diachrony 4 th International Conference on Maltese Linguistics Université Lyon 2 17062013 Christopher Lucas SOAS University of London Michael Spagnol Albert Gatt ID: 269649
Download Presentation The PPT/PDF document "Connecting /t/ in Maltese numerals: sync..." is the property of its rightful owner. Permission is granted to download and print the materials on this web site for personal, non-commercial use only, and to display it on your personal computer provided you do not modify the materials and that you retain all copyright notices contained in the materials. By downloading content from our website, you accept the terms of this agreement.
Slide1
Connecting /t/ in Maltese numerals: synchrony and diachrony
4th International Conference on Maltese LinguisticsUniversité Lyon 217/06/2013
Christopher Lucas
SOAS, University of London
Michael
Spagnol
Albert
Gatt
University of MaltaSlide2
2
Maltese cardinal numerals 2–10
Long
form
‘two’
tnejn‘three’tlieta‘four’erbgħa‘five’ħamsa‘six’sitta‘seven’sebgħa‘eight’tmienja‘nine’disgħa‘ten’għaxra
Short formżewġtlieterba’ħamessittseba’tmiendisa’għaxar
-t
form
żewġt
tlitt
erbat
ħamest
sitt
sebat
tmint
disat
għaxartSlide3
3
Outline
The synchronic distribution of the
-t
form numerals:
Existing claims and their inadequacy.Production experiment to produce more accurate generalization.
Reconstructing the evolution of this distribution.Slide4
4
Previous literature
Two main types of claim:
-t
is triggered by an initial
consonant cluster in the following plural noun.Not quite as contradictory as they may seem!
Both views tell part of the story.-t is triggered by an initial vowel in the following plural noun.Slide5
5
Previous literature
Consonant cluster views:
Cremona (1938: 204–5): plurals that are
CC-initial +
monosylabic trigger -t (+ epenthetic /i/), e.g.: ħabib ‘friend’, pl. ħbieb, 2 friends: zewġt iħbieb belt ‘town’, pl. bliet, 8 towns: tmint ibliet fenek ‘rabbit’, pl. fniek, 10 rabbits: għaxart ifniek But snin ‘years’ is exceptional: għaxar sninSlide6
6
Previous literature
Consonant cluster views:
Borg (1974):
-t
insertion, where licit, is always optional as one means of avoiding illicit consonant clusters, e.g.: kelb ‘dog’, pl. klieb, 7 dogs: sebat iklieb / seba’ klieb Also: -t is ungrammatical before a (non- epenthetic) initial vowel, and with “sound” plurals, e.g.: stampa ‘picture’, pl. stampi, seba’ stampiSlide7
7
Previous literature
Initial vowel views:
Sutcliffe (1936: 188–9):
-t
co-occurs with a following vowel, whether epenthetic or not.Aquilina (1965: 118): all V-initials, all monosyllabics, optional with CC-initial disyllabics.Hoberman (2007: 277–8): before any Arabic-derived V-initial, e.g. erbat aħwa
‘4 siblings’, all monosyllabics, optional with CC-initial disyllabics. Slide8
8
Experiment
Our intuition:
onset
is the key factor, with
number of syllables also contributing.Different phonological properties of various numerals (mono-/disyllabic, C-final vs. V-final) perhaps also relevant.
Not feasible to properly test further potentially relevant variables (sound vs. broken plural, Arabic- vs. Romance-derived) in a single experiment.Slide9
9
Experiment design
Subjects:
35
students of the University of Malta (
18 female, 17 male) + 5 older speakers (40+, 4 female, 1 male).Test items: 1 of 7 numerals (2, 4, 5, 7, 8, 9, 10) paired with a singular noun.
Fillers: a numeral (11–19) paired with a singular noun.
Task: say the phrase as you would in normal speech.
N.B.: 2–10 take a plural noun, 11+ a singular.
3 + 6 excluded. Slide10
10
Experiment design
Long
form
‘two’
tnejn‘three’tlieta‘four’erbgħa‘five’ħamsa‘six’sitta‘seven’sebgħa‘eight’tmienja‘nine’disgħa‘ten’għaxra
Short formżewġtlieterba’ħamessittseba’tmiendisa’għaxar
-t
form
żewġt
tlitt
erbat
ħamest
sitt
sebat
tmint
disat
għaxartSlide11
11
12 qasbaSlide12
12
12 qasba2 tifla13 bandiera7 artiklu
15 għalqa
10 larinġa
11 bejt
5 ħsieb19 ballun7 ħu/oħt19 kutra2 wildSlide13
13
Experiment design
56 plurals were tested, divided equally into 8 categories:
CC-initial: Mono-, Di-, Polysyllabic
CV-initial: Mono-, Di-, Polysyllabic V-initial: Di-, Polysyllabic Subjects were split into 7 groups. Each group had a different combination of numeral and noun, e.g.: Group 1: 2
tifla, 7 artiklu, 8 but Group 2: 4 tifla, 8 artiklu, 9 but Group 3: 5 tifla, 9 artiklu, 10 but etc. Slide14
Test nouns
Monosyllabic
Disyllabic
Polysyllabic
CC-initial
fniek
‘rabbits’bramel‘buckets’kmandamenti‘commandments’klieb‘dogs’platti‘plates’trakkijiet‘trucks’
bniet
‘girls’
stampi
‘pictures’
flokkijiet
‘shirts’
djar
‘houses’
ġranet
‘days’
dmirijiet
‘duties’
ħbieb
‘friends’
skejjel
‘schools’
ħsibijiet
‘thoughts’
snin
‘years’
ljieli
‘nights’
żminijiet
‘times’
bwiet
‘pockets’
kmamar
‘rooms’
studenti
‘students’
CV-initial
files
‘files’
kotba
‘books’
neputijiet
‘grandchildren’
films
‘films’
naħat
‘sides’
larinġiet
‘oranges’
xhur
‘months’
bozoz
‘bulbs’
pappagalli
‘butterflies’
jiem
‘days’
widnejn
‘ears’
kategoriji
‘categories’
gowls
‘goals’
żgħażagħ
‘youths’
pajjiżi
‘countries’
fonts
‘fonts’
diski
‘disks’
postijiet
‘places’
toasts
‘pieces
of toast’
kelmiet
‘words’
verżjonijiet
‘versions’
V-initial
ilsna
‘tongues’
arloġġi
‘clocks’
erwieħ
‘souls’
ajruplani
‘aeroplanes’
ulied
‘sons’
operazzjonijiet
‘operations’
uċuħ
‘faces’
għasafar
‘birds’
idejn
‘hands’
appartamenti
‘
appartments
’
aħwa
‘siblings’
artikli
‘articles’
oqsma
‘fields’
individwi
‘individuals’Slide15
Di
Mono
Poly
CC
123
1774CV
1301V93
n/a
31
Responses with
-t
by onset and no. of syllablesSlide16
Results for young (and older) speakers
Monosyllabic
%
-t
(young)
% -t(40+)Disyllabic% -t(young)% -t(40+)Polysyllabic% -t
(young)% -t(40+)CC-initial
fniek
94%
100%
bramel
74%
60%
kmandamenti
6%
0%
klieb
97%
100%
platti
7%
0%
trakkijiet
0%
0%
bniet
88%
100%
stampi
6%
40%
flokkijiet
3%
0%
djar
94%
100%
ġranet
77%
80%
dmirijiet
0%
0%
ħbieb
94%
100%
skejjel
56%
40%
ħsibijiet
3%
0%
snin
10%
25%
ljieli
80%
80%
żminijiet
0%
0%
bwiet
76%
100%
kmamar
71%
100%
studenti
0%
0%
CV-initial
files
0%
0%
kotba
0%
0%
neputijiet
0%
0%
films
0%
0%
naħat
0%
0%
larinġiet
0%
0%
xhur
3%
0%
bozoz
0%
0%
pappagalli
3%
0%
jiem
97%
100%
widnejn
0%
0%
kategoriji
0%
0%
gowls
0%
0%
żgħażagħ
3%
0%
pajjiżi
0%
0%
fonts
0%
0%
diski
0%
0%
postijiet
0%
0%
toasts
0%
0%
kelmiet
0%
0%
verżjonijiet
0%
0%
V-initial
ilsna
33%
40%
arloġġi
9%
0%
erwieħ
84%
100%
ajruplani
20%
20%
ulied
45%
40%
operazzjonijiet
3%
0%
uċuħ
24%
20%
għasafar
34%
60%
idejn
26%
80%
appartamenti
6%
20%
aħwa
45%
20%
artikli
9%
20%
oqsma
27%
20%
individwi
9%
20%Slide17
Di
Mono
Poly
CC
123
1744CV
111V66
0
31
Responses with
-t
by onset and no. of syllables, excluding
erwie
ħ
,
jiem
and
sninSlide18
18
Summary of results
Data were analysed using a linear mixed effects model, with random intercept and slope for test subjects, but not test items.
But no main effect of
numeral choice.
Main effect of onset type (z=-9.79, p<0.001) and no. of syllables (z=-9.96, p<0.001).
Interaction between
onset and syllables
(z=2.54, p=0.01). No pairwise interactions involving numeral choice, but an interaction between
all three conditions
(z=2.5, p=0.01).Slide19
19
Summary of results
Plurals with
CV- onsets
(other than
jiem ‘days’) are extremely hostile to -t, regardless of no. of syllables, broken vs. sound etc. CV-initialfiles0%kotba0%neputijiet0%films
0%naħat0%larinġiet0%xhur3%bozoz0%pappagalli3%(
jiem
97
%)
widnejn
0%
kategoriji
0%
gowls
0%
żgħażagħ
3
%
pajjiżi
0%
fonts
0%
diski
0%
postijiet
0%
toasts
0%
kelmiet
0%
verżjonijiet
0%Slide20
20
Summary of results
With
CC- onsets
there is a strong interaction with no. of syllables: monosyllables (all broken plurals) strongly favour
-t; disyllabic plurals are reasonably favourable when broken, apparently very hostile when sound; polysyllabic plurals (all sound) are very hostile. CC-initialfniek94%bramel74%kmandamenti6%klieb97%platti
7%trakkijiet0%bniet88%stampi6%flokkijiet3%djar94%ġranet
77%
dmirijiet
0%
ħbieb
94%
skejjel
56%
ħsibijiet
3
%
(
snin
10
%)
ljieli
80%
żminijiet
0%
bwiet
76%
kmamar
71%
studenti
0%Slide21
21
Summary of results
V- onsets
are somewhat favourable when disyllabic, less so when polysyllabic. Unclear whether sound vs. broken plural or no. of syllables (or both) is relevant here.
V-initial
ilsna33%arloġġi9%(erwieħ84%)ajruplani20%
ulied45%operazzjonijiet3%uċuħ24%għasafar34%idejn26%appartamenti6%
aħwa
45%
artikli
9%
oqsma
27%
individwi
9%Slide22
22
Summary of results
The interaction of onset, no. of syllables and
numeral choice
arises only with
V-initial disyllabics, where frequency and phonology of word-final segment of the numeral seem to be relevant:NumeralFreq. per million in MLRS corpusFreq. -t with disyllabics
Final segment
żewġ
893.42
50%
Obstruent
erba
‘
293.27
30%
Sonorant
ħames
276.2
43%
Obstruent
seba
‘
113.49
31%
Sonorant
tmien
93.04
33%
Sonorant
disa
‘
73.72
25%
Sonorant
għaxar
130.63
21%
SonorantSlide23
Questions arising
What kinds of rules do children acquire to produce these generalizations?
How did this system evolve from its early spoken Arabic ancestor?
23
Is the apparent inter-speaker variation (partially) a reflection of optionality for individual speakers?Slide24
Questions arising
Yes. Of the 7 subjects we retested (1 from each group),
none
were consistent in their use of
-t
.
24Is the apparent inter-speaker variation (partially) a reflection of optionality for individual speakers?
Across all test items, consistency of
-t
insertion ranged from
50%–84%
(for the least/most consistent subjects).Slide25
Optionality of
-t
for individual speakers
25
Mono-
Per subject consist-ency
Overall freq.Di-Per subject consist-encyOverall freq.Poly-Per subject consist-encyOverall freq.
CC-
fniek
86%
94%
bramel
86%
74%
kmandamenti
86%
6%
klieb
100%
97%
platti
100%
7%
trakkijiet
100%
0%
bniet
100%
88%
stampi
71%
6%
flokkijiet
86%
3%
djar
86%
94%
ġranet
33%
77%
dmirijiet
100%
0%
ħbieb
100%
94%
skejjel
57%
56%
ħsibijiet
100%
3%
snin
100%
10%
ljieli
86%
80%
żminijiet
100%
0%
bwiet
86%
76%
kmamar
83%
71%
studenti
100%
0%
CV-
files
100%
0%
kotba
100%
0%
neputijiet
100%
0%
films
100%
0%
naħat
100%
0%
larinġiet
100%
0%
xhur
83%
3%
bozoz
100%
0%
pappagalli
100%
3%
jiem
100%
97%
widnejn
100%
0%
kategoriji
100%
0%
gowls
100%
0%
żgħażagħ
100%
3%
pajjiżi
100%
0%
fonts
n/a
0%
diski
100%
0%
postijiet
100%
0%
toasts
100%
0%
kelmiet
100%
0%
verżjonijiet
100%
0%
V-
ilsna
86%
33%
arloġġi
83%
9%
erwieħ
57%
84%
ajruplani
57%
20%
ulied
60%
45%
operazzjonijiet
100%
3%
uċuħ
86%
24%
għasafar
71%
34%
idejn
71%
26%
appartamenti
86%
6%
aħwa
100%
45%
artikli
86%
9%
oqsma
57%
27%
individwi
71%
9%Slide26
Acquisition
Acquirers must posit something like the following rules:
numerals followed by
CV-
onsets obligatorily lack
-t.
CC- broken plurals at least optionally trigger -t.numerals followed by sound plurals with CC- (and CV-) onsets obligatorily lack -t.
26
V
- plurals
(of all types?)
optionally trigger
-t
.Slide27
Frequency
Acquirers are clearly sensitive to the token frequency of nouns and numeral + noun strings.
27
Apparent effect of no. of syllables could perhaps be explained as an epiphenomenon of frequency and/or as parasitic on the fact that
polysyllabicity
is associated with sound plurals.
This will be investigated further with a planned corpus study and a second experiment focussed on distinguishing the role of no. of syllables vs. sound/broken plural.Slide28
Exceptions
Why treat results for
erwieħ
‘souls’,
jiem ‘days’ and snin ‘years’ as discountable exceptions?
28
Erwieħ
: etymologically (perhaps synchronically)
erwieħ
is CC- monosyllabic.
Vowel epenthesis occurs whenever a word-initial
sonorant
+ consonant cluster follows something other than a vowel.
Usually this is /
i
/, e.g.
bla
lsien
‘without a tongue’ vs.
i
lsien
, but before /r/ it is /e/, e.g.
rġajt
‘I redid’ >
erġajt
But epenthetic /e/ is usually retained in all contexts (i.e. is no longer epenthetic).Slide29
Evolution
29
Classical Arabic (like
Akkadian
and Canaanite) exhibits gender polarity in numeral agreement from 3–10, e.g.
yawmun
(m.) ‘day’ > xamsatu (f.) ʾayyāmin ‘5 days’.
In many contemporary Arabic dialects this construction has been radically
refunctionalized
:
Independent numerals are etymologically feminine, e.g.
xamsa
‘
five
’.
Determiner numerals are (usually) etymologically masculine, e.g.
xam
a
s
sinīn
‘5 years’, but a reflex of
tāʾ
marbūṭa
reappears with high-frequency V-initial plurals, e.g.
xamas
-t-
iyyām
‘5 days’.Slide30
Evolution
30
Assuming Maltese started from this kind of system, the following evolution seems likely:
Stage I: only
V
CCV(V)C plurals take
-t, e.g. aħbāb ‘friends’.Stage IIa: VCCVVC and CVCVVC collapse into CCVVC, e.g. aħbāb ‘friends’ > ħbieb,
kilāb
‘dogs’ >
klieb
.
VCCiCa
>
VCCCa
, e.g.
ilsna
‘tongues’ (<
alsina
)
Stage
IIb
: among CCVVC plurals, only originally V-initial ones take
-t
. Initial V preserved after
-t
.Slide31
Evolution
31
Assuming Maltese started from this kind of system, the following evolution seems likely:
Stage
IIc
: initial V after
-t reanalysed as epenthetic, except for ilsna etc. Stage IIb: among CCVVC plurals, only originally V-initial ones take -t
. Initial V preserved after
-t
.
Stage III:
-t
+ epenthetic V generalized to all CCVVC plurals.
Stage
IVa
: all CC-initial
broken
plurals take
-t
,
obligatoriness
increases with frequency.
Stage
IVb
:
all
V-initial plurals take
-t
,
obligatoriness
increases with frequency.
Stage
IVa
: all CC-initial broken plurals take
-t
,
obligatoriness
increases with frequency.
Stage
IVb
:
all
V-initial plurals take
-t
,
obligatoriness
increases with frequency.
Present-day: CC- broken plurals identified as prototypical context for
-t,
V-initial contexts perhaps on the way out?Slide32
Jiem
and
snin
32
High token frequency (of lexical items, constructions) is commonly associated with irregularity (
Bybee 2003).This is because high-frequency
items are resistant to analogical (but not phonological) change.
Jiem
‘days’ <
ayyām
, i.e. a
-t
-inserting pattern. It is CV-initial but retains its V-type
morphosyntax
due to v. high frequency.
Snin
‘years’ <
sinīn
, i.e. a
-t
-resistant pattern. It is CC-initial but retains its CV-type
morphosyntax
due to v. high frequency. Also not a broken plural!Slide33
Conclusion
33
Despite having no clear function, and often being optional, the Maltese numeral
-t
morpheme has survived for 1000+ years and expanded its range to include both V- and CC-initial following plural nouns.
This serves as a reminder that functional pressures are not the only drivers of linguistic change.
At least as important is the need for acquirers to establish generalizations which are consistent with their input, but also transcend it, allowing them to produce novel utterances.
This is true of all aspects of the input, even those as functionally redundant as
Maltese numeral
-t
.Slide34
References
Aquilina, Joseph. 1965. Teach Yourself Maltese. London: The English Universities Press.Borg, Alexander. 1974. Maltese Numerals. Zeitschrift der Deutschen Morgenländischen Gesellschaft 124, 291–305.Bybee, Joan L. 2003. Mechanisms of change in grammaticalization: The role of frequency. In Richard D. Janda & Brian D. Joseph (eds.), Handbook of Historical Linguistics, 602–23. Oxford: Blackwell.Cassar, Carmel. 2000. A Concise History of Malta. Msida: Mireva.Cremona, Antonino. 1938. Tagħlim fuq il-Kitba Maltija. It-Tieni Ktieb. Oxford: Oxford University Press.
Fabri
, Ray. 1994. The syntax of numerals in Maltese. In Joseph M.
Brincat
(ed.), Languages of the Mediterranean. Proceedings of the Conference held in Malta, 26–29 September 1991, 228–39. Msida: University of Malta.Hoberman, Robert D. 2007. Maltese morphology. In Alan S. Kaye (ed.), Morphologies of Asia and Africa, 257–81. Winona Lake, IN: Eisenbrauns.Sutcliffe, Edmund. 1936. A Grammar of the Maltese Language, with Chrestomathy and Vocabulary. Oxford: Oxford University Press.Taine-Cheikh, Catherine. 2012. Numerals. In Lutz Edzard & Rudolf de Jong (eds.) Encyclopeida of Arabic Language and Linguistics Online. Brill Online. http://referenceworks.brillonline.com/entries/encyclopedia-of-arabic-language-and-linguistics/numerals-COM_vol3_023834