

3.4.2.1 Multiple Plural Association

In a good number of plural classes, there exist items
which belong to one or more other plural classes. In al.
most every case this multiple association can be explained
on semantic grounds. That is. an item may be perceived
in more than one way. For example. gig; 'penis may be
classified as [round (rod)] and thus be associable with the
plural class FMaaL (52222) or as [male genitalia] and thus
be associable with the plural class FMuuLa (zbuuba). And

 

*

m ‘hill'
necesociabi
(heel). or
nth the pi
here such c
chaotic ass
ilcrity or
plural clan
necrganizat
center, in
tc-one some
been shift:
on the has!
A 5001
angular ]
in moving

 

3latch, Ii.
c1 nasculi
assimilate
Plural. e.
‘tclevisic
liens ante
numerous 1
ill w
ll: 11': 3
betinne c

It 18 has

 

116
gal; 'hill‘ may be classified as [round (convexfl and thus
be associable with the same Eound] class as above. FhaaL
(flail). or as [female genitalia] and thus be associable
with the plural class FHuuL (3211131.)- In the instances
where such an explanation is not possible the apparent
chaotic association can be explained by phonological sim-
ilarity of the singular form to other items in the same
plural class. It is hypothesized that. since the semantic
reorganization to a binary. hierarchical. anthropccentric
system. in which there is no longer an emphasis on a one-
to-one semantic-morphological correspondence. items have
been shifted around the various broken plural classes
on the basis of phonological shape.

A good number of items with broken plurals also have

 

a regular plural. This gives evidence that the system

is moving toward an ultimate regular (suffixed) plural
system, with the only distinction made being the grammatic-
a1 masculine/feminine dichotomy. Borrowed words not fully
assimilated will always be associated with the feminine
plural. e.g.. 321;; 'beer' p1: M; W
'television' pl: palaviziyamp. Examples of native

items entering into the regular plural system6 are so
numerous that we list Just a few here: w 'queen'

pl: w. 3.11251 'chair' pl: ggsiyaatr mu'allim ‘teacher'
pl: gu'allimiig. To list all such items would not only

be time consuming but nothing would be‘gained by so doing.

It is best to Just say that more and more items

fl

ct undo
than pre

The
hccecl ca
clternat
exceptic
with in

lieu

fuiir

heath

caill

nxnnnac

 

 

117

at random and with much greater frequency of occurrence

than previously enter into the regular pluralization system.

The following is a detailed listing Of all semantically

based cases of multiple plural association except where the

alternate plural is a sound (regular) plural or is an

exceptional plural.

with in Chapter V -- Residue.

Item

faajir
kaafir
rajil
mxannaO
faaris
mugasg
kaagada
qubqaab
hajim

bahiG

Plurals
fajara
fujjar
kafara
kuffaar
riyaajil
rjuula
maxaaniiO

mxaanOa

fawaaris
fursaan
mgaasiis
mgaasa
kwaagiid
kwaagid
qbaaqiib
qabaaqiib
hujuum
?ah3aam

buhqu
?abhaa0

Semantic

weak. low
agent
weak. low
agent
male

male

low
low

strong
male

instrument
round
square
square

round
round

bulk
abstract

abstract
abstract

Singular
Canonical
Fozms

PaaMiL
FaaMiL

FaaMiL
FaaMlL

FvaL

FvMMvL

FvaL

FvavLa

FvMMva
FvMMva

FvaL
FvaL

FvaL
FvaL

Exceptional plurals will be dealt

Gloss
libertine

infidel

husband

effeminate

man

knight

scissors

sheet

Pair of wood-

en clogs

bulk

research

 

 

Ln;

ceiliq

iiiia

nuance
rial
rna'i
'iahmaq

ia'zal

ia'ncar
Ougul
blir

huhur

tall

 

 

Item

sadiiq
filis
?axras
xuswa
xisi
raa'i
?a).1maq

?a'zal

?a'war
Ougul
biir

juhur

tall

Plurals

—————

?asdiqaa?
sidqaan

filsaan

flaas7
fluus
xirsaan

xurus
xisyaan
xasaawi
xisyaan
xasaaya
ri'yaan
ru'aat
humuq
humaqaa?
'uzul
'uzzal

'uur
'uuraan
Ggaal
?a0qaal
byaar
?aabaar
Jhaar
jhuur

tlaal
tluul

118

Semantic

strong
close

low
round

female genital

weak
weak

weak. low
weak, low

weak. low
weak. low

weak. low
agent

weak
weak. low

weak

Singular
Canonical
Forms

 

FaMiiL
FvavL

FvM L

?aFMaL

FvMLv
FvMLv

Fva

Fvav

Faahi

7aFMaL

?aFMaL

weak. child-

like

weak
male

round
abstract

round
round

round
female gen-

?aFMaL

FvaL
FvaL

Fva
Fva

FvML
FvaL

italia. holes

round

female gen-
italia

FvML

v
ia. holes

Gloss
friend

one file

mut e

testicle

eunuch

shepherd

dumb

defenseless

one-eyed

weight

oil we11

anus

hill

 

 

zibb

kuux
uiiu
iiful
huuz
inch
iarcnn

Scenic}

)iriif

zibb

kuux
piip
quful
kuuz
Euub
sariik

saayih

gariif

qlss

dayl

 

Plgrals
hbaab
hbuub

ctaaf
ctaafaat

zbaab
zbuuba

kwaax
kwaaxa

pyaap
pyaapa

Qfaal
Qfaala

kwaaz
kwaaza

cwaab
cwaaba
surkaan
surakaa?

siyyaah
suwwaah

?a§raaf
surafaa?

qsuus
Qasaawisa
dyuul
dyuula

119

Semant c

round
female gen-

Singular
Canonical

Fozgg
FvML

italia. holes

round
paired body
part

round

male gen-
italia
round
round

round
round

round
round

round
round

round
round

male
agent

agent
agent

abstract
strong.
respected
death
agent

elongated

male gen-
italia

 

G;O§§

large. per-
ous pottery

vessel
FVMVL shoulder
FVML penis
Fva hut
Fva
Fva barrel
Fva
FvaL lock
FvaL
Fva clay urn
Fva
Fva tube
Fva
FvavL partner
FvavL
FaaMiL traveler
FaaMiL (2 diff. underlying

forms)

FvavL distinguished
FaMiiL
FvML clergyman
FvML tail
FvML

 

 

him
Th
Plural
ence in
cornea;
items n
reEula
iitilla
basis
Some n

hith "

inhc
i are

Iint

 

120

 

Singular
Canonical
Item Plurals Semantic Forms Gloss
sayr syuur elongated FvML leather
syuura male gen- FvML strap
italia
siffa sfaayif round lip
sfaaf round FvMLv
sariit 7asrita instrument FvavL ribbon
saraa?it instrument FvavL

3.4.2.2 Borrowed Items

 

The assignment of borrowed items to various broken
plural classes probably provides the best synchronic evid-
ence for the still partially operative semantic-morphological
correspondence of the pluralization system. Those borrowed
items which are not assigned to the morphologically marked
TGSUIar plural category [feminine are assigned to par-
ticular broken plural c1ass(es) in many instances on the
basis of their culturally perceived semantic content.

Some of these items also have an alternate association
With the sound feminine plural -a§§. The following is a
list of borrowed items assimilated into the broken plural

SYetem on the basis of semantic content.

 

Singular
Canonical
2‘2! Qaisiu Plural seam rare-s...— glass
kuuliis French FaMaaMiiL square FvMMva ggegfizg
side of
a stage
makiina Italian FaMaaMiL instrument maFMvLa machine
'arabaana Turkish FaMaaMiL instrument cart
hospital

xastaxaana Turkish FaMaaMiL place

 

 

H
:0
I3

i

feylasui

daktawr
qunaul
teller:
xariitc
karma

tuub

huub

nuhur

banid

butil

kiti:

clan
Fain.
the

Item

 

faylasuuf

daktawr
qunsul
tallaxaana
xariita
karxaana

guub

kuub

muhur

banid

butil

kitli

Origin

French'

English
English
Turkish
German

Turkish
English

English

Persian

English

English

English

121

Plural

FaMaaMiL
FaMaaMiLa

FaMaaMiL
FaMaaMiL
FaMaaMiL
FaMaaMiL
FaMaaMiL
FMaaL
FMaaLa
FMaaL
FMaaL

FMuuL

FMuuLa

FMaaLi

 

Singular
Canonical

Semantic Forms Gloss
strong philosopher
agent .
strong doctor
strong FvMMvL consul
place casino
place map
place factory
round Fva tube
round Fva
round Fva cup
round FvaL personal

stamp
female FvaL bonnet
genitalia,
holes
male gen- FvaL bottle
italia
round FvMLv kettle

Notice for example that the Turkish items which are

classified semantically as places take the plural pattern

FaMaaMiL even though they contain too many radicals to fit

the Arabic plural pattern. Examples:
xastaxaana 'hospital' p1: xastaxaayin
tallaxaana 'casino' p1: tallaxaayin
karxaana ‘factory' p1: karxaayin

 

 

nhere gag
Arabic wor

maxba
maxza
madba

where the

'the place

 

122

where -xaana is the Turkish morpheme meaning 'place'.

Arabic words which fall into this S-class are words like:

maxbaz 'bakery' pl: maxaabiz
maxzan 'storeroom‘ pl: maxaazin
madbag 'tannery' pl: madaabig

where the prefix mg— is the derivational affix indicating

'the place at which'.

 

 

 

 

l. Ernea
probably
nouns of
class. l
of [place
clean ) .

u- ([inc
Third, In

Plural cI

2: A000
is nm
the shap

bi regul
Wheat.

3. hip
In p
5' But
6. For

part101i

lg

 

FOOTNOTES
Chapter III

1. Ernest N. McCarus' hypothesis is that there were
probably three stages which led to the massive number of
nouns of [place] and [instrument] in the FahaaMiL plural
class. First. broken plurals existed (maybe some nouns

of Eflace] and [instrument] were in the FaMaaMiL plural
class). Second. nouns beginning with mgr (l?lace]) and

21' ([instrumentl) evolved via the derivational system.
Third. such newly formed nouns were assigned to the FahaaMiL

plural class on the basis of phonological shape.

2. According to Ernest N. McCarus. the shape of the singular
is usually analyzed as FaaMiY (1.9.. a FaaMiL agent). and

the shape of the plural is FuMaLat (where L n y) and which

by regular morphophonemic loss of intervocalic y gives

FuMalt e

3. In private communication.
4. In private communication.

5. By the present author.

6. For native items. the types which are associated with
sound plurals may be delineated. e.g.. all derived form

Participles, FaMMaaL-nouns. etc.
7. fluus is also used collectively.

123

 

 

 

 

A
entails
dence a
the ca:
the m

Ir
esaary
the an
able t<
Harem
featum
but the
Other I
'1th t}

apeem
”3301- e
“ales:
ever. '

phone]. (

 

CHAPTER IV
A Theoretical Description

A synchronic theoretical description of broken plurals
entails describing the semantic-morphological correspon-
dence still extant in the system as well as accounting for
the cases of pluralization based on the canonical form of
the singular.

In order to effect such an analysis it becomes nec-
essary to reclassify the data. To obtain predictions in
the direction from semantics to morphology. one must be
able to uniquely specify the plural class to which. for
example. an [abstract] item belongs. Not only is the
feature[abstractlassociable with several plural classes.
but there are also items classifiable as [abstract] in
other plural classes. which are not overtly associated
with the feature [abstract] .

The following section is an attempt to uniquely
specify the plurals for those items associable with the
major semantic categories specified on page 57. The fina1
analysis also provides for all the broken plurals. how—
ever. whether they are classified on semantic grounds.

Phonological ones. or a combination of the two.

12h

 

 

 

m
FWLU

MVL
ML
M
Fva
thI

FVMLa
FVMLa

FVML
thL

M
3 radi

1’va

hm.
hm
FVMva

Fva
FVMvL
“We
FVLv

‘30met
Item 0
ML.

M.
W

125

h.1 A Classification of the Data

abstract Plural
Fva(L)v FaMaaLL (3.2.1.33)

FvaL

FVMLa FVMaL (Appendix)
FVMLa FaMaLaat (3.2.1.310

FVML } FuMuuL (3.2.1.12)

EQEEQ

3 radicals:
Fva
FvML(a)

FvaL
FvavL

FMaaL (3e2e101)

Fva

FVMVL FMaaLa (3.2.1.2)
FvML(a)

Fva

Remarks
uniquely specified

high frequency
of occurrence

only two members

lesser frequency
of occurrence

high frequency
of occurrence
(alternant)*

high frequency
of occurrence
(alternant)*

*Sometimes native speakers assign both plurals to a given
item or are not sure which of the two to assign.

FvMLv
FvaLa FMaaLi (3.2.1.3)
Fva1(yya)

uniquely specified

 

hradic

F(V)Hv1
F(V)Mefl

FflMvL
Fvale
FMva1

FVMMV
FvMMv

FvMMg

Fva
FvMM

Fen)

Fle

 

126

rgggg. cont'd

Singular Plural Remarks

7 FaMLaat (3.2.1.11) only two members
FVMLa FVMaL (Appendix) P-shape class
F(v)Maaya FuMi (3.2.1.9) uniquely specified
a radicals:

 

 

F(v)Mvaa 1 d
F(v)MvMLa FiMaaMiL (3.2.1.u) uniquely specif e
FvMMvL(a)

FVVMLa FMaaMiL (3.2.1.7) uniquely specified
FMva(a)

FVMMVL(a) FaMaaMiiL (Appendix)

FvMMva(a) uniquely specified
FvMMaaL(a) FiMaaMiiL (3.2.1.5) only three members
FVMVL } FuMaaMiiL (3.2 .1.6)

FvMMva

FvMMva(a) FMaaMiiL (Appendix)

FvMMvL(a) FaMaaMuL (3.2.1.16) only two items
FuMLv(v)L FMaaMuL (3.2.1.8)

sguare

3 radicals:

FVMLa FVMaL (Appendix) P-shape class

FvMLa FiMaL (Appendix) uniquely specified

4 radicals:

FvMMva(a) FMaaLiin (Appendix) uniquely specified
FVMMVL

 

 

$3229.:
Si 5
FvMMvvl
FvMMw]
FvMMvv]

FvMMvL
FMwLa

m
2) radi

 

FvMMv
FvMMvL

FVMHv-v‘
FvMHvL'

3 radi
Wm
WM,
FVMLV

FvaL
FVML

FVMVL
WV

FaMnL

Fa(11L
Pam

he

3 ram

“PM
”at

 

 

s uare. cont'd
Singular
FvMMva(a)
FvMMva
FvMMva(a)

FvMMvL(a)
FMvaa

strong

4 radicals:
FvMM(v)vL (high)
FvMMvL(a)

FvMMva
FvMMva

3 radicals:
FVMLa (high)

FvaL (war)
FvMLv

FvaL
FvML

FvaL
FvML(v)

FaMiiL (high)

FaMiiL
FaMi

weak
3 radicals:

?3FMaL (low)
FMaaL

 

12?

Plural Remarks

—————-

FaMaaMiiL (Appendix)
FuMaaMiiL (3.2.1.6)
FMaaMiiL (Appendix)

FMaaMiL (3.2.1.7) uniquely specified

FaMaaMiiL (Appendix)

FahaaMiL (3.2.1.17)
(high. numerous. large)
FaMaaMiLa (3.2.1.39)

FVMaL (Appendix) P-shape class

FiMaaL (3.2.1.1u)
FuMuuL (3.2.1.12)

F1MUUL (3e2e1e13)

FuMaLaa? (3.2.1.37) uniquely specified

(respected? ?aFMiLaa? (3.2.1.18) uniquely specified

FiMiL (3.2.1.21) uniquely specified

 

 

 

 

me.
my
YaFMaI
FaMiiI

FvMiiI
FaaMiI

emu)
?
Fv(v)l
1’va
FvMLv
FvaJ

FvMLa
FVMLII

1) radf

FVMMw
FVNMV]

mam

ML)
3 rad:

NW)
FWM)

)an

 

128

weak. cont'd

Singular Plural

?aFMaL (defective) FuMuL (3.2.1.22)
FuMaLaa? (3.2.1.37)

 

FaMiiL (low)

FvMiiL

(childlike)}FuMMaL (3.2.1.20)
FaaMiL

FaaMiL (10W)
? (low)

FaMaLa (3.2.1.24)
FaMiiL (3.2.1.25)

Fv(v)Mv

Fva
FvMLv (1°W) FIMLaan (3.2.1.23)

FvavL

FvMLa

(low) } FaMaaLi (3.2.1.40)
FvMLiya

4 radicals:

Remarks

 

uniquely

uniquely

uniquely

uniquely

only two

uniquely

uniquely

FaMaaMiiL (Appendix)
FVMMVVL(3) (low. base. common. insignificant)
FvMMvL(a) uniquely
FMaMLat (low) FMaaMLa (3.2.1.26) uniquely
male enitalia
3 radicals:
F(V)MVVL FuMLaan (3.2.1.29) uniquely
Fv‘7M(v)L
FVML FMuuLa (3.2.1.28) uniquely
FvML
h radicals:
FvMMaaL FiMaaMiiL (3.2.1.5) uniquely
FvaL

specified
specified

specified

specified

members

specified

specified

specified
specified

specified

specified

specified

 

 

 

 

in
em

FvaL
FvML(

use

FvaL
FvML

F1M1L
FvaL

m
FvaL

FVMLa

mi

?aFMa

Mpg:
3rad
)5an
(am
F33M1

FaHie

Singular
female genitalia

FvaL

FvML(v)

other body parts
FvaL

FvML

FiMiL (paired)
FVMVL (paired)

plants
FvaL

FvMLa

colors
?aFMaL
?aFMaL

nationality
?

2213 (minor class)

FvaL
FvML

exam
3 radicals:
FaaM1
FaaMiL
FaaMiL
FaMiiL

 

FMuUL (302e1e27)

FMuuL (3.2.1.27)

FMEaLaat (3e201e31)
FVMaaLaat (3.2.1.30)

FMuuL (3.2.1.27)
FVMaaLaat (3.2.1.30)

FuMuL (3.2.1.22)
FiMiL (3.2.1.21)

singular minus -1
(Appendix)

FuMuuL (3.2.1.12)

FuMaat (3.2.1.38)

FuMMaaL (3.2.1.36)
FiMMaaL (3.2.1.35)
FuMaLaa? (3.2.1.3?)

 

Remarks

uniquely specified

uniquely specified

only two members

only two members

uniquely specified
uniquely specified

uniquely specified

only three members

not really a broken
plural

uniquely specified

uniquely specified

uniquely specified

 

 

 

 

 

 

my;
§iggg
h rad
FVMMV

FvMMv
WNW

lice

FvaL

FvMMv

mm
FVMMV
FVMMV

F‘7va
FMva

Dhenol
First
Tune):
”8th
eccen1

Figur.

 

130

252329 cont'd

 

 

 

Singular Plural Remarks
4 radicals:
FvMMva FMaaMiiL (Appendix)
FVMMVVL(V) E FaMaaMiLa (3.2.1.39)
FvMMva
location
FvML as?

ya FaMaaLi (3.2.1.uo) uniquely specified
FvaL
FvMMvL(a) FaMaaMiL (3.2.1.17) uniquely specified
instrument
FvMMva FMaaMiL (3.2.1.?) uniquely specified
FvMMvL(a) FaMaaMiL (3.2.1.17) uniquely specified
FVMVVL ?aFMiLa (3.2.l.hl) uniquely specified
FMva

It appears that there are actually three co-existent
phenomena in the plural system of Modern Iraqi Arabic.
First there is a major dichotomy between those nouns which
function in the older semantically—based internal plural-
ization system and those which function in the newer anthro-

pocentric suffixed system. This dichotomy iszhown in

Figure 13 e

 

 

 

81
appear:

Ft
there 1
““0015
t(ten)
assWe!
lying a
are 0&1)
“thin
(10) 31

lung 8

 

131

nouns

  
 
 
  

+e Gnos+o+uc+ks A

Su$¥ix System
MOEPHO" (Sauna Plurals)

TACTIC S

Simulfix S stem

(bI‘OKfiK P uroJS

Figure 13
Broken/Sound Morphotactic Dichotomy

Since the suffixed system is the productive system it
appears as the unmarked choice.

For those nouns that function in the older system
there is a second dichotomy between those whose plural
association(s) are semantically defined (i.e., those still
totally within the older system) and those whose plural
association(s) are morphologically defined by the under-
1Ying shape of the singular morpheme (i.e., those which
are caught midway between the older and newer systems).
Within the first type there is frequently a multiple associa-

tion with particular plural classes on the basis of under-

lying shape.

 

 

 

Cons
lexi
whicl
grow
with
mine
whom
fire
is 0:

claim

not I
Fin!
are 1
mi:

singi

 

132
Consider the semantically defined [plant] class. Those
lexical items which may be so semantically defined and
which enter into the internal plural system on semantic

grounds are associated either with the plural FMuuL or

 

with the plural FVJMaaLaat. The choice is uniquely deter-

mined by the shape of the singular, however. Such an item

whose singular is of the form FvaL is associated with the

 

 

first plural class cited (FMuuL) while one whose singular
is of the form Egggg is associated with the second plural
class cited (FVlMaaLaat).1

Examples of items entering into the older system but

not on semantic grounds are numerous. Actually the entire

FVIMaL plural class seems to be of this sort. While there
are—many semantically definable groups in this class, the
entire membership is represented by the form Egan; in the
singular.

Figure In is a diagram indicating how these various
dichotomies might appear in the morphotactics of a stratifica-

tional analysis. It implies that the choice of broken and

semantically-based plurals is ”cognitively motivated.” The
semantically-based plurals are easily Justifiable as so

motivated by their very nature. That is. such plurals are

semantically (in the sense of gnostotactic hierarchies) or
OOgnitively defined. Broken plurals are assumed to be

"cognitively motivated" because they are the marked of the
broken/sound dichotomy, and they are neither grammatically
nor semantically (in the sense of the sememic stratum) mo-

tivated. They appear rather to be a conscious (to the ex-

tent that language is conscious) choice.

 

 

 

 

 

hmoth
“er
rhat

 

P'uralé) O'F nou n5

  
   
  

sin-MW”
system
CbroKen plum ls)

Su-F—Fix System
(sound Flum is)

Seman‘l'imuy .
based

System P— base—J aid-em

Mo RPHOTACTIC. 5

Various Semavd'ic. Cahcbories

Figure 1b
Semantically/Phonologically-based

Morphotaotic Dichotomy
If one were to make predictions on the basis of the
hypothesis that language changes in the direction of the
unmarked or productive system, then one could say that the

first simplification of the pluralization system of Modern

 

 

 

Iraqi
sssig
than

hers

broke
the E
will

as is

tics

ized

ihis

 

 

134

Iraqi Arabic will probably be that broken plurals will be
assigned on the basis of their phonological shape rather
than on semantic grounds (c£.FVJMaL -- all of whose mem—
bers have the same P—shape in the singular. Elflég).

The ultimate simplification will probably be that
broken plurals are eliminated altogether. leaving only
the sound or suffixed plurals. The only remaining dichotomy
will be the masculine/feminine (marked) grammatical gender.

as is the situation in many Indo-European languages today.

4.2 The Gnostotactics

A description of the relevant parts of the gnostotac-
tics of Modern Iraqi Arabic must account for the hypothes-
ized current binary hierarchical, anthropocentric semantic
taxonomy. At the same time gnostemes within this con-
figuration which correspond to the older semantic non-
hierarchical taxonomy must have direct connections with the
morphotactic plural classes which they define.

The present work posits an interconnected gnostemic
system -- one hierarchical and the other not -- to account
for the data. In order to make the final diagram com-
prehensible, however, we will first discuss various prob-
lems encountered in constructing it. The binary hierarchy
which involves features relevant to subject selection pre-
sents the first problem.

There have been two alternative systems posited to
account for the surface phenomena related to subject
selection. Fillmore (1968) posits a deep level argument

Infiizgmggt with subject selection rules of the types

 

 

 

 

 

in

hand. 1)

account

Th
Th

The fee
better
like a.
like b.

a.

W:
I Fillmo:
Bentem
Accord

thus b
Em. ‘

8mm

like

 

135
;4 Agent-—9-Instrument. Goal. Chafe (1970). on the other
hand. posits an additional feature in the hierarchy to

account for sentences like

The flood destroyed the city.
The key opened the trunk.

The feature he posits is [potentlwhich seems to offer
better explanation for the grammaticality of sentences
like a, and the deviance from grammaticality in sentences

like he

a. John Opened the door with a key.
b. *John opened the door with the wind.

With only the instrument-as-case hypothesis (of.
Fillmore. 1968) the grammar should be able to yield a
sentence such as b if it yields a sentence such as d (below).
According to Chafe. however. both kg; and gin; are [+potent],
thus both can be selected as the subject of verbs like

open, yielding sentences such as

c) The key opened the door.
d) The wind opened the door.

There will still be further semantic restrictions
governing particular verbs. of course. since sentences

like

e) The key destroyed the city.

 

 

 

are not

context ‘
s)

are.

An
that by
subject
may simp

An
Lockwood
Kramer.
archs vi
the inst
lelotact
the subs
by Verbs
be hand]
Still a.
does 00c

Con

 

136
are not entirely grammatical without an extension of

context while sentences like

The wind destroyed the city.

v

f

are.

An additional point in favor of Chafe's analysis is
that by adding the feature [potent] one may eliminate a
subject selection rule (or in stratificational terms. one
may simplify a part of the lexotactics).

An approach similar to Fillmore's was adopted in
Lockwood's (1972) introductory text in stratificational
grammar. In the present work, however. Chafe's feature hier-
archy will be expanded, thereby eliminating the need for
the instrument subject selection construction in the
lexotactics. Rather only [potent] nouns can occur as
the subject of a certain subcategory of verbs. represented
by verbs such as 222E and destroy. This phenomenon will
be handled in the gnostotactics. The lexotactics must
still select which of two [potent] nouns. for example.
does occur as the subject, however.

Consider the following sentences:

 

cone:
coun‘
pote:

00M:

 

cone:
potes
[tenet

eoun‘
potes

r———\

count

r‘_‘

Deter

r")

("limes

lotto:
titles

“I the

Witch
term
eonnee
mm
Dream

“it a]

 

 

 

137

 

concrete

count Theb bulldozer destroyed the building.

potent The R fiopened the door.

concrete *The sidewalk destroyed the city.

count

concrete Nate; destroyed the city.

potent

[concrete] *WOod destroyed the house.

count My proof destroyed his argument.

potent

[count:] *Ideag destroyed the city.

[hOtht] §lazgrx destroyed the South.
Expgzigngg opened my mind.

(unmarked) *Hazm destroyed the city.

Starred items are not grammatical without expanded
contexts. Note that only the [potent] nouns yield gram-
matical sentences in a "normal" context when occurring
as the subject of verbs such as gpgn, or degtgoy.

F1sure 15 shows that portion of the gnostotactics
Which accounts for the taxonomic hierarchy of features
d°f1n1n8 nominalia. Each of the classes shown will have
connections to other hierarchies as wellz. such as the
mumslia hierarchy shown in Lamb a971b'p-221) and the
Predication structures including verbalia and deep level

case shown in Ikegami (1970. p.170ff.)

 

 

 

ovum

huoivo

 

138
nomimliq
“M‘s v , assess
CD”“+ V v v V "Oh-Couyd"

Potevd’ V V V V V v v V non—Poker)

Sedewalk wider ”“4 Fr“; ’46“ Sl‘Ve'lme “Arm

0”“er V V imwimf‘e

human 7 v m- Key
human. 00‘"
(medullae
‘FQerLina V V an“! V V Vac-Faint)“
wfimv V nut—Suah‘exe.
”“3“ that
v v V V V V V € ‘tree.
flower
_. \ H
m" Maw 6a,“. cow WAIMAA by.
and“; Wis»! adm be D s b‘ihie Raver Ct,"-
Zwba -' mass : : - - -
Figure 15

Taxonomic Hierarchy of Features
Defining Nominalia

 

)

 

Thi
hie

ree
But

nee

 

139
This hierarchy is an expanded version of Chafe's (1970: pp.108ff.)
hierarchy. The features [concrete vs. abstract) and
Enimal vs. vegetation] have been added for the obvious
reason that they are necessary in verbal argument selection.
But it appears that there is still a further dichotomy

needed to explain why a sentence like

John smashed the bridge with his car.

is grammatical but not
*John smashed the bridge with a hurricane.

It seems intuitively straightforward that what is needed
is a feature [controllable]. and that this feature
[controllable] is what determines which nominals can
function as instruments.

Consider the following sentences.

 

concrete *The man destroyed the building with wind.
potent Wind destroyed the building.

concrete The man destroyed the building with £222-

potent Fire destroyed the building.

controllable

concrete *The man destroyed the building with a tidal
count wave.

potent A tidal wave destroyed the building.

boncrete The man destroyed the building with a bulldozer,
count A bulldozer destroyed the building.

 

potent
kontrollable

 

the}
cent
16 e
or

1’40

Because all of the underlined arguments are potent.
Only the two which are
Figure

they can function as agents.

controllable can function as instrument. however.
16 shows the extension of Figure 15 to include a class

of controllable] items .

Q1) KG refitt—

COuwl’ , V l0K-COMK+
Pohlf V Pb‘hvd' V
midweek v V imim+¢
Cowtrclbble' V Y Y uguwl'rpnable
+1.4“ “Fire wake"
”1"" hurricane. wheel

Key Skip bunclow

Figure 16
Extension of Figure 15

 

 

 

 

and
since
1090'

the e

srohy
only

[etrc

end i

eons
0011!

not i

into]
and e
ivids
entis
“Omit
two 5
the l
chicE
firm

in s:

 

141

The features [round vs. square]. [color us. non-color].
and [strong vs. weak] must also be added to the hierarchy
since they are justified by the form of Modern Iraqi Arabic.
i.e.. because these distinctions are necessary to explain
the current semantically associated morphological classes.

But these features do not seem to fit into the hier-
archy already established in any neat way. About the
only generalizations one can make are that the feature
[strong] implies the feature [potent] but not vice versa.
and the feature(s) round/square] imply the features
[concrete] or ”unmarked“, e.g.. :otunditz. gguagenegg. but

count
not vice versa.

The best explanation seems to be that there are two
interconnecting systems -- a hierarchy as given in Figure 15
and expanded in Figure 16. and a non-hierarchy such as
evidenced by the data as characteristic of the older sem-
antic system. In other words. any semantic “unit" in the
nominalia hierarchy will have three connections -- to the
two aforementioned (the animalia. etc. hierarchies and
the predication constructions) and also to a non-hierar-
chical system of cultural perceptions (round. square.

strong warring. weak childlike. etc.) such as that shown

in Figure 17.

 

 

 

 

has

dis

its

01:

Des

142
guHunLIPenfifHOKs

“round a I‘ more" It
" bshun+
class 0 its, ' a class
Cannaiiows t i h
+1 MMiKA—lida
)Llerdrcky -b Mgr-ADM“
V
7'
Key |
Figure 17

Cultural Perceptions in the Gnostotactics
of Modern Iraqi Arabic

while the members of each culturally perceived class
have connections to the nominalia hierarchy. each class
as a whole (signified by a diamond node above) has a
direct connection to the morphotactics three strata below
where it serves to define certain morphological plural
classes. In the case of multiple association the given

nominal would have connections to two (or more) culturally

perceived classes.

 

1&3

h.3 The Morphotactigs

When considering the morphotactics of Modern Iraqi

Arabic, two problems arise. First. in order to arrive

at a simpler (or more general) tactics it becomes neces-
sary to group together those broken plurals which are

morphologically similar. On the other hand. in order to

adequately account for the data. certain semantically
defined broken plural classes must be kept distinct re-
gardless of their morphological similarity to other plural

classes. The following is an attempt to classify broken plurals

by morphological similarity.
The notation here used signifies an ordered occurrence

of the first. middle and last radicals (FML) simultaneous

with (e) an ordered occurrence of the internal vowel

sequence (e.g.. flat). ’

One generalization which becomes apparent when viewing
the following morphological shapes is that an overwhelming
majority of internal plurals in Iraqi Arabic have either
a or a; as the final vowel of the internal vowel sequence.
This seems to be a fairly common internal plural type in
the Hamito-Semitic family (cf. Greenbergsl955. pp.198-20h).
while it may not be impossible to group all a/g; final
vowel plurals together in the morphology and enable the
various initial vowels to be realized in the morphemic
Sign pattern. such an analysis presented other complexities

and thus was not used in the present description.

F

 

PM

a

P

M

M

mu
wk

 

FML' a:

FMaaL
FMaaLa
FHaaLi
FHaaLiin
FHaaLaat
?aFHaaL

FaHaLa
PahaLaat

FHL'uunJi;
FuMuL

F1M1L
FuMuLaa?

FML'ia
FiHaL

FHL'ua
FuMaLaa?

 

FML'Vg:
FVIMaaLaat

FML'

FiMLaan
FihLaat

round
round Sgt FML'Vfl

round

square

paired body parts
abstract

P-based Sgt FMLOVfl a

weak
abstract

color. weak
color. weak
weak

square

agent. strong. weak

paired body parts. plants

weak. close
meal

male
male

 

m

If?
I:

 

 

FML'a
FaMLaat

FML'aa:
FaMaaL
FaMaaLa
FaMaaLi

FMLO a:
FiMaaL
FiHaaLa

FML°uaI
FuMaaLi

FML'

7aFMiLa
?aFMiLaa?

FML- u
?aFMuL

FML°uuanJiu3

FiHuuL
FuMuuL
FuMuuLaat

FML-

FMuuL
FMuuLa

FML‘ais
FaMiiL

 

145

1‘ ound

female relative
?

place. time. weak

strong
strong

strong

cover. statement. instrument
strong. respected

strong, war
strong. war. bulk. abstract

?

female genitalia. death. elongated. plant
male genitalia

weak

 

 

M\

M»
W

 

 

FMHL‘aani
FaMaaMiL
FaHaaMiLa

FMML~aasu
FaMaaMuL

FMLL'aa:
FahasLL

FMHL'iauiAJuagg

FiMaafliL
FuHaaMiL
FMML- aninJ nu

FHaaHiL
FMaaMuL

FMML'lallI/Uuasia

FifiaafiiiL
FuMaaHiiL

EflfiL-Qatia

FMaaMiiL

Fifi—'aami;
FahaaMiiL

FMML'ugagflJiga.

FuHMaaL
FIMMaaL

M

F1MHaL
FuMMaL

1&6

place. instrument. strong
agent. strong

strong
abstract

round
round

round
round

male . round
round

agent. instrument. square. round
(P-class)

round. a uare. weak. strong. statement
(P-class

agent
agent

weak
weak. childlike

 

 

M

simi]

PD! 1

are a
stror
that
overt
that
lhile

For a

tan 1
tfittt
Mtt

 

1h?

FMML’ a:
FMaaMLa weak. low
FMouNi
FiMi 7
FuMi round
FuMaat agent

The previous categorization on the basis of morphological
similarity must be integrated with the categorization (on
pp. 123ff.) on the basis of semantic similarity.
Before a stratificational diagram is attempted. there
are several notions which need explanation. however. A
strong point in the theory of stratificational grammar is
that it has a mechanism which allows one to capture an
overt generalization present in Semitic. namely the fact
that tri- and quadri-consonantal roots are lexical in nature
while many of the vowel patterns are grammatical in nature.

For example. (Erwin: 1963. p.#?) the items

gigs; 'he studied'
?ag;u§ 'I study'
gagig 'lesson'
giraaga 'study'
gagrag 'he taught'
mugagrig 'teacher'
magraga 'school'

:an be seen to involve the same lexical triconsonantal

battern interdigitated with different derivational vowel

:atterns.

 

 

148

With the mechanism of tactic precedence stratification-
al grammar can separate the lexical “morpheme" from the

grammatical ”morpheme” at the S-morphemic level thus.

olrs a6

Figure 18

The Separation of Lexical and Grammatical
Morphemes in the Morphotactics

Take the noun daris for example. In stratificational

 

terms one can say that the root 9;; occurs simultaneously
(cf. the downward unordered egg above) with the derivational
vowel sequence 5;. However. g. z and g must be ordered

with respect to one another in the order given (cf. the
downward ordered eggs above). The tactics of the next
lower stratum. namely the phonotactics. takes precedence

over the ordering supplied above. It is the phonotactics

which interdigitates the consonants with the vowels.

 

 

 

The p
while

the 0
way t
Hheth
vowel
as do

inter

nomin
of th
1101111!
@1355
name
like

tliég
nonet
we ,1
Which
Speak
(Sthe

tub

deriv
1“it

Vote}

 

1&9

The phonotactics only allows a CVCV...(8) sequence. Thus
while it receives the 9;; and 5; simultaneously drs V).
ai

the order it imposes is d g r i g. i.e.. CVCVC. In this
way the root gr; need only appear once in the morphOtaotios.
Whether it functions nominally or verbally (i.e., what
vowels are interdigitated) is significant only insofar
as derivation supplies the exact meaning intended. by
Lnterconnecting with the lexical root.

Derivation is an important and integral part of the
10mino-verbal system of Arabic. But while a good number
m the nominals in Arabic can be derived from verbs. these
nouns are in the main abstract nouns like zurrizza 'freedom'.
irggs§;%tudy' (e.g., the £222! of Latin) or agentive
ouns like mu'allig 'teacher'. kggtig 'writer' or adjectives
ike naggi; 'dry'. £32; 'free'. Concrete nouns like
itggb 'book' and Eerie 'lesson' are significant but
onetheless are in the minority. Host of the nominals
s will be dealing with. however. will be of the type
nich do not enter into derivation except when a native
peaker consciously coins a new term (for example. if
5)he were to make a verb 'to cup' from the borrowed item
.1312 'cup’ ).

However. nouns like kuub are not presently in the

 

:rivational system. Their vowel patterns are provided
dividually. though they fall into a finite number of

mel pattern categories.

 

 

 

 

and
syt
The
ape

act

 

150

The problem lies in separating adjectives. agentive
abstract nouns. which do enter into the derivational
:em. from the majority of concrete nouns which do not.
former type of nominal must have its vowel pattern
led out after the tactics has assigned a vowel pattern
rding to the derivational system. while the latter

must have its vowel pattern spelled out idiosyncrat-
1y.

In the diagrams which follow. each broken plural class
be diagrammed separately. An integrated diagram
be presented at the end of the section on morphology.

l Semantically-based Plurals

In order that the reader may more fully understand
aemantically-based plural diagrams which follow. a
Lag of frequently occurring phenomena is presented
n
a) prefixes and suffixes

Figure 19 shows the general nominal config-
on in the morphotactics. Some stems are preceded by
xes. e.g.. ?aswad 'blaok (m.)° where 22- is a prefix.
stems are followed by one or two suffixes. e.g..
. pl: hyaaya. where <5 in the first case is a femin-

a_____

uffix, and in the second case is a determined plural

to

 

 

 

 

811

Vow

 

151

preFacs . \
Sumxes Pro MM'M
S (A F“ Xes
SflMu
74 ,- a I a MORPHO‘FAC'FULS
Figure 19

Prefixes and Suffixes

b) simulfixes
The general simulfix configuration is shown in
:ure 20, where the singular stem §gggp 'tube' occurs
ultaneously with the plural preemptive (of. e below)
e1 sequence 2:54: in the morphotactics.

$46M P‘“m\ Stimu‘RX

v hNoeFuhoTfichnzs
CVwfib ¢+ag+

Figure 20
Simulfixes

 

 

 

 

152
c) neutralization
Figure 21 details the neutralization of the pro-
minal suffixes -; (1 sg.) and -§_ (3 m.sg.) e.g.. c’uubi

' tube'. with the determined suffixes -; and -a_ as in

L1 p1: bwa_ari 'bugle‘.

   

A‘s Runomim' smt—Fi res

4e]? ‘ued ““fl| InoRPHEhfirc ALTERNAIION
gum,“ PATreeu

Figure 21
Neutralization

d) alternation
The alternation of the suppletive root m which
LIB in the environment of plural with the singular root
'woman' is shown in Figure 22. The conditioning line
ates from the plural path in the morphotactics and

M
les the "unit" )ns?/ to be realized in case plural is

en.

 

 

153

I- Plum,

us? ’
MORPHEM’EC .A LTERNATI‘OM
nu. ? PA‘WERN
Figure 22
Alternation

e) preemption

Figure 20 (above) shows an example of the simul-
:aneous occurrence of a stem with preemptive plural vowels.
figure 23 (below) shows how the phonotactics first checks
0 see if a preemptive vowel has been generated by the
orphotactics. and only in case one has not is a "normal"
owel generated. If a preemptive vowel has been generated
hen whatever "normal” vowel was also generated is realizad
3 zero. Thus. for example. fiyggb is realized as filigalib.

fa.

{2&2 'tubes'. (Interdigitation is explained in Section 4.“,

1e Phonotactics.)

 

can

{c1

ah
hrs
Th;

31:

 

15h
V
MDEPHourQ.
Aurenxnmn
'PAnTEFuA
FHONOTRIITRJS
a+ a
Figure 23

Preemption (Simplified Version)

.1.1 Round Plurals

 

.l.1.l FMaaL-tzpe

The diagram of Figure 2h is based on the following
mptions and generalizations.

1) Inflectional simulfixes (broken plurals in this

) appear to be replacives Just as they are in English
Lockwood: 1972). In English plurals like men. 2223
2. migg, etc. are handled in a similar manner. (The
llar vowel patterns of some nouns in Modern Iraqi

c are given by the derivational system. however.)
lural vowels gf and 51+ take precedence over the
tar vowels g. and a! respectively. Thus the gen-
:ation is captured that the singular and plural forms
he same consonants. The phonotactics accepts the

vowel (§+ or 21?) if it is available. and only in

 

155

e the plural vowel has not been generated does it allow
singular vowel to be realized. Plurals in Arabic are
e posited to behave in like manner.

While the phenomenon of pluralization in Arabic
ders on ”derivation". it is analyzed as distinct from
generation of derived vowel sequences for the follow-
'reasons. First. the phenomenon of pluralization in—
ves not only simulfixation but also suffixation. That
some realizations of the lexcn LN/p1ural/ are S-mor-
mic suffixes. e.g.,-a t. -;;n and -g. Second. the
nomenon of broken (or internal) pluralization involves,
addition to simulfixation (which is also involved in
ivation), preemption. Once the "derived"
gular has been generated. the Option is open to simul-
eously generate a preemptive plural vowel sequence.

2) Any member of the category of nominals semantically
:ified as round and having the vowel pattern v . for
Iple. conditions the choice of the plural FhaaL (i.e.,
vowel sequence gfaa+v* -- zero preemptive vowel. fol-
d by long ; preemptive vowel. followed by the phon-
ically predictable preemptive alternation i533 (V+)).
singular vowel sequence is present in the singulars
1e noun types designated G) through on the
am. Thus what were previously listed as five sep-
plurals can be seen to be predictable variants of
lural “allomorphs”. The plural types FHaaL. FHaaLa,
L. FMaaMuL. and FMaaMiL are all characterized by the

m GNyround/ and they are all morphologically related.

 

The
par
sir
pre

de1

 

156

Their individual surface realizations are for the most
part predictable from the shapes of their corresponding
singulars. The singular items (9 through and the
prediction of their corresponding surface plurals are
detailed in part h) below.

3) As an example. take the item ggggg 'tube'. The
vowel signified by V is predictable by phonological en-
vironment (here /__w) and its point of articulation is
therefore specified as a determined element by the phono-
tactics. If a speaker wishes to express the idea 'tube +
plural' in Iraqi he would generate the above lexon for
'tube' simultaneous with a lexon LN/plural/. The morpho-

tactics guarantees that the realization of LN/plural/ is (in

 

this case) simultaneous rather than sequentially ordered
vith respect to the root (of. the topmost downward un-
irdered and). The speaker must then decide if the plural
.s semantically-based or phonologically-based, i.e..
hether the plural is defined on semantic grounds or
honological ones). If within his/her gnostemic level
he given item has been classed as having a semantically-
ased plural. then the next ”choice" is which semantic
aature defines the class to which this item belongs. In
1e case of §ggb it is the [round] class. The exact
:oice of preemptive plural vowels in determined by the
wel pattern of the singular.

4) The simplification achieved with this analysis
a to collapse what were identified as five different

 

cla
had
her

880

the
sin

86!!

tu1

thf

 

15?

.asses semantically (all of them were [round] but each

1d a different surface form) and four different classes
arphologically into two stratificational plural morphemes.
ich sharing the configuration flfgn+v*. This configuration
? plural preemptive vowels enables us to group tOgether
iose lexical and derived (not shown) singulars having the
lngular vowel pattern I! and predict that all such items
amantically specified as [round] will take a plural with
is pattern 1;; as a factor. It is apparent that this may
a viewed as an anatactic relationship. but such an analysis
Ld not provide the simplest description. Perhaps with
beater analytical refinement. this generalization could

: captured in the description. but it could not be cap-
lred here. The various types of singulars taking part in
[is general plural class will be described individually

low.

 

 

158

 

mds+>§®

MdeQ

\ I +>+8+Q mezdnw
// §E®
II maria

(I I/
l // ‘I
’l/

 

¢12>+

a women

4@

era
@

wemswmw

‘

Figure 21+
r‘MaaL Plurals (Semantically-based) Bound

 

159

(:) The type of singular characterized by the item

p 'tube' is generally described as having a weak

le radical. i.e.. either 1 or 1. Its characteris-

surface form is Fva. e.g.. Euub 'tube'. 2112 'barrel'.

 

 

forms have here been analyzed as lexical nouns (i.e.,
derived) of the form FVWflL. where w signifies a weak
cal. V signifies a vowel alternation uoii. whose
ace alternants are contextually predictable (in this
1 before 1. 2 before 1). fl indicates a vowel ”place
er” considered a vowel by the CVCV... pattern of the

atactics. but given no phonic substantive realization.

 

 

160
The plural vowel pattern (giau+V+) will be given
:edence over the singular pattern 12 in the following
er. If the plural pattern has been selected by the
hotactics then the phonotactics will receive the singu-

form and the plural pattern simultaneously. 3ngb .
+ +
fl+at V
e the phonotactics imposes the order CVCV...(C). it

t selects c. then the precedence-taking ¢+ (in pref-

ce to E. which is realized as fl). then 1. then aa+

preference to fl). then b. In case there are only three
consonants (as is the case here). the last preemptive
L (V+) will be realized as h. Thus nouns related in

to £ngb (cuwbucuub) in the singular will have the

 

 

ll form §Qwaxbfl (cwaab). given abstractly as FMaaL.
:2) - (:) Forms such as tall 'hilr and filis 'fils'.

 

 

I having dissimilar surface forms are related to forms

tuub at a more abstract level. These forms too may

 

alyzed as having the vowel pattern _Q. Forms such as
_are here analyzed as having the singular form rv1gs
e morphemic sign level. The 1 here is the same mor-

X (hvu) present in forms like chub and 2112. The

 

nt is predictable here also by context. (Actually

the underlying (morphonic and phonemic) vowel to which
phoneme” "labial" is added in the appropriate en-
nents of labials. velars. emphatics. and r in the

nic alternation pattern.) It is here posited then

:he stratificational phonemic sign (C-phonemic) form

 

 

 

161

filis is file (the same shape as tall above) and that

 

e phonetic tactics inserts an epenthetic copy of the
rst vowel between the last two consonants of a stem if

ey are not the same and if they occur word finally

 

f. filim 'film'. a borrowing from English). At the
unique phonemic sign level where i in the environment
_C# is not distinctive. all stratificational phonemic
s in this position are realized as fl. The phonetic
:tics may then insert the anaptyctic vowel everywhere

this environment. Thus items such as tall (tglfll) and

 

;;§ (illflg) have their plurals realized in the manner
flained for £2122 under C) .

(:) There are certain nominals which take the plural
LaL. but which have a final vowel in the singular, e.g..
fig 'narghile‘ (garfisa. i.e.. a zfl_a pattern). In order
handle the plural mechanism for such nouns it is nec-
ary that the final singular vowel which denotes the
mmatically feminine be suppressed (i.e., not be realized)
"plural" is chosen. This is done in Figure 2n by having

feminine suffix as a possible suffix only in the singu-

 

configuration (of. the right branch of the topmost down-

5 ordered 2;). The morphon T is realized as P/t/ before
uffix beginning with a consonant. Elsewhere it is
lized as fl. For example. the phonotactics. according

?1%ure 2b would receive agrflg (since the feminine
n+3 1+V+

?ix -3 would not have been allowed to be realized) and

 

 

162

i generate g¢+rax+§. with V+ being realized as n since
S were only three stem consonants. Thus the plural
Sass. This guarantees that. whether the singular forms

l plural is FMaaL end in a vowel (garsa) or not (cuub).

 

; is no final vowel in the plural pattern.

g) There are certain other nominals which end in a

I and which keep this vowel in the plural. These

als represented by gamflin 'hose for narghile' pose
blem since their final vowels must somehow be kept
not from the final vowels of nouns like ggréa above
are suppressed.

In order to keep them distinct only one difference
:ded: the suffix -; attached to some forms as a deter-
sign is allowed to be realized in the environment
lral. The phonotactics generates the plural vowel

'ns of such nouns in the following manner. Receiving

iguration such as gamg§ 1 . the phonotactics would
fut-v4-
a:

the order gflfmaa+ci. V+ being realized as 5 since
are only three stem consonants. Thus the emerging
_form is gmaal; (FMaaLi).
) - There is one additional problem here. how-
kThere are nominals whose plurals are of the form

. but whose singulars not only have an i after the
Foper but also an additional a (ggzgjiiiacgaajxg

ttion channel') or sometimes also an additional ya

 

 

163

QO+i+yastquizxa 'nightstick‘). This additional materi-
lso signifies ”grammatically feminine”. The only addition-
estriction needed is a subclass governing the realiza-
. of 1 (cf. ). The subclass of nominals represented
quizxa 'nightstick' would enable the additional I to
ealized in the singular.

(:) and (:) Nominals of the form hayfiya 'snake' and
gigggg 'vulva' whose plurals are of the form FMaaLa

a different kind of problem. The singular form may

ay not have a suffix yet the plural form always does.
resolution of this problem lies in making separate

lessee of these nominals as in Figure 25.

 

 

 

 

,4!

 

 

16h

 

.3“
ha‘
A . G
.i
4‘.
4r
\
\
\\
— /
/ /
/ ///
A /// /
./ A V
/
/ ‘
/ a ‘

(:) ¢fi-‘/+
@ a”
8

Figure 25
Subclasses of FMaaL-type Plurals

Phonotactics on receiving the configuration

+
supplies the order kfligaa+§a. As before V is

as d yielding the surface form ksaasa.

 

 

 

 

 

165

The final form classified as a FMaaL-type plural
epresented by thflbag (mutbag) 'double-tubed flute'.
__a____ __a___

form actually represents two phonologically predictable

l1 alternants FMaaHiL and FHaaMuL. The phonotactics.

receiving thflbag yields the form m¢+tat+bV+g (mtaabug).
__AL___— ——e4 .—L
¢+as+V+

time because there are more than three stem consonants

 

r ultimately realized as one of the alternants 1 or g.
Lding on environment.

Exactly how the derivational vowel patterns (see the
tst left unordered 91) are assigned is not dealt with
is treatment. Examples of adjectives in this plural
are shown. Their simultaneously realized vowel

rn is 251:. where V1 is a morphon denoting the alterna-
aF°i, g in the environment k__ and 1 elsewhere.

The remainder of the semantically defined [round]

18 (eight in number) may be generalized to two

ional round] broken plural classes. They are:

Ls: (FaMaaMiiL, FiMaaMiiL, FMaaMiiL). M

aMiL, FuMaaMiL) and yjfl (FuMi, FiMi, FaMIaat).

3 26 (see u.3.1.1.2) shows how the first of these

3] allomorphs (i.e., morphemic sisHS)

1to the overall structure. Figure 27 (see u.3.1.1.3)

Lbes the second.

 

 

 

 

 

166

.1.1.2 We

We have reduced thirteen broken plural classes.

ntically defined by the component GN/round/, to

ee allomorphic (stratificational morphemic) classes

L, FvMaaMi(i)L. and FVML. Figure 26 describes the

aaMi(i)L plurals.

 

 

 

 

 

 

167

6”/mu we) /

SVda 7‘15" \ \
9V5 i 74>) \ '\
MV$¢M‘ \

n&?¢‘m :r

H7¢wa 1w ,1; V” 0+ a:+ 1'"

Figure 26
aMi(i)L Plurals (Semantically-based) Round

 

 

 

168
lownward unordered and labeled (:) defines the class
.ngulars which take FvMaaMi(i)L plurals. The enabler
:hed to this plural class enables this particular
L1 choice (FvMaaMi(i)L) to be realized.
(:) is the subclass of singulars whose plural is
.Mi(i)L FuMaaMi(i)L where the alternation inJu (V+)
onologically predictable. The generalization that
h of the final vowel of the plural is predictable
e occurrence of length with the final vowel of the
lar is captured in the following manner. Length
ice the final consonant in the morphotactics and is
ned thereto. so that length is present in both singular
Lural forms. The phonotactics will assign this length
: syllable nucleus (v) as a constituent. Thus, in

arphotactics the configuration is V :C while in the

 

:actics the configuration is !;_9. (Actually two

of abstract length will have to be set up in order
sure their separate treatment in the phonology. That
:ngth originally associated with a consonant must be
Listinot from length originally associated with a vowel,
:hough both are ultimately associated with a vowel on
mface.)

here are both masculine and feminine nouns taking
lural. An example is sVda?g; (sidaara 'common Iraqi
ar'). (The feminine suffix is realized as fl in the

nment of plural.) Its simultaneously realized plural

 

169

V+as+i+. The phonotactics would interdigitate the

emptive vowels to yield §V+danfzi+z (81958112)- Glottal

has several realizations in Iraqi. The singular fonm

 

his noun at the stratificational phononic level would

dat . The phonotactics enables ? to be realized
Length (I) in the environment before a consonant in
morphzmc alternation pattern. The plural form at the
utificational phononic level would be gidaszig since the
lotactics would enable ? to be realizad as l in the
wronment before (or after) 1 in themonmxnic alternation
ern. (It may also be realized as 1, En i, g in various
r environments.)

One thing which became evident in dealing with this
lass of plurals is that glottal stop (7) will have to
its alternants specified in the morphonic alternation
Ern above the phonotactics (i.e., before the phono-
ics assigns a point of articulation to X or !f(i.e..

e i is realized as g (in some environments) or i).

h +

zample, na? 'ua 'water wheel'. pl: nv+?a: iar.

a1 stop must first be realized as g in an ”elsewhere”
onment. Then V+ (neutralized with 1) can be

ctly predicted as 3 before 1 in the phonemic
nation pattern (nuwaa'iir).

Nouns in this subclass take the plural FMaaMiiL.
her words the first vowel is 9+. The plural of

3k (mankuuk ‘bobbin') is mnaakiik.

 

 

 

 

 

 

 

170

Class (:) is the ”elsewhere” class (i.e., the class
hose members take a: as the first vowel of the plural.
he plural of eranss (gurgggsa 'blister') is faraagiis.
ouns in all three of these subclasses ( (:) . (:) . (:) )
ave similar underlying shapes so there is no way to pre-

Lct class membership on such a basis.

,3010103 FVML‘tRBe
Figure 27 describes the FvML-plurals defined by the

toston GN/round/. (In general discussions defining
:mantic components are referred to as [round] . for ex-
ple; in theoretical descriptions these components are

ferred to as gnostons. for example, GN/round/.)

 

 

 

171

LN/Pl/

4‘...

- v/ -
/ N

® ‘ P—baSeA

\ GHA-b‘SCJ/ G’éeh'owd/
\ O
\

C)bwnn2n

\ .na

\/
A /
\ .
9 as) , \ ,
A A @A CqunJ/\ . / QT
\ \\ ' /- k1
\
\ /%
\l \ \\ //' ///
\a§¢? \ \///¢/:}
—\ an.\ A._{ ,,”’./”’

-_ - / \ /

7 \ If A /\
Uh¢ha?\$im?kd\ali3 bam¢7q?\\ A
‘\\:‘\\\\

.__

4

_.,__\ ‘

3" (1* a*' 55+

Figure 27
FvML Plurals (Semantically-based) Bound

 

 

 

 

172

<:) Here we see that the broken or internal (simulfixed)
.rals. which we have been so far discussing, form a di-
tomy with the sound or external (suffixed) plurals.
broken plurals are defined by the downward unordered
,indicating simultaneity, while the sound plurals are
ined by the downward ordered and indicating consecu-
ity.

(:) This node defines the class of FvML—plurals as e
1e. All of the members of this class take a plural
h FvML as a factor. Here we have an instance of an
bling line connecting to an enabler within the the
tics rather than below. Currently in stratificational
Dry, enabling lines are allowed to connect to enablers
V below the tactics in the ”emic" alternation pattern.
a restriction was found to be too stringent in such
as as (:> . The Iraqi data requires both semantic (or
Itotactic) conditioning (see the diamond node connected
:he gnoston GN/round/) and morphological conditioning.
urder to adequately account for the data. conditioning
in the morphotactics was necessary. In other words.
e not all items characterized by the gnoston GN/round/
any one particular plural shape, but are further
itioned by the morphotactic c1ass(es) of the singulars
which they are associated. it proved necessary to

v conditioning within the tactics of the morphology.

 

 

173

The members of the class defined by the node designated
(:) form a subclass. The surface shape of the plural is
i. As an example take the item £2123 (mragyg 'mirror').

shape of the singular with the feminine ending ya (the
sing allowed to be realized in this case) is mfl;a?z§.
phonotactics receives the 1; after the stem proper.
5 beginning again the CV sequence and allowing the
sonant cluster 31 (mfira? and ya). Glottal stop is

VCVC CV
[ized as length before a consonant. yielding the form

flag. The plural is 32:33 which the phonotactics

‘1“

u
ardigitates to yield gu+rfli . The surface form (£321)
1chieved when glottal stop is realized as 1 in the en-

)nment u.iC #. In other words, mur? is realized as

 

 

174

(:) defines the subclass FiMi. The usually predictable
i (l) alternation is not the case here. Thus two
.arate subclasses had to be set up (FuMi and FiMi).
3 the item L222: (Lag; 'aba'). The singular is 12223
:h glottal stop being realized as a in the environment
__#. In other words 'ab? is realized as :222. The
Lral is $2222 . The interdigitated form is 'i+bg!?.
1+9? .

>tta1 stop is realized as i in the environment u.iC #.

 

'ib? is realized as 'ibi.

 

(:) This subclass has the plural shape FaMLaat ~-
t is. the internal preemptive vowel sequence 2:g_: and
suffix -aat. Take bamgya? (baglaaxa 'okra') as an
mple. The singular is straightforward at this point.

plural is (pgggzgl a'é). The interdigitated form is
*T
a

ngfya?aat. It here becomes apparent that we need a

n boundary between the stem proper and suffixed material.
that this boundary must function as an environment

Lvalent to that of a consonant for the realization of 1.

>ther words, if we posit a stem boundary (+) between

1:123 and sit. we can say that Z is realized as L in the
ronment ___C,+. Then we would get the form bamxasas .
phonotactics or the phonetic tactics could insure that
one of a sequence of two long vowels is realized. A

nd choice for the solution of this problem is to con-

:t the phonotactics in such a manner as to have all

 

175

aterial after the last consonant able to be interdigitated
lth a preemptive vowel truncated (i.e., realized as t).

n this case the a: of ba+mgfya? would be truncated yielding
lth the addition of the sound plural suffix all the correct
arm bamyaa . The second solution seems to be the more
eneral case, i.e.. it seems to account for more data.

or example bantalawn 'trousers' p1: banggtiig. The ggg
are seems to have been truncated.

(:) The item here rahfiha? (rahhaaya 'grinder') is
O A I.

 

robably derived but it is the only feminine example aVaiI-

ale for this plural subclass.

 

176

3.1.2 Square Plurals
3.1.2.1 FvMaaMi‘izL-type

The semantically based plurals defined by the gnoston
’square/ are four in number: FMaaLiin. FiMaL. FMaaMiiL
1 FaMaaMiiL. The last two are also defined by the gnoston
’round/ and were discussed above. Actually there are
tare items in all of the FvMaaMiiL plurals. Figure 28 shows
r the component GN/square/ is interrelated with the

haMi(i)L plural system.

A
Gu/guare/
6%und/
' Q
in» A.» ‘5
:umffi-i.k : ‘
dsqgum 1r \ \ ’ FMmL RML FIMQL FMaaLiin
\ \ :
\ I
55.41.41 \ampazb X
SOJ9$JQ :d\
\\ \\
\\ \ AIL
\ \
\
\\ \
\ \ 4
Figure 28

'MaaMl(1)L Plurals (Semantically-based) Square

 

177

h3.1.2.2 FMaaLiin-type

The square plural FMaaLiin is morphologically similar
:0 the FhaaL round plurals and thus must be integrated
ith them. Figure 29 details this integration.

    

43.1:

; ‘Ii L
dumfaKam F MQL FVMMM

bdzyit

Figure 29
FMaaLiin Plural (Semantically-based) Square

r'

 

178

The noun dukkaan 'shop' pl: dkaakiin is subject to

 

a dual interpretation. It may be analyzed as belonging
to the plural class FHaaMiiL or to the class FMaaLiin.
If the second analysis is correct than dukkaan provides
another example of truncation of final singular material.
The noun haayit 'wall' p1: hyaatiin also presents
a____m :___4___
an interpretation problem. Normally one would assign it
an underlying shape of ha?¢yit. In that case the plural
H
should be hwaayiin. To obtain the correct plural (hyaatiin)
w %-
one must analyze the underlying form as halyit.
M
These nouns form a separate class and enable the Qfas+
alural (related to the round plurals) to be simultaneously
realized with the noun stem which is obligatorily followed
>y the sound masculine plural -ii .
For example: halyit is realized as h¢+yas+tisn
;:-:& isn 4_.____4___
as

>r hyaatiin.
4...a___

L.3.1.2.3 FlMaL-txpe

The last plural class characterized by the component
'N/square/ is FiMaL. This particular morphological shape
;as not been introduced yet. Figure 30 shows this plural.
he plural of ga§g§ (g§§§g 'ledge'L for example is g;§§§.
s usual the :a feminine ending of the singular is not
ealized in the environment of plural. All of these nouns

ave similar phonological shapes and they are all feminine.

 

179

   

Panama, Fv MaaMli L

H (1"

Figure 30
FiMaL Plural (Semantically-based) Square

14.3.1.3 Paired Body Parts Plurals

The next semantic class to be considered is that of
[paired body parts]. There are only two morphological
classes so related: FMaaLaat and FVMaaLaat where V a
UN u). The only problem is with the plural of ?ubut

___c.

'armpit'. The singular is predictable. Its underlying

Phonemic shape is ?Vbt. 1 would then be correctly pro-
“

vided as g in the phonemic alternation pattern. and would

also be copied by the phonetic tactics to prevent dissimilar

 

180
consonants from surfacing contiguously. However. the

underlying phonemic shape of the plural is ?V+ban+talt.

 

1 would then be incorrectly predicted to bepi since the
emphatic (t) is not in the immediate environment. I can
see no way at the present time to resolve this conflict
without assigning the surface vowel to the underlying
singular forms and have the plural use that vowel as its
first vowel. It seems that such an analysis would lose

much in.generality.

l1'0302l.03e1 FmfiaLa-aE-EIB:

Figure 31 details a context sensitive solution to
the [paired body parts] plurals. It captures the general-
ization that both plurals share the constituents 3;: and
Eli: and differ only with respect to their initial preemptive
vowels. On the other hand, it misses the generalization
that one of these plurals is related in shape to the
FMaaL ( [round] ) plurals.

 

181

Figure 31

FMaaLaat and FVMaaLaat Plurals
(Semantically-based) Paired Body Parts

 

':n

 

182

Figure 32 details another context-sensitive solution
which captures the generalization the first solution
failed to capture but misses the generalization the first
solution made (i.e., that both paired body parts plurals
share 2l£)- The final Judge is the simplicity meaSure

which prefers the first solution.

A
A‘L .
A
6" use:
(’N/rbu . 4M9 N
\ a
\ ¢ . s
\\ D'

\ \\\ \ A A
7MJ¢7\ rVJ4w1 ‘\\ \\ _ 51h
mpg? due: a“

A.»
‘L
v+ \\\ I
We? A a‘“
¢+ai+
Figure 32

Alternate Diagram to Figure 31

 

183
b.3.1.h

Abstract Plurals

The next semantic class described is [abstract].

There are four morphOIOgical shapes associated with this
class: FaMaaLL, FaMaLaat. ?aFHaaL and FVMuuL.

4.3.1.“.1 FaMaaLL-txge

  
 

xa7y5§¢5 \\

4—0.4"
:§:3¢s¢s \\
'a ?¢M¢W\

4+ a:*

Figure 33
FaMaaLL P1ura1 (Semantically-based) Abstract

 

— ’

18h
As an example take the item 'a7QQO (zagmgi' common
man' ). The plural is 'a?flmflm . The -L (nisba suffix3)
a Hat

is not realized. because in the environment of plural the

speaker does not generate the nisba suffix for this item.

It is a determined constituent of the singular form.

however. This is a different -i than that encountered

in the FMaaLi plurals. Actually this (Lgamml) is the same surface
form used for 'gz (paternal) uncle' (where the -i is a

pronominal suffix). It is also used as a term of respect

for any man. The phonotactics would interdigitate the

preemptive vowels to yield the form Leilalimflim. The

phononic shape would be iagagmn.

“-3-1-“-2 Esflséss£:£xss

Another [abstract] plural is FaHaLaat. Figure 3a
describes this plural. The results are fairly straight-

forward and need no further explanation. The plural of

Zlflflfl (algae 'a service') is xadamaat.

 

185

A‘L
A‘L
N
“Alma,
A \ O
\\ ‘0’
\ l /
\
A
Xialoém \ \ A
‘ 5
J0” ¢k :zn
cut
A
060."
Figure 34

FaMaLaat Plural (Semantically-based) Abstract

 

186

403010403 7aFHaaL-tzge

A third abstract] plural is ?aFMaaL. It is described
in Figure 350

 

Figure 35
?aFMaaL Plural (Semantically-based) Abstract

 

187
This plural class enables a prefix 3a- to be realized
followed by the simultaneous realization of the singular
item and the relevant part (fliéli) of the ‘round] plural
pattern. The plural of ryggg (£1211 'livelihood'). for

example, is (}a :¥£2% or ?arzaag.

“03.10“.“ FWU‘JL‘txpe
The last of the [abstract] plurals is FVMuuL.

This is described in Figure 36.

 

 

 

 

 

Figure 36
FVMuuL Plurals (Semantically-based) Abstract

 

 

188

Most abstract nouns belong to the derivational
system. 1.e., they have their singular vowel pattern
specified separately and simultaneously by the derivational
system of the morphology. For example. Egg; 'doubt' is

of the form _'

§KK a9!

Figure 37
An Example of a Noun in the Derivational System

where the root Egg can be interdigitated with other

vowel patterns to achieve verbal or adjectival stems,
e.g.. 'to doubt' and 'doubt'. and afi is the vowel pattern
for abstract nouns. Nouns like 32:29 'promise' still have
the vowel pattern 22.“ However, whether one treats them
as derivational or lexical (for simplicity of description)
Preempting and interdigitation still function in the same

manner 0

 

189

4.3.1.5 weak Plurals

There are many [wean] plurals: FaMaLa, FuMuL. FiMiL,
FuMuLaa?, FuMaLaa?. FiHLaan, FaMaaLi, FaMiiL, FaHaaMiiL,
FiMMaL. FuMMaL, and FMaaMLa. These twelve classes can
be generalized to eight stratificational morphemic signs.
however. Some of them may be integrated with plurals we
have encountered previously or will encounter on the

following pages. Figure 38 describes the plural FaMaLa.

h.3.1.5.1 FaMaLa-tzpe

0" (1*

Figure 38
FaMaLa(at) Plurals (Semantically-based) Weak

 

190

The previous figure shows how the weak plural FaMaLa
is integrated with the abstract plural FaMaLaat. The
noun kaafiz and the class it represents have a derivation-

ally determined singular vowel pattern a:i. The plural

of 52; (kaafi; 'infidel') is kfr . The same inter-
a:g
8'1 ++B
a a

digitation and preemption as before takes place yielding

the form kafara.

 

h.3.1.5.2 FvML-type

Figure 39 shows the generalization of three plurals
FuMuL. FiMiL. and FiMLaan to one stratificational morphemic
pattern: i‘g_fl. Sometimes the i‘g can be generalized
to a phonologically predictable alternation. but not in

this case.

 

191
A
O
A O
1..
AL Gwaou/
\
\ . N A out im
\ ,. .l
C) .Ala 6i)‘.t \\ //.1‘
’3’ / o
A @ um n's

@ a M‘

?

a 75 Sip“ 4”- A
~£:
: (:Z“ if; \\
m‘é \\ \ an:
El
q H U|+ 95+
Figure 39

FvML Plurals (Semantically-based) Weak

 

192

(:> The nominals in this subclass take either the plural
FiMiL or the plural FuMuL. Both plurals have a IQ pattern
since the second vowel is inserted as a copy of the first
Just in case different consonants occur contiguously
word finally.

(:) signifies a subclass under (:) which enables the
preemptive vowel 2? to be realized. Both subclasses (:)
and (:) have a simultaneous vowel sequence ea in the
singular which is derivationally specified.

(:) specifies the subclass which enables the FiMLaan
plural pattern to be realized. The FiMLaan pattern has
the same vowel sequence as the FiML pattern followed by
a suffix gig. This suffix has the same shape as the ending

on certain adjectives represented by the item (n's am)
a5

(na'saan 'sleepy').

The subclass @ comprises items like which are
derivational Lay (‘amya 'blind’). (we are using the

a!!!

feminine derivational vowel pattern here; the masculine
vowel sequence (23) bears an anatactic relationship to
the feminine (22): the masculine form is preceded by an
affix 3a-) as well as items like (:) 2922; (Lag; 'mouse')
which are lexical in nature. (Notice that they both have the

same vowel pattern. however.)

 

193

h.3.1.5.3 FquLaa?-type

Figure no generalizes two plural classes to one stratifica-
tional morphemic vowel pattern. The second vowel of
FuhuLaa? and FuMaLaa? is phonologically predictable. If
the middle and last radical provide the proper environment
a u is selected. otherwise an a is selected. For example.
the plural of haqisr 'low, base' is huqura:? (W is realized
as 2 between 1 and 1) and the plural of zabino 'troublesome'
is xubaGafl’ (where V is realized as e)-

.‘dl‘IIII-—-

\\ AL
A.»
\ ..
\ a?
\
W" A
b
x9 A
a,h
u*9;+
Figure no

FquLaa? Plurals (Semantically-based) Weak

 

 

 

191+

4.3.1.5.4 FgJiaaLi-type

Even though it appears at first glance that the
plural FaMaaLi fits neatly into the FaMaaMi(i)L plural
group. it proves simpler to set up a separate class at
this point and integrate it with plural classes encounter-
ed on the following pages, e.g., FaMaaLa, FaMaaL. FiMaaL,
FiMaaLa. FuMaaLi: and the previously described FVMaaLaat.

P‘umlg Sinfiu‘ws
G
(“v/clue!
Pp. ;
® ‘ \\o g / /
CD A \ \ / 0
\ /
\ // ‘ /
@A \ / % f, A r aT‘
been“ «\/ .
l
4 wt
da?¢k': \
V* W 07 az+
Figure 1&1

FaMaaLi Plural (Semantically-based) Weak

 

 

 

———————————————————————————————————————————____________T:::]IIIIIIIII"

195

Figure 41 illustrates the integration of the FaMaaLi
[weak] plural with the FVMaaLaat [paired body part] plural
encountered previously.

(:) and.(:> define the FaMaaLi class as a whole with

the general internal vowel sequence v+a:+ and the suffix 1i.

 

(:> signifies the subclass represented by balflg

(balwa 'affliction'). This subclass enables the vowel a:

 

to be realized as the first vowel of the plural vowel

sequence. The plural of balwa is Mlflg :)orm1§awi.

8H8!

 

(:) signifies the subclass represented by da?gh;
(daahiya 'disaster'). This subclass presents a small
problem. In order to capture the generalization that the i
is preserved in the plural. it is necessary to have this
subclass not allow the suffix -1 to be realized. The

81nsular is @72hiza where y is permitted to be realized.

The plural 18<m ) or dawaahi.

 

h.3.l.5.5 FVMHaL-type
The next [week] plural we will discuss is the FiMMalrv

FuMMaL plural. It is described in Figure 42.

    

4NQiH'

KVdi :§ \

mpdih’ve
(uufltha

Figure #2
FVMMaL Plurals (Semantically-based) Weak

This is a rather unusual plural class in that it
employs not only simulfixed vowels but also reduplication
of the middle radical. This phenomenon of reduplication
is common in derivation as a realization of causation and
repetition.

The morphon MN/*/ is realized as a construction choice

in the phonology. Figure 43 describes this phenomenon.

  
 

“Faunv017\CTIC£5

Figure 43
The Phonotactic Realization of the Morphon MN/*/

If the morphology has generated the morphon MN/*/
then the immediately preceding consonant. repetitively
reduplicated. is allowed to be realized. For example,
if we take the plural of thi:r (futiir 'foolish').

..._r___.. .4...

[thi:r) , we get fV+ t£a+ r. The ultimate form is.

V+*a+
CV CCV C

of course, futtar.
—_l..l—-—

4.3010506 FMaaflé-txpe
Another [weak] plural is FMaaMLa. It is described
in Figure 44.

plural s SICNTAIAY‘S
A A

 

Figure 44
FMaaMLa Plural (Semantically-based) Weak

 

199

The subclass of nominals enabling this plural to be
realized also enables the realization of a prefix 31 (gr)
which is present in both the singulr and plural forms.
The plural vowel pattern is 3;! plus a suffix -g. The
plural of mgaggi (the reduplication in the singular forms
has not been detailed) is (ml 35% a) (M).

a:

“03010507 FaMllL-Exge
FaMiiL is yet another [weak] plural class. It
is described in Figure 45. The plural of hmaar 'donkey'
s....
is hamiir.
4......

 

Figure 45
FaMiiL Plural (Semantically-based) Weak

 

 

200
4.3.1.5.8 FaggahiiL-type
The last of the [weak] plurals is FaMaaMiiL.
We have encountered this plural previously. Figure 46
shows the integration of the Beak] with the [rounzfl and square]
plurals of the same morphological shape. The plural of

'asfl'u:s ('as'uus 'coccyx‘) is 'asaa'iis.
l 4.4 “-

 

  
 

'a§¢‘u :s
{d'plu :K

5* w a" 01+ 1+

Figure 46
FaHaaMiiL Plural (Semantically-based) Weak

 

201

4.3.1.6 Color Plurals

The next semantically defined group of plurals is
[polor]. There are two color plurals, FuMuL and FiMiL
but they have already been described underiweaklandlrouna.

“03-19601 hull-£22:
Figure 47 describes the integration of Eolor with

those plurals owaeakland‘roundltaking the same morphological

shape.

 

lfi’“ V+ ¢+

Figure 4?
FvHL Plurals (Semantically-based) Color

 

r,‘

 

 

202

While the weak and round alternants were not phonolog-
ically predictable, the color plural alternant. with one
exception. is. so the color S-class enables the E: to be
realized. The plural of xadra 'green f.’ is xudur while

..... ..a..
the plural of malha 'grey f.’ is milih. The epenthetic
...4.. ._._i
copy vowel is predictably inserted in case two different

consonants occur finally.

4.3.1.? Strong Plurals

The strong class contains very many plural shapes:
FuMaLaa?, FuMuLaa?, FiHaaL. ?thiLaa?. FuMuuL, FuMuuLaat,
FiMuuL, FaMaaMiL. FaMaaMiLa, FaMaaMiiL. and FaMaaMuL.
These thirteen classes can be analyzed as five stratifica-

tional morphemic classes.

[403010701 FuMVLaa?-t!Re
The first of these is FquIaa? described in Figure 48.

This is the same plural as previously described under
' ' is hukumaa?

weak plurals. The plural of hakiim wise

while that of najiib 'noble' is nujabaa .

 

 

203

A.
A ‘
A‘l
(”a /
5 Can
A \\ ({Te/ ‘
\\ 4" 4'
A \ \\__
\
\\
At
gm.
KJE a.
.A
At
61:?
u*"*
Figure 48

FquLaa? Plurals (Semantically-based) Strong

l+-3.1.7.2 FVheeL-type

There are three plural classes subsumed under this“
general morphemic shape: FiMaaL, FiMaaLa and FuMaaLi.
Figure 49 describes these plurals. Each class shares

the same simulfix vowel sequence V§:. They differ only

204

with respect to their suffixes. The plural of ggxgy
(mizxg 'hundred'), for example. is man:y; (miyaazi).

A
O
A
l
\\~ 47
\
A A \ \‘ //
Maw) MVY¢Y
AIL |
VJr a°+
Figure 49

FVMaaL Plurals (Semantically-based) Strong

 

 

 

205

”03910703 78FM1Laa?-E!Ee
The next [strong] plural is ?aFMiLaa?. Figure 50
describes it. 1

' I
A .
“‘ 6¢Qnmj/
/ \ ‘
”// A‘ ‘\ ‘7
\Wfi—
\ ._ __ __ ..
\.
\
1‘. \\‘_
AL ‘.L A.

flat ‘ 3W * A
CH: 035 523*}?

Figure 50
?aFMiLaa? Plural (Semantically-based) Strong

 

_

206
Both nouns with the derivational vowel sequence 5!
and those with sequence EL; are found in this plural class.

The plural of tbb (tabiib 'doctor') is ?atbibaa?.
a

4.3.1.7.4 FVMuuL-type

The fourth type of [strong] plural is FVMuuL. The
morphological shapes related under this class are FihuuL.
FuMuuL and FuMuuLaat. The first vowel is the phonologic-
ally predictable inJu alternation characterized by 1:.
This plural is described in Figure 51.

   

hVééw
stoh stySl

Figure 51
FVMuuL Plurals (Semantically—based) Strong

 

-A

 

 

v———————————————————————————————————————————_______iZZZJIIIIIIIIIIIIIIIIEEFV
207

The plural of thgn (husin 'fort') is hV+su:+n (husuun)
..... a.4_.. i..4_... 4.4...

while that of gyjfln (filJAB 'prison') is §V+Ju:+n (sijuun).
The vowel insertion for the singular presents a problem.
The two most frequent inserted vowels are the usual copy
of the first vowel and a non-alternating ;. Sometimes,
however. one finds a u in the proper environment. In
order to obtain the correct results while maintaining
the epenthesis generalization. the phonetic tactics will

have to be rendered more complex.

“03010705 FamaaMVL-t!2e
The last set of strong] plurals contains the mor-

phological shapes: FaMaaMiL. FaMaaMiLa. FaMaaMuL and
FaMaaMiiL. These plurals are described in Figure 52.

 

208

A
GN/SMnfl/ '
\.
\ 9
\\
\\
A \.. I .—
'd “‘ ,,.»/ " ”
spKar A f
\ \ on 5am
\\r\‘\\
”Mihfithr ‘\. \:>
\\
\\ \\ .
(1+

2+ u+ a+a:+ i”:

Figure 52
FaMaaMVL Plurals (Semantically-based) Strong

While all the members of the FaHaaMuL plural class
Provide the correct environment for the choice of g:
as the final vowel of the plural vowel sequence. this
vowel cannot be 3: since many of the items taking the

FaMaaMiL plural also provide the correct environment.

 

 

 

209

Thus the if—u+ dichotomy was set up. The plural of

na7¢de (naadum 'dam') is nawaadum of 'as he ' ks ' '
,_3 _£_s Les—r. army )

I 0 O

agaakizg of qudsain (qursaan pirate ) qaraasina: of

 

masfihusr (mashuur 'celebrity’) masaahiir.

 

“03.1.8 Agent P1E3;8

There are six [agent] plural classes: FuMMaaL.
FiMMaaL, FuMaat. FuMaLaa7, FaMaaMiLa and FMaaMiiL. Since
the FuMMaaL and FiHMaaL plurals are phonologically pre-
dictable alternants. there are actually only five different

morphological shapes represented.

4.3.1.8.1 FVEEEQL-txgg

Figure 53 describes this plural. The plural of

W' (waa'id-' reacher') is " while that of syh
3% . P M an

(saayih 'traveler') is siyyaah. The morphon MN/*/ enables
“ g
the C (reduplicated consonant) construction to be

realized in the phonology. (See Figure 43. p. 197.)

 

210

 

Figure 53
FVMMaaL Plurals (Semantically-based) Agent

“03.1.8.2 FuMaaE-tyge
The FuMaat plural is another [agent] plural. The

roots belonging to this class are biconsonanta1.5 The
Singular items all have a determined -; suffix. while the
Plural has a determined -§at suffix. The description

is given in Figure 54.

211

Plurals
A
A
C’Veuf/
A\ d 41
0 :
\\_._.__ /’
1“ \\
\
5‘9 a- '
“1'
Figure 5h

FuMaat Plural (Semantically-based) Agent

The plural of Jasbi (Jaabi 'collector') is jgp2§_.

h.3.1.8.3 FuMaLaa'I-tme

This plural is defined also by the gnostons GN/strong/

and GN/weak/. Figure 55 shows the integration.

 

 

212

“lager/1
A \\ 9 e,
\ — \ \i
A \ \
\
A;
WZJ‘ A
a i.
0:?
A
ur"+
Figure 55

FuMaLaa? Plural (Semantically-based) Agent

The plural of gazing (waziir 'minister') is 122a3327.

“ad-8.1+ mm
It is evident that this plural can be integrated
with the FaMaaMVL strong plurals and also with the FaMaaL-

 

 

213
type plurals. (We have only discussed the FaMaaLi weak
plurals in this category). Figure 56 details this in-
tegration.

uv¢g¢1zk. \\
Wu1|¢’7¢K;

fligiwgés

CV06r V+

Figure 56
FaMaaMiLa Plural (Semantically-based) Agent

 

 

21a

The plural of qurflsaan (qursaan 'pirate') is qaraasina.
Q 0 :

 

The glottal stop in gallg 'angel' and giglg 'clergyman' is
peculiar. It is needed in order to insure the realization
of E: (1) in the plural forms (of. malaa71ka and gasggwisa)
but where one predicts a x realization. a 1 and 1 appear.
In the singular the realization is neutralized with d

and the phonetic tactics produces the correct results

by inserting the epenthetic capy vowel for gala; and by

not inserting anything where the final consonants are the

same (of. giss).

“0301.805 F M 111-1'; 0
This plural. the last of the [agent] plurals. is also

characterized by squar§.[roung, and Enstrumeng. It has
previously been integrated under b.3.1.2.1. Here another
diagram is unnecessary since the only addition to the
diagram cited above would be a class of items including
muxgtaaz (mgtaar 'mukhtar') and rawguzs (:aaggug 'dancer')
enabling the g: as the first preemptive vowel. This class
is a subclass of those items enabling length (3) to be

realized.

h.3.1.9 Instrument Plurals

The instrument plurals are three in numbers FaMaaMiL,
FMaaMiiL and ?aFMiIam The first two have already been

discussed in some detail under u.3.1.7 and 4.3.1.8.

 

 

215

“-3-1-9-1 W
The FaMaaMiL-type includes such items as xanflja;

(xanjaz 'dagger') and gargbas (£3122; 'reel').

The FMaaMiiL—type includes such items as migflgass

(mingaag 'tweezers') and maggkuik (mankuuk 'bobbin').
Note that length is predictably preserved in the final

vowel of the plural.

“03.1.9.3 7aFM3La-tlpe
The 7aFMiLa plural is diagrammed in Figure 57. It

is integrated with the 7aFMiLaa? strong plural. The

plural of aflnash (Jnaah 'wing’) is 7ajniha.
—_—__L .____a .____L_

A

 

¢+i+ 0:?

Figure 57
?aFMiLa Plural (Semantically-based) Instrument

216
h.3.1.10 Location Plurals
The two location plurals are FaMaaLi and FaMaaMiL
both of which have been discussed with respect to other

plural classes. They are detailed below.

b.3.1.10.1 FaMaaLi-type

Both FaMaaLi and FaMaaMiL can be integrated using
the same technique as was used with the FMaaL-type round

plurals. Figure 58 details the integration.

 

ma +al¢laxa: \
A A Kanns / '1 A //
Gib ‘I’ w— " .
A ‘
\\ A.L
tabs 9“ sat.» \\
0* w [Jr
Figure 58

FaMaaLi Plurals (Semantically-based) Location

 

 

217

The a and 2 forms appear on the surface as the same
plural shape .

The gnoston GN/location/ enables both the may prefix.
and the particular plural class to be realized. (2&-
does not occur with all location items. however.) Bi-
(with gay) and tri-consonantal surface forms take the
FaMaaLi plural (a subtype of FaMaaHiL) while tri- (with gay)
and quadri—consonantal surface forms take the FaMaaMiL
plural. It seems that perhaps the general ”locative“
singular pattern was 22 making the biconsonantal stems
(i.e., underlying triconsonantal with final weak radical)
appear to have a feminine suffix. and the general plural
pattern was g;i_making the biconsonantal stems appear to
have the suffix ~L. The ma- locative prefix was constant.

While it would be simpler if one were analyzing only
this one plural to treat it as described above. it proves
simpler when viewing the total plural system to integrate
this plural with previous analyses. such as FaMaaMuL, FMaaMiL.

etc.

4.3.1.11 hale Plurals
There are four [male] plurals: FuhLaan. FuMLa.

FMuuLa and FiMaaMiiL. The morphological shape of the
first three has not previously been discussed. The first
two (FuMLaan and FuMLa) are morphologically related. All

of these plurals are discussed below.

218

“03.1.1101 FuML-txge
This type comprises two plurals FuMLaan and FuMLa.
They are described in Figure 59.

¢
.‘L
1.;
‘ma/
4
\\\\ "
\ ‘ / /
/ / /
A A / / / /
\ __ \~ /
AI.
73&? A
V
b ~0r3 EEE
2;;— 0:1 a3; G'Tl a
u" 95+
Figure 59

FuML Plurals (Semantically-based) Male

219
The class of items enabling the FuMLaan plural to
be realized is represented by 2;; (faaris 'knight').
Its plural is £2;saag. An examgl: of the items representing
the FuMLa plural is 3; ? (15; 'brother'). The glottal
stop is realized as fiainthasingular in the environment
C___# and as g in the plural in the environment before

the -g suffix. The plural is 3213;.

“030101102 quuLa-txae
This male plural type is morphologically related

to one of the female types and will be integrated under

4.3.1.12.2. Figure 60 describes this plural.

 

{6* uzf

Figure 60
FMuuLa Plural (Semantically-based) Male

 

220

This plural is enabled to be realized by singular
items of the general form FdeL. e.g.. gibgb (zibb 'penis')
and M (m; 'bastard‘ ). The expected epenthetic
insertion occurs when the final consonants are not the

same. The plural of 21be is zbuuba.

 

4.3.1.11.3 FiflfliiL-tne

This plural class has been described previously
under the round plurals (of. 4.3.1.1.2). Figure 61
details the integration of the male variant. The plural
of 17m (aflgasnsfiigaan 'neighbor') is Jiyaariin.

a:

 

 

or j?!” :H

5an
with no Cond'ctionin5
en 43am») lenaJrh

 

 

 

 

fl} Vf an!“ if :

Figure 61
FiMaaMiiL Plural (Semantically-based) Male

j

 

221
4.3.1.12 Female Plurals
There is only one [femalel plural: FMuuL. It
can be integrated with one of the [male plurals (FMuuLa)
which is morphologically similar. Figure 62 shows the

integration.

4.3.1.1201 FMuuL¢the

 

Figure 62
FMuuL Plural (Semantically-based) Female

 

222

This If poses the same problem as encountered before.
Sometimes the phonetic tactics inserts the expected copy of
the first vowel in the correct environment: sometimes a
non-alternating i, and sometimes 1 whose alternants are .1233-

The plural of 9.51119 (La—1359. 'female breast') is nhuud.

 

There are two [plant] plurals u FVMaaLaat and FMuuL .

both of which can be integrated with previously encountered

plurals. e.g.. FMaaLaat and FMuuLa.

“0301.13e1 FVMBaIaat-tzae
The FVMaaLaat-type is described in Figure 63. The

plural of :2th (nabta ’plant') is papaataa .

 

\ 91./lam” “fr- bon
\ ' ’ par +/

A
\
\ -
\ “~
nab/t \*~.,

A‘L
a: t

A

f—
Figure 63

FVMaaLaat Plural (Semantically-based) Plant

4.3.1.13“? FHuuL-tzpe
The second of the [plant] plurals and the final
semantically based plural is FMuuL. It is the same

moI'Phological shape as the female plural. Figure 64

 

 

 

224

shows the integration. The plural of hirfls (hiris 'plant')
is hguus.

 

 

 

¢V'u:*

Figure 64
FMuuL Plural (Semantically-based) Plant

The plurals not discussed were either of an undeter-
mined common semantic characteristic or did not comprise

a major semantic class. These plurals would be integrated

 

44_____________fIIlIlllllllllllllllll'iiiifiw

225
and diagrammed in the same manner as the plurals presented

herein in a fuller treatment.

4.3.1.14 Semantically-based Plurals -- Integrated
This diagram describes not only an integration of

all the semantically based plurals detailed above. but
also the phenomenon of multiple plural association. That
is. some items will be members of more than one class.
Figures 65 - 75 present the overall morphological des-
cription of semantically based broken plurals in Iraqi
Arabic. Due to the size and complexity of the overall
diagram. it had to be presented in sections. End points
(or terminals) of one diagram can be matched with begin-

ning points on succeeding diagrams.

 

 

 

Figure 65
S-based Plurals Integrated

  

6”” o ca+i on/
I .

 

Figure 66
Prefixes

 

 

228

      

. s ageij
01136 as ms. flame?
\ A+$+ 52‘. I
A¢$+u6 233V
352/1 . . A#,:+> 136?
A+ +556 w\m&mfix +us$w£dwzz
+uE+mc_svr/
r.
3.19:. 3. ts”. <
«+61. > ”as“? “(Magi ‘
r s +6 .
sassy. $«+....LS Listed. ‘ 016.7% XN>L Musttseva. §*1.X
, .3, $3., 4 3, +33?
‘ { ‘ { ‘ numb Anne, 1‘
‘ ‘
AoPcL
.4qu sum.»

Figure 67
Morphotactic Classes of Singulars (a)

 

   

«1&s. "6*
«1 +s2+sei :,x> (3v. him?
fiSSoLV I
,. £33,
a+>+u +a
fisspwz
r... 35 9 as.
9 matey, 9.3%.»8.
& 4 ,
n+ia§oz
4.

 

{a are; are
$1.52 Q‘s/{Oz
« > s .3. . WNW» e
was at is , l, c,
/ +5§<£+nt~vl
{ m
«afiwr as? m50 L% km
. r. a >
3.37... 4 Ce, as a are,
$3 4 Sewage {$539+ 9s.“ 4
+<erux AKVLRG ujiw/
A+>+.s+6 c.5300 I
4 tarsv/ 4 4
3, 4
4‘
means.
/ 9:43?“— 03V

Sufi

     

Figure 68
Morphotactic Classes of Singulars (b)

 

 

 

a a. . sec
m > x was
A+ ..+:.< + if”) dQJ‘ns. M «+. V 5* as», vflJdi
are: +01%; A+.:+> «to +HW>eu { A+ns+>
fa, fer, a i .31
< .3 V, y,
fl+>¥5 $03 ‘ ‘ A vus
:. 9.th .3.” 1 memes
”A ”3% sumas‘ { AA u:++ Sissies. Aivu 3+$ { +s+..
2 Wusswmv/ «26V / ++e oivxoiuwx +£s1w/ «Y! a U
.4 14‘ p¢ws
aiucl ‘
‘ ‘
Anudv / AWNV
{

Figure 68a
Morphotactie Classes of Singulars (c)

 

 

+:.+U+us+> Jeans its .6 “3;

“‘(‘y

231

+9 +5

#&.to+w

+6.L

I .‘

 

+ . +6 +6 +> +$ +> +u6+§

‘

+fl +>

 

\L c\8\20

Figure 69
Corresponding Plural Classes (a)

 

 

232

ia.meie+>.e “ .ts+w +3 1,

 

+.. .45
+is+e +~ +5
x +6
11“. 11“ As ,
_ >+5
+fl+3 +>+5+6 +U +> «T +& +: L“: +fi+u6+> 1.23 +> +6 in +K+HQ+Q 1‘ +§
dfil<l A's AI. .A‘ Ads A‘. 1|. AI‘ 1" AI.
1“
ADV \
bats?
\io.r*cuo\5

Figure 70
Corresponding Plural Classes (b)

 

 

Figure 71
Adjectives

 

 

6%."...84 /

 

Figure 72
Major Sound Plurals

O
(E) (e) (0)
AL
‘
a t I: n a V1 '
Figure 73

Suffixes

(G)

 

£>.______4[
‘\\\\\

 

 

Figure 74
Pronominal Suffixes

 

 

 

Figure 75
Feminine Suffix

 

44______—___________:III-IIIIIIIIIIIIIIIIII"

238

4.3.2 Phonologicglly-based Plurals

In almost every plural class there are items which
do not possess the semantic characteristic common to the
majority of the members. They appear rather to have been
assigned on the basis of phonological shape. This section
details such items grouping them in order that predic-

tions might be made based on the canonical form of the

 

 

 

 

 

 

singular.
Singular Plurals Freguency (i.e., number
of items in
FvMMvL FaMaaMiL 126 the data as-
mm. 9 arms?
FaMaaMiiL 3
FaMaaMuL 3
FMaaMiiL l
FvMaaMiiL 1
FvML (fem) FvMaL 153
(mass) 7aFMaaL 95
FMuuL 29
FMaaL 19
FaMaaLi 25
FMaaLa 12
FuMuL 12
FMaaLi 12
FuMuuL 8
FiMLaan 4
FiMiL 4
FaMaaLi 4
?aFM1La 3
FMuuLa 3
?aFMuL 3
FaMaaMiL 3

Singglar
FvML.

FvavL
(i.e.,

(i.e.,

(1080.

(1080.

FanL

cont'd

WsV?)

WBV?)

vvsv?)
WSV?)

Plurgg

FiMaaL
FiMuuL
?aFMiLaa?
FiMLaat
FMaaMiL
FiMi
FuMaaLi

FaMaaMiL
?aFMiLa
FuMuL
?aFiMLa
FiMaaMiL
FiMLaan
FuMuLaa?
7aFMaaL
FuMaaMiL
FMaaMiL
?aFMiLaa?
FaMaLa
FuMLaan
FiMLaat
FMaaLa

?aFMaaL
FMuuL
FaMaaLa
FuMuuL
FMaaL
FuMuuLaat
7aFMuL
FiMuuL

Freguencr

.p
ta H r4 H +4 N nit» N M)-P s'xacn N H r4 H thv N

H
'\)

H t4 t4 c--: U\tm .

 

 

maids;
FvMaL

FvavL

(1090' WSV?)

(1.90, WxV?)

FVMMVVL

(i.e., last 0

omitted)

.FMva

(1000. Vii-2V?)

(lees. W3V?)

FMVMMiL

FvaL

(i.e., VVaV?)

240

P1ura1§
?aFMaaL
FiMaaL
FMaaLa

FMaaMiL
FaMaLa
FiMLaan
FaMaaMiiL
?aFMaaL

FaMaaMiiL
FMaaMiiL
FiMaaMiiL
FaMaaHiL
FMaaMLa
FuMaaMiiL
FaMaaLa
FaMaaLi
FMaaLa

FMaaMiL
FuMuL
FuMLaan
FiMaaMiL
FiMLaan

FMaaMLa
FMaaL

FaMaaMiL

 

Ereguencz
4
1
1

Id H h‘lv a

22
16

HHNNme‘

wawfl

H

241

The above facts would be integrated into the broken
plurals in the following manner. The less frequent of the
plurals for a particular singular phonological shape will
be the marked preemptive pattern. enabled to be realized
by the few items comprising each class which takes a par-
ticular marked plural. The most frequently occurring (i.e.,
the expected) plural for a particular phonological shape
will be the unmarked plural for that phonological shape.

Figures 76-78 detail the phonologically-based plurals.

4.3.2.1 Plurals of the ghonological Shape EzMMvL

4.3.2.1.1 FaMaaMiL and FaMaaMiiL-trpeg
What Figure 76 details is the unmarked nature of the

plural FaHaaMiL. thus explaining its relatively frequent
occurrence as the plural for the singular shape thMvL.
The choice of Q: or I: as the first preemptive vowel in
this series is more marked than the choice of 5:. The

final vowel is either the marked choice 3: or the un-

 

marked 1: with "long" r being more marked than "short” l-
If one takes the least marked path of the plurals for the
singulars of the shape FvMMvL ( ) the resultant plural

is FaMaaMiL. For example, the plural of laglag 'stork' is

lagaalig.

242

A
A
601W

A“.
\\
'\
\

\
\
10> .‘L \\ \\
(:> \\ .‘L \\ \
’ \ \\ \
\ \
e to \ \\ .
\ \ x,
\\ \+/ u+AI
\\_${/V*a+ 49‘ <
‘+
I
Figure 76

FaMaaMiL Plural (Phonologically-based) FvMMvL

 

213

4.3.2.2 Plurals for the Phonological Shape FvML
4.3.2.2.1 FvMaL-tme

This plural type poses a slight problem since the
first vowel of the plural is the same as the first vowel
of the particular singular item taking this plural.
Figure 77 details this plural type.

 

Figure 77
FvMaL Plural (Phonologically-based) FvML

 

 

 

244
The blank (_) is not a preemptive. Instead the

normal vowel takes precedence over it. This is detailed
in the phonotactics. The plural of nisgb (nigba 'relation-
ship') is nisab. All of the singular items in this class

 

are grammatically feminine. i.e.. they have a suffix -3.

“0302.202 7aFMaaL-txpe
The second most frequent plural type for the singular

phonological shape FvML is ?aFMaaL. This plural is de-
tailed in Figure 78. All of the items entering into this

plural type are grammatically masculine.

A
II
. 1‘ C”é+xsa£/
1.. “L ‘ii I'
A \ \_ 1 \
P‘") \
qu¢§
? a his¢t>
4.. AL
¢+a:+ O\+
Figure 78

7aFMaaL Plural (Phonologically-based) FvML

 

 

2l&5

The remainder of the plurals for the phonological
shape FvML are diverse but may be grouped under four
major plural types: FaMaaMiL, FvMuuL. FiML and FHvL.
These are the marked types for this particular singular
phonological shape and would be integrated as choices
under the first line of a downward ordered 2;. Particular
nominal subclasses would enable certain plural subclasses

within these types to be realized.

4.3.2.3 0the ho 010 all -ba ed P1 ls

The remainder of the phonologically-based plurals
would be integrated as above with the most frequently
occurring being the unmarked choice for the particular
singular shape. Since they are quite straightforward
they need not be detailed here.

4.3.3 Summary
This section (4.3). in dealing with the morphology

of both the semantically-based and phonologically-based
broken plurals of Iraqi Arabic, has been complete insofar
as possible and insofar as the data studied are concerned.
Due to time and space limitations a few of the descriptions
were succinct. leaving minor details and previously ex-

plained material to be filled in by the reader.

4.4 The Phonotactics

 

The phonotactics must account for the following
phenomena in order to be compatable with the suggested
treatment of plurals: interdigitation of consonants and

vowels. preemption of singular vowels by plural vowels. and

 

 

 

246

syllable structure. Since there are no consonant clusters
at the phonological level. the phonotactics does not have
to account for them. Consonant clusters may however appear
on the phonemic sign level as a result of the empty realiza-
tion of t vowels. For example. PS/bflussha/ 'with her
money' from gg 'with' followed by 351313 'money' followed

by fig ‘her'.
CV

A preview of the general structure of the phonotactics
has been given in discussions in the section of morphology
(of. 4.3.1.5.5) Since it is the responsibility of the
phonotactics of Iraqi Arabic to account for syllable
structure. interdigitation and preempting. we must structure
our model so as to include these phenomena. Figures 79-84

detail the phonotactics in general.

 

 

 

24?

 

I
Q
(a)
o t '
Lu
Yo
@ O 0 ~
(u) (a)
Lb Fr

Figure 79
The Phonotactics

 

 

 

Figure 80
Liquids and Glides

 

 

249

 

Figure 81
Vowels (Preemptive. Long and "Normal")

250

@

o
O 0 I

I- o ’ I 4' 4r 0 0 0

(e a) 01> <9) . ’

Lb Fr C" SF

‘7 «’

E2. Af’ 7>kr

 

Figure 82
Fricatives

 

 

251

 

 

 

Figure 83
Stops

 

 

(M)

252

(K)

Figure 84
Nasals

N5

 

253

The diagram above (Figure 81) accounts for two
different types of syllable structure: a long vowel
alone, e.g.. 2; ‘and'. 1; 'yes'. or the now familiar
CV(CV)...(C) pattern. With the CV(CV)...(C) pattern
it thereby accounts also for interdigitation since it
can only accept a series of consonants and vowels in
that order. Thus. receiving simultaneously the
sequences £22 and 15;, the phonotactics can only supply
the order kitasb 'book'.

The phenomenon of preemption is accounted for by
the downward ordered 2r over the vowels. The marked
series are the preemptive vowels. That is. first the
tactics checks to see of any preemptive vowels have been
generated by the morphotactics. If so. then the
preemptive vowel is realized and the vowel it preempts
is realized as fl (in the morphonic alternation pattern
above the phonotactics). Only if there is no preemptive
vowel does the “normal" vowel get realized. The
preemptive vowels neutralize with their ”normal"
counterparts below the phonotactics in the phonemic
alternation pattern. The case of non-preemption (i.e..

the blank (_)) is handled as in Figure 85.

 

 

254

MN/f/

Preamf+th

*0

to
normal " voWels

Figure 85
Phonotactic Realization of the Morphon MN/_/

The tactics first checks to see if MN/__/ has been
generated: if so. instead of a preemptive vowel. the
"normal" vowel takes precedence. Otherwise the tactics

functions as described above. An example is the FVMaL

 

 

255
plural. The plural of xidma 'a service' is xidflm .
+
_a

 

The phonotactics. on receiving this configuration. allows
xrda+m. The blank insures that the singular vowel
takes precedence. Thus the plural r;___.

The phenomenon of consonantal reduplication (or
gemination) has been discussed previously (cf. 4.3.1.5.5).
but we will reiterate it here. If the morphology has
generated MN/*/, then a reduplicative construction
in the phonology is allowed to be realized. See page
197 for figure and examples. This is also illustrated
in Figure 79 where the MN/*/ enables the reduplicated

consonant construction to be realized.

4.5 Intervenrgg Realrzatronal Phenomegg

Some of the realizational phenomena which were
mentioned informally in morphological discussions are
formalized here. He mentioned before that the preemptive
vowels neutralize with their "normal" counterparts. This
configuration is part of the phonemic alternation pattern

(below the phonotactics). Figure 86 details this.

 

 

 

256

_

. \
\

o a a“. .9'
‘ LhVo

/
a

u I
Figure 86
Alternation and Neutralization of the "Normal"

and Preemptive Vowels
Figure 86 also shows that if the phonotactics

does not accomodate y: (in case there are less than
four radicals). then it is realized as t in the morphonic
alternation pattern. That is, y: in the plural pattern
£211: is only realized when there are four radicals.
There are three additional alternation phenomena.
One of these occurs in the morphonic alternation pattern

(above the phonotactics) and is detailed in Figure 8?.
This figure describes the alternate realizations of glottal

 

257

6
stop: length (s), frontality (y). labiality (w), and
7
glottality (?). The environments are stated rather than

diagrammed: they would occur as part of the phonotactics.

MN/7/ marl—e; 0.8
# m, excefhoml

l a,»

PHoumvmtns

 

Figure 87
Alternate Bealizations of Glottal Stop

The second alternation is that of‘ggii or 1 described
in Figure 88. It occurs in the phonemic alternation pat-
tern. The environment forms a ”natural class" in the sense
that all of the items enabling £2 (labial) to be realized

are performed with a raiSed tongue back as is the vowel g.

 

 

[
258
This alternation must occur in the phonemic alternation
pattern (below the phonotactics) since the specification
of 1 as i or 2,must succeed u.e.. occur lower in the

realizational chain than) the realization of glottal stOp

as x or g. (If this were not the case we could handle

 

the alternation in the morphonic alternation pattern as

in Figure 89.

PuoucrrACchs

Pfiouevurc.
ALTsauAnou
PA'n'eRN

 

Figure 88
Alternate Realizations of 1

 

 

 

Figure 89
Possible Alternate Analysis of 1

The third alternation phenomenon occurs in the

phonemic alternation pattern. This phenomenon. the realiza-

tion of glottal stop as 19 (a). must be spec-
ified as occurring in the environment Of ¢C #, as in

the phonotactics consonants do not occur contiguously.

 

 

 

260

/
WONEMtC. / / / '
ALTERNA‘I‘XON
PATTERN

v0 (5'

Figure 90

Realization of Glottal Stop as a

Another phenomenon which occurs in the phonemic
alternation pattern is the related phenomenon of the
realization of 1 and 1 as g and 1. respectively, in the
environment ¢C____#. This entails only the generation
of 12 as a determined element in this environment. Note
that all the so-called weak radicals (1.1.1) are specified
as Vo(calic) in this environment (¢C___#). An example

of this is P/filw/ 'colt'. The phonemic sign realiza-

tion of this item is zilu.

 

 

 

 

261

u.6 The Phonetic Tactics

Since the phonetic tactics has never been formalized,
the types of phenomena it handles with respect to our
analysis of broken plurals will only be explained
verbally.

There are two related phenomena which it is assumed
that the phonetic tactics handles. One is the insertion
of a determined vowel (1) between the first two consonants
of a three consonant cluster. The other is the insertion
of a determined vowel (usually ;, sometimes 3 or a) between
two different consonants which occur phonological word
finally.

The first is a quite general Arabic phenomenon (i.e..
it occurs throughout the dialects). An example of it
is PS/bf1ussak/ 'with your money'. realized on the sur-

face as bifluusak. Often speakers of Arabic as a native

 

language insert this vowel in English words containing
triconsonantal clusters. for example, [lipistik] ‘lipstick'.
The second phenomenon is a characteristic of Iraqi
Arabic but not of all the dialects. e.g.. not of Egyptian
Arabic. It is the vowel insertion between dissimilar
consonants phonological word finally. for example, filim
(Iraqi) versus film (Egyptian). -—-_-
The phonetic tactics also handles phenomena like
phonetic assimilation, but this has no real bearing on

the general problem of broken plurals.

 

 

262
u.7 Implications for the getatheory

The major implication for the metatheory which this
study has made is that language is a much more integrated
and much less well stratified system of relationships
than it was previously thought to be.

While stratification is still an important concept
upon which to base the theory, the requirement that units
on one level relate only to the units of the 1evel(s)
immediately contiguous is too strong. It needs to be
made less stringent, in order to allow relationships
between units on non-contiguous levels. This study has
utilized the less stringent requirement in the analysis
of Iraqi broken plurals. It has posited a direct relation-
ship between portions of the gnostotactics and the mor-
photactics in accounting for the semantically-based broken
plurals.

The less stringent requirement was used in this
analysis in order to achieve descriptive adequacy. In
other words, the data itself requires the analyst to
posit direct links between semantics and morphology. It
must necessarily follow that the theory, in order to
maintain explanatory adequacy. must allow this less
stringent requirement (i.e., that direct relationships
may exist between non-contiguous levels). For, in order
for a theory to be explanatorily adequate, it must first
allow descriptively adequate grammars to be written and

secondly, it must present a means of choosing the simplest

 

 

 

 

263

of n descriptively adequate grammars.

The empirical evidence presented by the data from
Modern Iraqi Arabic thus requires the theory of stratifica-
tional grammar to incorporate a less stringent requirement
regarding the existence of direct relationships between
non-contiguous levels. That is. the theory must change

requirement 1 (Hg 1) to requirement la (Rq la):

Rq 1: Direct relationships only exist between ”units“

on immediately contiguous levels.

Bq lat Direct relationships usually exist between
“units" on immediately contiguous levels.
In some cases (as required by the particular
description) direct relationships may exist

between ”units“ on non-contiguous levels.

 

 

FOOTNOTES

Chapter h

1. In these and all following examples V represents the

vowel alternation ifl’u. while v represents any vowels

V1 - the same vowel as in the singular pattern.

2. A problem arose in constructing Figure 15 in that

the fact that one member of each dichotomy is a marked
choice could not be reconciled with the fact that un-
marked choices also have connections to the predication
structures. Therefore the downward gas appear as unor-
dered. If this problem could be resolved, then an analysis

in which the downward 25s were ordered would be preferred.

3. The nisba suffix is a suffix indicating relationship,
for example. someone from. of. or in the ”relationship"
of nation to national with the country of Iraq is termed

Iragi. where the -; is the nisba suffix.

#. It should be recalled that the phonetic tactics has
a restriction on dissimilar consonants finally. and in-

serts an epenthetic copy of the first vowel in such cases.

5. These are traditionally analyzed as based on ”defective
roots according to Ernest N. McCarus, and so are considered

triconsonantal according to such an analysis.

26h

 

 

 

 

265

6. The present analysis uses glottal stop (1) as the under-
lying representation in many instances where other analyses
have used underlying g. z or (g, i). (Cf. Brame (1970) and
Levy (1971)).

7. This analysis misses the generalization that g is

more "marked” than i. In order to capture this. the

analysis would have to show 3 as having more components

:3

than _1_, e.g.. ;_ - P/Hi/ and ‘_ = P/Hi /. Since the
/ Lb/

component 31 is not necessary elsewhere in the phonology
and the component E; is, the analysis given (; 2 £1: E = _Q)

was chosen.

 

 

 

CHAPTER V

Residue

5.1 Theory of Exceptions

Levy's (1971) dissertation on noun plurals in Modern

 

Standard Arabic provides a good example of the use of the
theory of exceptions as posited by Lakoff (1965). Nouns
which prove "exceptions“ are accounted for in one of four
ways:

1) ordering. i.e.. by ordering two rules so that
after the first has applied to an item. the item no longer
meets the structural description of the rule following.
(The later rule is the more general).

2) -rule i] . i.e.. minus rule lexical features. or
marking lexical items so that they do not undergo a partic—
ular rule for which they meet the structural description.
(They may undergo a later more general rule for which they
also meet the SD.)

3) suppletive. i.e.. truly exceptional items which are
spelled out idiosyncratically in the lexicon.

h) minor rules. i.e.. [} qatal Pl] . These mark
exceptional plurals in that they are not the expected
plural for the particular semantic and/or canonical form.
These apply optionally if they are an alternate plural for

the given item.

266

267

5.2 A Stratificational Treatment

Since there are no rules (but only relationships) in
a stratificational description. none of these notions on
the handling of syntactic irregularity is apropos in the
case of the present analysis.

Markedness in the sense of downward ordered gas does
play a role in stratificational grammar. The most frequent
(or the expected) plural for a particular semantic or
canonical subclass is interrelated in an unmarked manner
as the last branch of a downward ordered 21. The less
frequent plurals for such a class are related via the
more marked (leftmost) branches and only occur when they
are enabled to be realized by the small subclass of items
which takes them. Multiple (or optional alternate) plurals
are handled by multiple class membership. That is, an
item taking more than one plural will simultaneously be
a member of several morphOIOgical classes. The plural
which is actually generated will depend on several factors
(i.e., whether the speaker is basing his plural on semantic
grounds or on canonical shape. or whether the speaker
chooses a more marked alternant as opposed to a less marked
one with respect to various broken alternants or with
respect to a broken versus a sound alternant).

There are a few items whose plurals are truly exceptional
in that they serve as the plural for Just the given item.
These plurals will be shown to be "exceptions" by the fact

that the class which enables their realization contains only

 

 

vr:::lIIllIIIIIIIIIIIIIIIIIIIIIIIIIIIiIiIIIIIIIIIIIIiiiiF? f

268

one member. They are an integral part of the system

of relationships nevertheless.

The following nominals were found to be classifiable as

exceptions.
Singular Plural
?ustaad ?asatida
?umm 7ummahaat
badir bduuraat
*xara xaryaan
saa?il saa?il
“saffaar safafiir
sana siniin
*‘adu 'aadaa?
*‘amya (f.) 'imyiin
*qiss qissiisiin
*qasmar qasaamra
qanaat qanawaat
mara niswaan

Glgsg

professor
mother'
seed

feces
liquid
coppersmith

 

year
enemy
blind
clergyman
fool
canal
woman

Starred items have alternant non-exceptional plurals.

An attempt is made to describe the treatment of each.

The plural of 7us§aad 'professor' will be related

to the FaMaaMiLa agent plural but will enable length

following the second vowel to be realized as fl as illus-

trated in Figure 91 yielding 7agatida.

 

269

\

\

 

Figure 91
Exceptional Plural 7asatida

 

The plural of ?umm 'mother' is a sound feminine
plural. The only idiosyncracy is that ah must be enabled
to be realized as a determined non-final suffix in the

environment of plural for this particular item as in

Figure 92 yielding 7gmmahaat.

 

 

 

270

 

Figure 92
Exceptional Plural ?ummah§at

The plural of bgdiz 'seed' (also baziz) is predictable
inasmuch as it belongs to the gun plant plural type (FMuuL).
It is exceptional in that it requires the suffix -aat. This

is shown in Figure 93.

 

 

 

271

Gu/P lau+/

\___'_’l: //
/
/
dti

(>445? ¢ u:

Figure 93
Exceptional Plural bdugraat

The exceptional plural of gar; 'feces' is xagyaan.
The non-exceptional broken plural is xiryaag (weak, low).
(It also has a sound plural xaryaat.) The plural 22125;;
can be viewed as exceptional in the choice of the first

plural vowel as described in Figure 94.

 

 

Figure 9h
Exceptional Plural xazygag

The plural of gaggi;_‘liquid' is the same form as
the singular. It can be handled the same way as the
English plural of bee , i.e.. by letting the singular
item enable the plural morpheme to be realized as t as

illustrated in Figure 95.

 

 

273

 

    

GN/S Joasei/

P—ba: eJ

Figure 95
Exceptional Plural saa?;1

The exceptional plural of saffaar 'coppersmith' can
‘ a_____.
be integrated with the FaMaaMiiL plural type. The singular
enables length following the second vowel to be realized

as I! as in Figure 96 yielding safafiir.
/ .I————

 

 

274

 

Figure 96
Exceptional Plural safafiir
a_______

The plural of gap; 'year' is detailed below. This
plural is unusual in that it takes the plural suffix -iin
and simultaneously fails to have final consonantal
material realized in the plural stem. This second
phenomenon was detailed previously (see pp.l7H-5). The

plural is siniin.

 

 

  

   

P~bas Cc,
GN
4b“ eel / / /

.d/

_i:|’\

Figure 97
Exceptional Plural siniin

The exceptional plural of 'adu 'enemy' is 'aadaa?.
It is similar in exceptionality to gang above in that it
not only takes a suffix but fails to realize a final stem

consonant. Figure 98 describes this plural.

 
 

 

    
 

FLBaseA
C%'bqsed/

  

Figure 98
Exceptional Plural 'aadaa?

The exceptional plural of lgmxg (f.) 'blind' may
be integrated with the FiML-type weak plural and is ex-
ceptional only in that it takes the suffix -iin. Figure
99 details the integration. The plural generated is :igyiig.

 

 

Figure 99
Exceptional Plural ‘imxiin

The exceptional plural of gigs 'clergyman' may be
analyzed as a sound masculine plural with an exceptional

affix -;i§r. FigurelOO illustrates this. The plural
generated is gigsiigi; .

 

 

Figure 100

Exceptional Plural gisgiigiin

The exceptional plural of gaémag 'fool' may be
integrated with the FMaaMLa-type weak. low plural. It
is exceptional only in that it has a as the first vowel

rather than E. Figure loldetails this yielding the plural

g a éaamra a

 

 

6%eak/

¢7ag¢nM1r
\
\
\
\
\ \
a a; Q!
Figure 101

Exceptional Plural gasaamra

The plural of ganaat 'canal', ganagaat. is open to
several alternate analyses. Firstfi it is possible that
it is a sound feminine plural: singular. gana? i t (fem-
inine suffix): plural. gana? + sat (feminine plural suffix)
in which case it would not be exceptional. Second. it
is possible that it is unexceptionally associated with the
broken plural FaMaLaat in which case the final 3 of the
singular fails to get realized. Third, it is possible
to analyze it as an exceptional plural of the form
FaMaMaaL. In case it is analyzed as an exception. it

would be handled as in Figure 102.

 

280

P~bA§eJ
6%“:ks eJ /

 

Figure 102
Exceptional Plural ganagaat

The plural of £212 'woman'. piswaan. is interesting.
Synchronically it appears as a suppletive plural completely
unrelated to the singular form. If one looks at the
underlying radicals (2:1 and Egg) it is not too difficult
to construct a plausible diachronic history given the
strong relationship between the nasals and the affinity
of z and g in historical linguistics in general. However.
in a synchronic study such as this. one is compelled to

treat the forms as unrelated. The plural of mar; would be

 

444-______‘_______________________________:3IIIIIIEEEEEEFT'"

 

281
handled as in Figure 103. Thus in the environment of
plural the morpheme alternant gigflz is realized. This
particular morpheme enables the exceptional plural _aun

to be realized. thus yielding pigfll asn for the

plural. The blank insures that nisg? is realized and

the suffix ash is then appended.

LN/PI /

 

(ginfluhrs)

LEX mm, A MOWCYKS
ALTan-ms
PmrEEN ,

Figure 103
Exceptional Plural niswaan

 

FOOTNOTES
Chapter 5

1. It is this interpretation which is most commonly given
in Arabic grammars. The suffix -t is posited as the
feminine suffix (rather than -;t as is commonly given

for underlying forms using E and 1 (or g and i). for
example, 9355! -;t) on the grounds that -t is used
elsewhere in the system to designate the grammatically
feminine, for example. compare 222 'son' and 212 -3

'daughter'.

282

 

 

 

CHAPTER VI

Conclusion

This dissertation. in substantiating the hypothesis
that the morphology of plural formation in Arabic re-
veals an underlying multi-gender system upon which the
more recent morphological/agreement genders seem to
have been superimposed. makes wider claims for the
Semitic language family in general. Talmy Givén in
suggesting the extension of his hypothesis from the
Bantu language family to the Semitic acted as the
catalyst for this study. The research and reported
results are original. The study has made implications
for the analysis of Arabic both synchronic and diachronic
in substantiating the semanticemorphological corres-
pondence of the broken plural system. It has also
made implications for the theory of stratificational
grammar in positing a less rigidly stratified model to
achieve a more descriptively adequate analysis of the
Iraqi Arabic broken plural system.

Levy (1971). while presenting a thorough phonologically-
based description of Modern Standard Arabic broken
plurals, failed to note the semanto-morphological relation-
ships which exist within the system. The present study
improves on a treatment such as Levy‘s 1) by noting the
semanto—morphological relationships which exist and 2) by

283

 

28h
explaining some of the peculiarities of the system
(e.g.. multiple plural association) by means of culturally-
perceived semantic characteristics.

There seems to be a partial meaning/form correspondence
extractable from the system. For example. tri-consonantal
plural patterns FHaaL are all associated with the

FMaaLa

FMaaLi
semantic component GN/round/. Thus one might say that
the broken plural pattern 25; signifies rotundity.

Also. triconsonantal plural patterns FMuuL are
FMuuLa

both associated with the semantic component GN/genitalia.
body parts/. Similarly one might say that the broken
plural pattern 22; signifies procreativity. No attempt
was made in the present study to discover to what extent

such meaning/form correspondences could be said to exist.

6.1 Unresolved Problems

There are several generalizations which, while
easily stateable in an informal manner. were difficult
to formalize due to the nature and amount of the data.

One is the fact that a number of the broken plurals
share a or a; as a final plural preemptive vowel. It
was decided in this presentation to keep the diagrams
readable rather than to try to capture this generaliza-
tion formally.

Several anatactic relationships were also not dealt
with satisfactorily: the £2 [round] singulars and their

flat plural counterparts. and the fig masculine adjectives

‘W

 

 

285
of color and defect with their a! feminine counterparts.
Anataxis in its usual sense is used to describe
different or opposite ordering of constituents between
one stratum and another. Both of the anatactic phenomena
described above. however. appear on the same stratum,
and it is not clear Just how such Jixtaposition is

handled within the same stratum.

6.2 Suggestions for Further Research

There needs to be much more work along the lines
developed in this study using other dialects of Arabic
and other Semitic languages to verify and perhaps broaden
the hypotheses made herein. When more information is
gathered from other dialects. perhaps a more fully
detailed hypothesis can be made on the semantic-morphological

structure of broken plurals in the Proto-language.

 

 

BIBLIOGRAPHY

 

 

 

BIBLIOGRAPHY

Azzawi. Theresa. 1969. ”A Stratificational Discussion of
Stress in Modern Iraqi Arabic.” MS.

Bateson, Mary Catherine. 1967. Apabic Language Handbook.
Washington, D.C.

Bennett, David C. 1969. Spapial and Temporal Uges of
English Prepositions. Unpublished Ph.D. Dissertation.
ale University.

Blanc, Haim. 196U. Commuppl Dialects of Baghdad.

Boston. Mass.

Brame, Michael K. 1970. Arabic Phonology. Unpublished
Ph.D. Dissertation, M.I.T.

Chafe, Wallace. 1970. Meaning and the Stpucture of
Langppg . Chicago.

Chomsky. Noam. 1965. Agpects of the Theory of Syntax.
Cambridge. Mass.
and Morris Halls. 1968. The Sound Paptern
pf Engligh. New York.
Clarity, B.E.. Karl Stowasser. and Ronald G. Wolfe, eds.

196D. A Dictionary 0: Iragi Ayabics English-Ayabio.

Washington, D.C.

Erwin, Wallace M. 1963. Referenge Gpammap o: Iragi Apabig.

Washington, D.C.

1969. A Basic Course in Iraqi Apabic.

Washington. D.C.

 

Fillmore, Charles. 1968. "The Case for Case" in

Univepsals in Linguistic Theory, Bach and Harms. eds.

New York.

Fleisch, Henri. 1961. T aité de hilolo ie a abe. Vol. I
(Preliminaires. phonLtique, morphologie nominale).
Beirut.

286

 

287

Givén, Talmy. 1971. "Some Historical Changes in the
Noun Class System of Bantu; Their Possible Causes

and Wider Implications" in Papers in African Lingpis-

tics. Urbana.

Gleason, H. A. 1965. Linguistics and English Grammar.

New York.

Goodenough, Ward H. 1956. "Componential Analysis and the
Study of Meaning." Language 32.195-216.

Greenberg, Joseph. 1955. "Internal p—Plurals in Afro-
Asiatic (Hamito-Semitic)." Verbffentlichun en des
Instituts fur Orientforschung, No. 26. Berlin.

1966. "Language Universals" in Cpppgpp
Tpends in Linguistips III. pp. 61-112. The Hague.

Hjelmslev. Louis. 1961. Prolegomena to a Theory of

Langpage. (Translated by Francis J. Whitfield .
Madison. Wis.

 

Hockett, Charles F. 1961. ”Linguistic Elements and Their
Relations.” Langpage 37.29-53.

1968. The State of the App. The Hague.

 

Hyman. Larry. 1970. ”How Concrete is Phonology?"
11.!!Eulge “'6 a 58-76 a

Ikegami, Yoshihiko. 1970. The Semological Structure of
the En lish Verbs of Motion: A Stratificational
Approach. Tokyo.

1971. "A Stratificational Analysis

of the Hand Gestures of Indian Classical Dancing."
Semiotica b.365-91.

 

 

 

Katz, Jerrold J. and Jerry A. Fodor. 1963. ”The Structure
of a Semantic Theory." Langpgge 39.170-210.

Katz. Jerrold J. and Paul Postal. 196k. An Integrated
Theor of Lin uistic Descri tions. Research Monograph
No. 2%. Cambridge, Mass.
Killean, Mary Carolyn G. n.d. ”Interesting Featuresnof
Gender-Number Concord in Modern Literary Arabic. MS.
Kiparsky, Paul. 1968a. ”How Abstract is Phonology?" MS.
1968b. "Linguistic Universals and Linguistic

Change" in Universals in Linguistic Theory, Bach and

Harms, eds. New York.

Kurylowicz, Jerzy. 1961. L'Apophonie en Sémitigue. Wroclaw,

POl‘ndp

288

Lakoff. George. 1965. On the Nat e of S ntactic I e lar t .
Cambridge ' M188 0 e

1969. "Empiricism without Facts."
Foundations of Language 5.518-27.

Lamb, Sydney M. 1966a. ”Prolegomena to a Theory of
Phonology.” Langpage “2.536-73.

1966b. Outline of Stpatificatiopgl Grammag.
Washington. D.C.

1969. "Lexicology and Semantics" in
Linguistics Today, Archibald A. Hill. ed. New York.

1971a. "The Crooked Path of Progress in I
Cognitive Linguistics." Geor eto Mono ra h S es

on Langpages and Lipguisticg, 2‘.
1971b. "Linguistic and Cognitive Networks"
in Cognition: A Multiple View. Paul Garvin. ed.
New York.
1972. "Some Types of Ordering" in Phonological
Theopy: Evolution and prrent Practicg. V. B. Makkai.
ed. New York.

Lekiaschwili, A. S. n.d. ”Zur Bildung des gebrochenen
Plurals im Arabischen.” 25th International Congress of
Orientalists, Proceedings. Vol. II. 136-Ul.

Levy, Mary M. 1971. The Plural of the Noun in Mpdern
Standard Arabic. npublished Ph. . Dissertation,
University of Michigan.

Lockwood, David G. 1969. "Markedness in Stratificational
Phonology." Language h5.300-8.

1972a. Iptroduction to Stratificational

Linguistics. New York.

 

1972b. "Neutralization, Biuniqueness,
and Stratificational Phonology” in Phonological Theory:
Evolution and Current Practice. V. B. Mak 1. ed.
New York.

 

1973. ”Replacives without Process" in

A Reader in Stpatificational Linguistics. A. Makkai

and D.G. Lockwood. eds. University. Ala..

 

Lounsbury, Floyd G. 1956. "A Semantic Analysis of the
Pawnee Kinship Usage." Language 32.158-19h.

 

 

 

 

 

289

Lyons, John. 1968. Introduction to Theoretical Lipgpistics.

Cambridge.

Makkai, Adam. 1972. Idiom Structure in Epglish. The Hague.
McCarthy, R.J.. S.J. and Faraj Raffouli. eds. 1969.
Spoken Apabic o: Baghdad: An Apthology of Texts.
Part Two A . Baghdad.
Murtonen. A. 1964. Broken Plurals: Origin and Development
of the Sygtem. Leiden.

Postal. Paul M. 1968. Aspects of Phonological Theopy.

New York.

Reich, Peter A. 1968. "Symbols, Relations, and Structural
Complexity." Linguistics Automation Project. Yale
University, New Haven. Conn..

Rice, Frank A. and Majed F. Sa'id. 1966. Eastern Arabic.
Beirute

 

Sampson, Geoffrey. 1970a. "On the Necessity for a Phonological
Base Component.” Langppge 46. 586 -626.

1970b. Stratifipgtional Gpammar: A
Definition and an Example. Janus Linguarum Series Minor 88.

 

The Hague e
Saussure. Ferdinand de. 1959. Course in General Lingpistics.
New York.

Sobleman, Harvey. 1962. Arabic Dialect Spudies: Bibliogyaphy.

Washington, D.C.
Sullivan. William J. 1969. A Stratificational Description

of the Phonolo and In leot onal Mor hole of Russ an.
Unpublished Ph.D. Dissertation. Yale University.

Trubetzkoy, N. S. 1969. Principles of Phonolpgy. Berkeley.

Van Ess, John. 1938. The Spoken Arabic of Ipag. London.
(2nd edition).

Weinreich, Uriel, William Labov and Marvin Herzog. 1968.
"Empirical Foundations for a Theory of Language Change“

in Directions for Histopical Linguistics. Austin. Tex.

Woodhead, D. B. and Wayne Beene, eds. 1967. A Dictionapy
of Iraqi Arabic: Arabic-Epglish. Washington, D.C.

 

44‘_‘__________TIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIEVwn

290

Wright, William W. 1896. A Grammgr of the Arabic Language.
3rd edition. 2 Volumes. Cambridge.

 

 

 

 

APPENDIX

Questionsble Plural Classes

 

 

 

APPENDIX

Questionable Plural Classes

These plural classes are questionable either 1)
because their defining semantic characteristics are
too numerous or 2) because the number of items com-

prising a particular plural class are too few.

?aFMuL
Sem Phop
? FvaL

This class does not seem to have a semantic character-
istic and its P-shape for the singular is a quite common

one. All the items of the class are listed here.

saham 'share'
bahar 'sea'
sahsr 'month'
nafis 'soul'
FaMaaLa
em Phon
? FaMiyya
FaMLaan

These items seem to be grouped on the basis of P-shape.
There is no apparent reason for gunuch to be in this class.

All the items of this class appear here.

291

 

 

 

292

xaliyya 'cell'

sariyya ’company'

dahiyya 'blood sacrifice’

qadiyya 'legal case'

maziyya 'merit'

kaslaan 'lazy'

nasraani ’Christian'

‘xisi 'eunuch'
FuMuLaa?
§sm 2mm

7 (weak) FaMi iL

There does not appear to be any semantic character-
istic for the two members of this class. The plural P-
shape is morphologically related to other [weak] plural
classes. however. The singular P-shape FaMiiL is common

to both members of this class.

hakiim 'wise'
haqiir 'low, base'
FuMaaLi
Sem Phon

? (strong) ?

These are the only two members of this class. There
does not seem to be any semantic or phonological Justifica-
tion for the class. The plural P-shape is morphologically

related to other {strong} classes. however.

 

 

293

mulls 'tutor' (in Islam)
fuxtiyya 'turtledove'

 

 

FuMLa
Sem Phon
7 (male) ?

The same remarks apply to this class as to the one
above. The plural P—shape of this class is morphologically

related to other [male plural classes, though.

?ax 'brother'
gariib 'strange'

FuMuuLaat
Sem Phon
? FvaL

These items have no semantic characteristic in common.

They have no other plural.

wasil 'receipt'
dasam 'fat'
FiMLaat
Sem Phon
[meal] Fva

This appears to be a mixed class. There are five
members total and these are the only two related items.

They form a minor class [meal] .

 

'aga 'dinner'
gada 'lunch'

FaMaaL a FaMaat

Sem Phon
[female 7
relative

‘- or<]

This at first appeared to be a broken plural class.
but on closer inspection. it seems to be a regular fem-
inine plural (-aat). Both hint and ?uxut have related

masculine forms:

2% 'son' 2% t 'daughter'
2% 'brother' 2% t 'sister'

where -p is a feminine suffix. The plural of the feminine
is the masculine stem plus the regular suffix -352. In
the singular there is a predictable epenthetic vowel in-
sertion between dissimilar consonants in final position
and a predictable vowel alternation inJu (V). The only
problem arises with the plural of 3; t 'sister' which is
xawpap. The radicals are in an anaZactic relationship
with those of the singular. In order to achieve the cor-
rect plural, the underlying form for the feminine plural
must be ;3 instead of 2; . Glottal stop is normally real-

a a
ized as p in the environment given.

bint 'daughter'
7uxut 'sister'

 

 

 

Sg. minus -1

 

m Phon
[nationality] ?

This is probably a fairly recent plural class and
it cannot be called a broken plural at all. for the
plural is merely the singular minus the (nisba or
"relationship") suffix -;. There is no different
vowel interdigitation. It is strictly a semantically-

based plural class.

?armani 'Armenian'
?a1maani 'German'
?amriikaani 'American'
?ingiliizi 'Englishman'

 

 

sruugi 'person from the rural areas of Southern
Iraq'
baanyaani 'Hindu'
turmaani 'Turkoman'
ruusi 'Russian'
'arabi 'Arab'
FiMaL
Sem Phon
[square] FvMLa

This may be a subclass of FVMaL. The fourth member

of this class is button.

 

sayla “woman's long head scarf'

xayma 'tent'
daéca 'low hedge'

 

Afl____________________________::::1IIIIIIIIIIIIIEIIEEEIIIIIIIIEEEEEEFT

 

296
FMaaLiin
[square] FvMMva(a)

FvavL

All of the items of this class are here listed.
Two of the members may also be analyzed as belonging
to the FMaaMiiL plural class. leaving haayit as an
a____a

exceptional plural.

haayit 'wall'
dukkaan 'shop'
darbuuna 'alley'

 

FaMaaMiiL

Sem Phon
[round] FvMMva(a)
Esq uare] FvMMvL ( a )

Eow, base, wean]
Etrong, high]
Etatement]

This is a plural class which gives evidence of more
than a few characteristic semantic features and which is
therefore probably a P—shape plural class. The possibility
of having morphological overlapping of separate corresponding
semantic classes was considered in cases of two or three
defining semantic characteristics. Given the fact that
this class has more than several and the fact that there

are large numbers of items associated with this plural

 

297
which are not defined by the given semantic characteris-
tics but which do correspond to the given canonical forms.
it seems more probable that this class is defined by
phonological characteristics. Five-sixths of the total
number of items are accounted for here. Round -- 1/3;
square. weak. strong -- 1/7 each: statement. a minor
class -- l/lh. Notice that the body parts here are the
coccyx -- the lowest point of the spine and the pharynx --

the lowest point (exclusive of the vocal cords) in the vocal

 

tract 0

 

 

298

The humans in this class are looked on with dis-

The animals contained herein are the common. or

insignificant insects and birds.

The items in the [strong] S-class denote respected

humans and institutions or a "strength-in-numbers“ concept.

[statementlappears to be characteristic of a significant

minor class.

related.

[round]

haawan
migrafa
girbiil
mashuuf
zanbiil
mirhaad
xaasuuga
cafciir
naaquus
furgaasa
?idfir
mihraae
mihraab
burnayta
burkaan
taawuus
naafuura
minxaar
?ikliil
firjaal
miftaah
mintaad
taahuuna
purgaal

The items listed under it are definitely

'mortar'
'large spoon'
'coarse sieve'
'long boat'
'large basket'
'toilet'
'spoon'
'concave spatula'
'church bell'
'blister'
'fingernail'
'plow'

'prayer niche'
Chat.
'volcano'
'peacock'
'fountain'
'nostril’
'wreath'
'compass'
Ikeyo
'balloon'
'windmill'
'compass'

 

mingaar
?ubuub

karkuusa

qurbaac
'amuud

tarbuus
saaruux
siryaan
mazriib
mirzaab
dambuus
xaazuuq
xartuum
mihraar
mijdaaf
pirdaag
barmiil
bardaag
?ibr11g
tinniin
jarbuu'
xanziir
'irbiid
timsaah
miidaan

[square]

paskiir
tirbaas
jizdaan

carcuuba
marjiiha

tasWiir
tafsiir
qaamuus

 

299

'telescope'
'tube'
'tassel'
'whip'

'post'

'fez'

'rocket'
'artery'
'drainspout'
'drainspout'
op1no

'shaft'
'elephant's trunk'
'thermometer'
'oar.

'glass'
'barrel'
'glass'
'pitcher?
’dragon'
'kangaroo rat'
'pig'

'large snake'
'crocodile'
'square (Open place)'

'hand towel'

'door bolt'

'wallet‘

'picture frame'

'swing'

'photograph'

'book interpreting the Koran‘
'dictionary'

 

taQriir
taqwiim

kuuliis

cirdaag

haanuut
dihliiz

diiwaan
santuur
qaawuus

*dbaara

*faargawn

[weak ]

'as'uus
O

zarduum

Ja‘muus

JarOuum
taa'uun

ba'buus
qaazuug

miskiin

miflis
bartiil
bahluul

majnuun
gaZguuza
mamluuk

tartuur
O O

sa'luuk

darwiis

barguuGa
bargasa

zarzuur
'asfuur
*mxannaO
'ifriit

 

300

'official report'
'calendar'

'opening at the side of a stage'
'summer cabin'

'canteen'

'narrow passage'

'guest house of a village'
'dulcimer'

'hospital ward'

'file'

'railroad car'

'coccyx'
'pharynx'
'feces'

ogemo

'plague'

‘act of lewd poking'
'the shaft'
'poor'
'bankrupt'
'bribe‘

'clown‘

'madman'
'well-proportioned girl'
'white slave'
'braggart'
'pauper'
'dervish'

'flea'

'small insects'
'starling'
'sparrow'
'effeminate man'
'clever'

*Items preceded by an asterisk do not correspond to the
established P—shape(s) for the given plural class.

[strong]

bilyawn
zangiin
tilmiid
Jiiraan
ma'miil
m'azzib
'irriis
mashuur
'anguud
Jamhuur
?unbuub
?ustuul
tadbiir
miiOaaq
miiraaG
maw'id

qaanuun
dastuur

{statement}

 

?inbiiq
?injiil
tasriih
matluub
ta'biir
ta'riif
FV1MaL
Sem
[abstract]
[round]

[sq uare]

[strong]

Phon

 

'billion'
'wealthy'
'student'
'neighbor'
'cuetomer'
'guest'
'bridegroom'
'celebrity'
‘bunch'
'crowd'
'tribe'
'fleet'
'organization'
'pact'
'inheritance'
'appointment'
'law'
'constitution'

'retort'
‘gospel'
'declaration'
'wish'
'expression‘
'definition'

FVMLa

 

 

302

This is another plural class which is characterized
by more than one semantic feature. Three-fifths of the
total number of items are accounted for semantically by
means of these S-classes. Abstract accounts for 1/6;
round. 14; square. 1/6: strong. 1/25.
One minor class. [elongated]. is a subclass of [round] .
Due to the number of defining semantic characteris-
tics, this is probably better described as a P-shape
plural class. (See note under Egflggflii_. p. 296.)

[abstract]

Jumla 'total'
tuhfa 'rarity'
bid'a 'innovation'
bu?ra 'focus'
junha 'misdemeanor'
hiJBa 'excuse'
hikma 'wisdom'
hiila 'trick'
xibra 'experience'
xidma 'a service'
xitta 'plan'

dimma 'conscience'
zubda 'essence'
zarga 'blue'

sil'a 'commodity'
siira 'behavior
sudfa 'chance'
:ibra 'warning'
'uqda 'quirk'
fursa 'opportunity'
nag. ~1/2-

giima 'price'

 

mihna
mudda
ni'ma
mihna
nisba
hudna

[round]

birka
nugta
nuqta
wusla
luula
kutla
giisa
gulla
guffa
gursa
girba
qus'a
qubba
fiina
r11§a
'uuda
'ilba
'ukra
'ugda
'uqca
tawba
tug'a
gum
siisa
sibha
xuuda
dugma

 

 

303

'tribulation'
'period of time'
'benefaction'
'profession'
'relationships'
'armistice'

'puddle'
'SPOt.} ”free" variants
’dot'

'piece'

'cylinder'

'lump'

'date container'
'cannonball'
'large basket'
'flat round loaf'
'water bag'

'large bowl for carrying food to the troops'
'dome'

'fez'

'token'

'amulet'

'wooden container'
'bump'

'knct'

'heel (of shoe)‘
'ball'

'stain'

'navel'

'glass bottle'
'prayer beads'
'helmet'

'button'

 

 

 

30h

 

xirza 'bead'

hilma 'nipple'

hufra 'hole'

hidba 'hump on the back’
31113 'basket of palm leaves'
jufra 'hole'

tunga ‘clay water Jug'

badra ‘spool' ‘
liifa 'pad of plant fibers used as scouring pad'
Qumma 'peak'
*kawm 'heap'

[elongated]

firca 'brush'

'unda 'splinter'

'usba 'tendon'

diixa 'date stalk'

dings 'column'

Gurma 'shred'

pulka 'spangle'

tides 'drawstring'

tufga 'rifle'

tikma 'pillar'

?ubra 'needle'

liita 'thin, flexible stick'
*hiim 'crowbar'

[square]

hizza 'large cloth slung over the shoulder by

workmen to carry sand'

xirga 'rag'

hujra 'room'

Suwwa 'patch'

juuma 'loom'

Junta 'suitcase'

Jinda

'porter's back pad'

 

305

Jufta 'dominoes piece'
tiiga 'wall around roof'
buqca ‘large cloth'
?awda 'room'
durga 'shield'
ruzma 'parcel'
rug'a 'patch'
sufra 'woven mat'
su'ba 'section'
¥1qqa 'apartment'
gutra 'headcloth'
gurfa 'room'
gutra 'headcloth'
fijja 'piece of cloth'
fuuta 'woman's head scarf'
girma 'pleat'
qisla 'barracks'
gubba 'room'
kulla 'mosquito net'
[strong]
hujja ‘authoritative source'
Tumma 'people'
?usra 'clan'
xulfa 'offspring'
zumra 'gang'
firqa 'team'
FMaaMiiL
§sm 2:122
[agent] FvMMva(a)
[instrument]
[square]

[round]

306

The items listed here account for about 3/5 of the
total number of items in this class -- agent. 1/5: instru-
ment, l/5; round. l/lh and square, 1/8. The semantic
classes which characterize this plural class are all major
categories whose members require no further explanation.
There are other [agent] classes. three of which are char-
acterized by different P-shapes for the singular. viz.
FaaMi (p1: FuMaat). FaaMiL (p1: FuMMaaL). FaMiiL (p1:
FuMalaa?) and one of which has the same P-shape as this
class FvMMva(v) (p1: FaMaaMiLa). There is also another
plural class for [instrument], FamaaMiL. whose character-
istic singular P-shape is FvMMvL(a).

Because of the number of semantic features needed to
define this class it is probably better analyzed as a P-
shape class. (See note under FaMaaMiiL, p.296.)

[agentfi

baggaal 'grocer'

bayyaad ’tinner'

hammaal 'porter'

xabbaaz 'baker'

raaguus 'dancer'
muxtaar ‘mukhtar'
raggaa' 'shoe repairman'
zabbaal 'garbage collector'
sabbaag 'painter'
saffaar 'coppersmith'
fallaah 'farmer'
gassaab 'butcher'
gawwaad 'pimp'

kannaas 'sweeper'

 

 

307

 

mallaah 'sailor'

naddaaf 'man who renovates mattresses'
nazzaah 'man who empties septic tanks'
naatuur 'watchman'

laa'uub 'player'

saa'uul 'fire tender'
saabuuh 'expert swimmer'
caawuus 'foreman'

Jaasuus 'spy'

[instrument]

saatuur 'cleaver'

mingaas 'tweezer'

maakuuk 'bobbin'

mi'caal 'sling'

maaxuud 'club, bat'
faddaan 'animal-drawn plow'
faanuus 'lantern'

Eaahuul 'plumb line'
sicciin 'knife'

zanjiil 'chain'

cingaal 'fastener'

cillaab 'hook'

Xaakuuc ’hammer‘

tannuur 'outdoor oven'
bismaar 'nail'

baapuur 'long cigarette holder'
*mhaffa 'hand fan'
*mugas 'scissors'
*jaawan 'large mortar'
[round]

lakluuka 'ball. lump'
karuuk 'baby cradle'

'stone'

salbuux
O

 

saahuud

xarmuus
mincaasa
maa'uun
finjaan

[square]
taabuut
daaguur
diilaab
raazuuna
sijjaada
sirdaab
sibbaac
sanduuq
sanduug
taabuuga
qur?aan
kaagada

 

308

 

'large bead on the end of a string of
prayer beads'

’bunch of dates'
'bowl'

'plate'

'(small porcelaid cup'

'coffin'

'removable center door post'
'wardrobe'

'shelf'

'prayer rug'

'cellar‘

'window'

'money box:}"free” variants
'box

'brick'

'Koran'

'sheet'