Malayalam - A Grammatical Sketch and A Text (Jiang) PDF
Malayalam - A Grammatical Sketch and A Text (Jiang) PDF
Malayalam - A Grammatical Sketch and A Text (Jiang) PDF
Haowen Jiang
Department of Linguistics
Rice University
April, 2010
1
Contents
2
4.2.4. Postpositional phrases..............................................................................37
4.3. Predicate nominals ...............................................................................................37
4.4. Existential/locational/possessive sentences .........................................................38
4.5. Expression of grammatical relations....................................................................39
4.6. Causatives ............................................................................................................40
4.7. Passives ................................................................................................................41
4.8. Reflexives ............................................................................................................42
4.9. Reciprocals...........................................................................................................42
4.10. Questions..........................................................................................................43
4.10.1. Polar questions .........................................................................................43
4.10.2. Content questions.....................................................................................44
4.11. Imperatives.......................................................................................................44
4.12. Negation...........................................................................................................45
Appendix I: Glossary ...............................................................................................................47
Appendix II: A Practical Orthogrphy of Malayalam ...............................................................54
Part II: A Text of Malayalam ...............................................................................................55
1. Introduction......................................................................................................................55
2. A bird’s-eye view of the text ...........................................................................................56
2.1. The text from top down........................................................................................56
2.2. Justifications for the transcriptions ......................................................................58
2.3. Preliminary analysis.............................................................................................60
3. A worm’s-eye view of the text.........................................................................................61
3.1. The text from bottom up ......................................................................................62
3.2. Justifications for the glossing...............................................................................70
3.3. Linguistic analysis ...............................................................................................72
References................................................................................................................................75
3
Tables
4
Abbreviations
?? unknown
1 first person
2 second person
3 third person
ACC accusative
AFF affirmative
ATTR attributive
CAUS causative
COND1 first conditional (-aal)
COND2 second conditional (-engkil)
COORD coordination
COP1 first copula (e.g. aaNE, aayiryunnu, aayiryikkyum)
COP2 second copula (e.g. aakunnu, aayi, aakum)
CVB converb
DAT dative
DNT deontic
EMPH emphatic
EX existential
EXC: exclusive
F feminine
FS false start
FUT future
GEN genitive
HON honorific
HAB habitual
HORT hortative
IMP imperative
INC inclusive
INF1 first infinitive (-uka)
INF2 second infinitive (-aan)
INS instrumental
INTR intransitive
LOC locative
5
M masculine
NEG negative
NMLZ nominalizer
NOM: nominative
PFV perfective
PL plural
POT potential
PRF perfect
PROH: prohibitive
PRS present
PST past
Q interrogative
QUOT quotative
RECP reciprocal
REFL reflexive
SG singular
SOC sociative
TR transitive
UQ universal quantification
6
Part I: A Grammatical Sketch of Malayalam
1. Introduction
Malayalam (ISO: mal) is a Dravidian language primarily spoken in the southwest of India.
According to Lewis (2009), it is the official language of Kerala state and Lakshadweep union
territory. Its alternative names include Alealum, Malayalani, Malayali, Malean, etc. Within
India alone there were over 35 million speakers of Malayalam in 1997, not including the other
nearly 500,000 speakers outside India.
The data used in this grammatical sketch mostly come from my consultant Sona Joseph, a
26-year-old female who speaks Malayalam natively and English fluently. She originally came
from Kochi (formerly Cochin), Kerala, India, and is currently a graduate student at Rice
University.
The methodology adopted here is primarily direct elicitation. To complement the elicitation
data from the sole consultant, I also transcribed a small portion of a TV interview on the Onam
Festival, which is the biggest festival in the state of Kerala (see Part II for details).
This grammatical sketch is divided into three major sections, namely, phonology,
morphology, and syntax. Aside from the sketch, a glossary of some selected Malayalam
words (about 420) is included in Appendix I.
2. Phonology
The starting point of data collection for Malayalam phonology is Comrie and Smith’s (1977)
Basic Vocabulary. Most of the time, the words given by the consultant were retrieved from her
memory. In some cases, however, the consultant used Gundert’s (1970) Malayalam and
English Dictionary in order to find out words with some particular syllables.
The organization of this section is as follows: Section 2.1 introduces the phoneme inventory
in Malayalam; Section 2.2 identifies some allophonic variations of the proposed phonemes;
Section 2.3 slightly touches upon the syllable structure of Malayalam; finally Section 2.4
discusses the typological status of the phoneme inventory as a whole.
7
alveolo-palatals, retroflexes, and velars. A bilabial plosive, for example, is either voiceless or
voiced. Within voiceless bilabial plosives a further distinction is made between aspirated and
unaspirated ones whereas for voiced bilabial plosives the distinction is between modal-voiced
and breathy-voiced ones. Additionally, an unaspirated voiceless bilabial plosive is either
singleton (i.e. short) or geminate (i.e. long). The same five-way distinction is also found in
dental, alveolo-palatal, retroflex, and velar plosives.
Alveolo-palatal
Labiodental
Retroflex
Alveolar
Bilabial
Glottal
Palatal
Dental
Voice
Velar
- p pː pʰ t̪ t̪ ː t̪ ʰ tʰʲ tɕ tɕː tɕʰ ʈ ʈː ʈʰ k kː kʰ
Plosive
+ b bʱ d̪ d̪ ʱ ʤ ʤʱ ɖ ɖʱ g gʱ
Nasal + m mː n̪ n̪ ː n nː ɲ ɲː ɳ ɳː ŋ
Trill + r
Tap/Flap + ɾʲ
Fricative - s ɕ ʂ h
- f
Affricate
+
Cen. Approx. + ʋ ɻ j
Lat. approx. + l lː ɭ ɭː
In terms of place of articulation, on the other hand, alveolars are the most complex because
they involve all manners of articulation except for affricate. Even an alveolar tap/flap is
distinguished from an alveolar trill, much like the case in Spanish (e.g. pero /peɾo/ ‘but’ vs.
perro /pero/ ‘dog’). Unlike in Spanish, however, the alveolar tap/flap in Malayalam has an
inherent palatal quality in it, which is indicated by the palatalized diacritic.
A conspicuous feature in the consonant inventory is that length is phonemic. Consonants that
show this short/long contrast include all voiceless plosives (except for alveolars), all nasals
(except for velars), and all lateral approximants. What these consonants have in common is an
occlusion phase before the release of a segment, during which the air pressure is built up, and
8
this is also precisely the phase that is prolonged in the case of geminate consonants.
9
Admittedly, more (near) minimal pairs are needed in order to justify the proposed phonemes.
Some generalizations, however, can still be gleaned from the (near) minimal pairs found so far.
For instance, Pair (2) and (5) show the phonemic contrast in length for voiceless bilabial and
retroflex plosives respectively. Since there seems to be no phonetic reason whatsoever that
would make bilabials and retroflexes the only voiceless plosives that demonstrate the
short/long contrast, I expect voiceless dental and velar plosives to behave just like their bilabial
and retroflex counterparts, even though I have not found any (near) minimal pairs to show this
yet.1 In the same vein, I also believe that alveolars would not be the only nasals that show the
short/long contrast, as in Pair (7). More importantly, crosslinguistically plosives and nasals
tend to share the same pattern because they form the natural class of stops. Thus, in a case like
Malayalam, where we find minimal pairs of the short/long contrast in voiceless bilabial and
retroflex plosives as well as alveolar nasals, we may well infer that length is also phonemic in
voiceless alveolar plosives as well as bilabial and retroflex nasals, which is confirmed by
acoustic evidence (See Section 2.1.3.1 for details).
Moreover, Pair (9) and (10) show that geminate nasals make a distinction among alveolars,
retroflexes, and palatals. Crosslinguistically geminate segments are more marked than their
singleton counterparts, and if the marked category makes a distinction along some parameter
(or length for that matter) the unmarked category is very likely to do so as well, but not vice
versa. Thus, I expect singleton nasals also make a distinction among alveolars, retroflexes, and
palatals, just as their geminate counterparts do, even though I have not yet found any (near)
minimal pairs for the former case.
Finally, Pair (22) shows the contrast between voiceless unaspirated and aspirated velars
while Pair (23) illustrates the contrast between modal-voiced and breathy-voiced velars. Here
we see aspiration and phonation type are phonemic in voiceless and voiced velars respectively.
Again, it is highly unlikely that velars are the only plosives that demonstrate these two
parameters, so I assume bilabial, dental, and retroflex plosives also make the distinction
between voiceless unaspirated and aspirated ones as well as modal-voiced and breathy-voiced
ones.
2.1.1.2. Distribution
Table 4 shows the distribution of each consonant phoneme within a word, i.e., whether a
particular phoneme occurs at the word-initial, word-medial, or word-final position.
1
The alveolar plosive is a bit tricky. Nevertheless, since it is the sole member in that category, I do not expect it to
demonstrate a phonemic contrast in length.
10
Table 4: Examples of Malayalam consonant phonemes
W-Initial W-Medial W-Final
/p/ paːle ‘milk’ apakaʈam ‘accident’
peːn ‘louse’ aːpat̪ ːə ‘dangerous’
para ‘rice barn’ apasurt̪ i ‘bad rumor’
paɻam ‘banana’ apasuaɾʲam ‘bad voice’
paːləm ‘bridge’ apamaɾʲad̪ a ‘bad manner’
/pː/ upːə ‘salt’
opːə ‘signature’
karupːə ‘black’
apːa ‘dad’
epːam ‘when’
/pʰ/ pʰalam ‘fruit; result’ safalam ‘come true’
(also falam) (but NOT sapʰalam)
pʰalikːuka ‘bear fruit’ d̪ uɹfalam ‘fruitless’
(also falikːuka) (but NOT d̪ uɹpʰalam)
falagam ‘(wood) plank’
faɹgunam ‘Aquarius’
faɳam ‘head of a serpent’
/b/ balam ‘strength’ aːba ‘Father’
balaʋaːn ‘strong man’ n̪ iɕabd̪ am ‘silent’
baːlika ‘girl’ kombə ‘cattle horn’
aːɖəmbaɾʲam ‘celebration’
kuʈumbam ‘family’
/bʱ/ bʱumi ‘earth’ (also fumi) abʱinajam ‘acting’
bʱajam ‘fear’ (also fajam) abʱajam ‘help’
bʱaɾʲa ‘wife’
bʱaɹt̪ aːʋə ‘husband’
bʱaːɾʲam ‘heavy’
/t̪ / t̪ ara ‘floor’ pat̪ a ‘bubble’
t̪ oli ‘skin’ paːt̪ a ‘road’
t̪ eːn ‘honey’ mat̪ i ‘enough’
t̪ iɾʲijuka ‘turn’ mat̪ il ‘barrier’
/t̪ ː/ kat̪ ːi ‘knife’
ʧiːt̪ ːa ‘dirty’
ʋit̪ ːə ‘seed’
kaɻut̪ ːə ‘neck’
ut̪ ːamam ‘perfect’
11
/t̪ ʰ/ at̪ ʰaʋaː ‘otherwise; also’
/d̪ / d̪ eiʋam ‘god’ nad̪ i ‘lake’
d̪ iɕa ‘direction’ maːd̪ aʋə ‘mother’
d̪ iʋasam ‘day’ end ̪ ̪ ə ‘what’
und ̪ ̪ uka ‘push’
ɕard̪ ikːuka ‘vomit’
/d̪ ʱ/ d̪ ʱanam ‘money’ ad̪ ʱipan ‘honor’
/tʰʲ/ kaːtʰʲə ‘wind’
maːtʰʲə ‘mat’
n̪ aːtʰʲəm ‘foul smell’
ʈatʰʲu ‘tatoo’
/ʧ/ ʧaːɾʲam ‘ash’
ʧiːt̪ ːa ‘dirty’
ʧuʈə ‘warm’
ʧeʋi ‘ear’
ʧoɾʲa ‘bloːd’
/ʧː/ murʧːa ‘sharp’
aʧːən ‘father’
kuraʧːi ‘few’
paʧːa ‘green’
iraʧːi ‘flesh’
/ʧʰ/ ?? ?? ??
/ʤ/ ʤiːʋi ‘creature’ aɲʤə ‘five’
ʤiːʋikːuka ‘live’ neɲʤə ‘chest’
ʤenanam ‘birth’
ʤaːm ‘jam’
ʤiɾʲakəm ‘cumin’
/ʤʱ/ ?? ?? ??
/ʈ/ ʈatʰʲu ‘tatoo’ kaʈi ‘bite’
ʈauɳ ‘town’ naʈi ‘actress’
iʋiʈe ‘here’
ɾʲaɳʈə ‘two’
aʈiʧːu ‘beat’
/ʈː/ kaʈːi ‘thick’
paʈːi ‘dog’
kuʈːi ‘child’
muʈːə ‘knee; joint’
muʈːa ‘egg’
12
/ʈʰ/ ʈʰipːu ‘name of a king’
/ɖ/ ɖokʈəɹ ‘doctor’ aːɖəmbaɾʲam ‘celebration’
ɖamaɾʲu ‘a small drum’
ɖakːu ‘rubber’
ɖaɖimam ‘pomegranate tree’
/ɖʱ/ ɖʱakːa ‘large drum’ aːɖʱen ‘powerful man’
ɖʱamaːnam ‘kettle’
/k/ kaːlə ‘leg’ maːsika ‘magazine’
kai ‘hand’ ahaŋkaːɾʲam ‘pride’
karupːə ‘black’ ʤiɾʲakəm ‘cumin’
kaɭikːuka ‘play’ ʤiːʋikːuka ‘live’
kaɾʲanam ‘because’ apakaʈam ‘accident’
/kː/ muːkːə ‘nose’
naːkːə ‘tongue’
ɾʲakːt̪ əm ‘blood’
t̪ akːaɭi ‘tomato’
kaɻikːuka ‘eat’
/kʰ/ kʰeːd̪ am ‘distress; sorrow’ mukʰam ‘face’
kʰaɹʧːu ‘expense’ sukʰəm ‘good; health’
kʰani ‘(gold) mine’
kʰananam ‘digging’
/g/ gaɳam ‘group’ aŋgəɭ ‘uncle’
garʋə ‘pride’ agn̪ i ‘fire’
guɳam ‘virtue’
guɾʲu ‘guru’
got̪ ambə ‘wheat’
/gʱ/ gʱaʈikaɾʲam ‘clock’ nagʱam ‘claw/nail’
gʱanam ‘heaviness’ meːgʱam ‘cloud’
(also kanam) aːgʱoːɕam ‘celebration’
gʱaʈːam ‘landing place’ aːgʱoːɕikːuka ‘celebrate’
gʱoːɾʲam ‘frightful’
gʱoːɕam ‘noise’
/m/ maɕi ‘ink’ bʱumi ‘earth’ d̪ iʋasam ‘day’
maɻa ‘rain’ kombə ‘cattle horn’ ʋanam ‘forest’
meliɲːa ‘thin’ maːmsəm ‘flesh’ paɻam ‘banana’
muʈi ‘hair’ got̪ ambə ‘wheat’ ʋaɹɕam ‘year’
muri ‘room’ ʋiɕamam ‘sadness’ paːd̪ əm ‘foot’
13
/mː/ namːaɭ ‘we’
amːa ‘mom’
/n̪ / n̪ anaɲa ‘wet’ end
̪ ̪ ə ‘what’
n̪ anaʋə ‘wetness’ und
̪ ̪ uka ‘push’
n̪ eːɾʲe ‘straight’ ʧand̪ ̪ rən ‘moon’
nei
̪ ̯ jə ‘fat; butter’ agn̪ i ‘fire’
n̪ isaːɾʲam ‘unimportant’
/n̪ ː/ on̪ ːə ‘one’
mun̪ ːə ‘three’
ʋan̪ ːu ‘came’
kon̪ ːu ‘killed’
ʧuʋan̪ ːa ‘red’
/n/ nad̪ i ‘lake’ manuɕen ‘person’ manuɕen ‘person’
naːkːə ‘tongue’ eŋəne ‘how’ aʧːn ‘fat̪ her’
niɻaɲə ‘full’ ʋaːnam ‘sky’ ɲaːn ‘I’
neɲʤə ‘chest’ aːna ‘elephant’ peːn ‘louse’
n̪ eːɾʲe ‘straight’ aɲʤə ‘five’ t̪ eːn ‘honey’
/nː/ anːa ‘Anna’
panːa ‘bad’
manːa ‘manna’
pinːe ‘then; later’
/ɲ/ ɲaːn ‘I’
ɲaŋaɭ ‘us’
/ɲː/ maɲːa ‘yellow’
n̪ anaɲːa ‘wet’
ʋiːd̪ ikuraɲːa ‘narrow’
niɻaɲːə ‘full’
kuɲːə ‘child’
/ɳ/ ʋiːɳu ‘fall’ ʈauɳ ‘town’
maɳam ‘smell’ muɳ ‘moon’
paɳam ‘money’
aːɳə ‘male’
peɳə ‘female’
/ɳː/ kaɳːə ‘eye’
ʋaɳːam ‘fat’
maɳːə ‘soil; sand’
14
/ŋ/ t̪ aːŋkəɭ ‘you’
aŋə ‘you’
t̪ eŋə ‘coconut tree’
uraŋuka ‘sleep’
uɳaŋia ‘dry’
/r/ raʋa ‘Rava flower’ t̪ ara ‘floor’
ramsaːn ‘Festival’ ʧorə ‘rice’
raːn̪ ːi ‘place name’ muri ‘room’
raɳi ‘queen’ mərkam ‘animal’
ruːpa ‘rupee’ apasurt̪ i ‘bad rumor’
/ɾʲ/ ɾʲakːt̪ əm ‘blood’ ʧoɾʲa ‘blood’
ɾʲaʋile ‘morning’ ʧiɾʲikːuka ‘laugh’
ɾʲaɳʈə ‘two’ t̪ aɾʲu ‘give’
ɾʲat̪ ːri ‘night’ ʈuɾʲe ‘far’
ɾʲaʤaʋə ‘king’ peːɾʲə ‘name’
/s/ sanʈoɕam ‘happiness’ maːsika ‘magazine’
samajam ‘time’ panʤasaːɾʲa ‘sugar’
sukʰəm ‘good; health’ kaseɾʲa ‘chair’
simham ‘lion’
st̪ riː ‘female’
/ɕ/ ɕaɾʲiɾʲam ‘body’ maɕi ‘ink’
ɕaːɾʲiɾʲam ‘beauty’ ʋaɕi ‘stubbornness’
ɕaɾʲi ‘correct’ manuɕen ‘person’
ɕard̪ ikːuka ‘vomit’ ʋiɕam ‘poison’
ʋiɕeɕam ‘news’
/ʂ/ iʂʈam ‘liking’
/h/ haːnikaɾʲam ‘dangerous’ mahaːn ‘expert’ ’
himam ‘snow’ ahaŋkaːɾʲam ‘pride’
hərd̪ ajam ‘heart’ aːhaːɾʲam ‘food’
mahaɾʲaʤaʋə ‘great king’
/f/ fajam ‘fear’ safalam ‘come true’
fumi ‘earth’ d̪ uɹfalam ‘fruitless’
faɳam ‘head of a serpent’
falagam ‘(wood) plank’
faɹgunam ‘Aquarius’
/ʋ/ ʋiːɳu ‘fall’ aʋan ‘he’
ʋaʈi ‘stick’ ʧeʋi ‘ear’
ʋere ‘other’
15
/ɻ/ maɻa ‘rain’
paɻam ‘banana’
kaɻiʋə ‘skill’
kaɻut̪ a ‘donkey’
d̪ aɻe ‘below; under’
/j/ joːni ‘skin’ kaijə ‘hand’
samajam ‘time’
ʧerija ‘small’
majil ‘peacock’
t̪ iɾʲijuka ‘turn’
/l/ laikːuka ‘dissolve’ mala ‘mountain’ kuʈal ‘intestines’
liːna ‘person name’ mula ‘breast’ kaʈal ‘ocean’
ʋalikːuka ‘pull’ t̪ uʋal ‘feather’
ʋalud̪ ə ‘big’ kuʈil ‘hut’
kaːlə ‘leg’ mat̪ il ‘barrier’
/lː/ pulːə ‘grass’
kalːə ‘stone’
mulːa ‘jasmin’
malːi ‘coriander’
kolːuka ‘kill’
/ɭ/ kiɭi ‘bird’ aʋaɭ ‘she’
muɭakə ‘chili’ kaɾʲaɭ ‘liver’
n̪ aːɭə ‘day’ epːoːɭ ‘when’
keɭkːuka ‘hear’ namːaɭ ‘we’
ʋaɾʲaɭʧːa ‘dry’ ɲaŋaɭ ‘us’
/ɭː/ ʋaɭːi ‘string’
kaɭːi ‘female thief’
uɭːi ‘onion’
ʋaɭːam ‘water’
uɭːa ‘have’
16
4. /tʰʲ/, /ŋ/, and /ɻ/ occur only at the word-medial position.
5. No consonant occurs only at the word-final position.
6. /ɭ/ and /ɳ/ never occur at the word-initial position.
7. The voiceless aspirated dental and breathy-voiced dental plosive (i.e. /t̪ ʰ/ and /d̪ ʱ/) are
extremely restricted (in terms of both the consultant’s personal lexicon and the words that
can be found in the dictionary), and so are their retroflex counterparts (i.e. /ʈʰ/ and /ɖʱ/).
With regard to the last point, two hypotheses can be made. For one thing, it may suggest that
some historically vigorous phonemes have now been marginalized over time so that they can
only be found in some archaic words. For the other, this may also be the result of language
contact through which some new phonemes are created. Since the lingua franca in India is
Hindi, which is famous for its four-way distinction in plosives (i.e. voiceless aspirated,
voiceless unaspirated, modal-voiced, and breathy-voiced), the language contact hypothesis
seems to be more plausible.
Also relevant is the fact that the consultant has a hard time pronouncing breathy-voiced
plosives and tends to substitute them with voiceless aspirated ones in the same place of
articulation. During an independent elicitation session with the consultant (see the file
Haowen_11.18.2009 in the database), I asked her to come up with words that have the
voiceless aspirated dental plosive (i.e. /t̪ ʰ/) in it, and she gave me [t̪ ʰanam] ‘money’ and [at̪ ʰipan]
‘honor’. However, after looking up these two words in the dictionary, she confessed having
made a mistake. As it turns out, they should in fact be [d̪ ʱanam] and [ad̪ ʱipan] respectively. She
also admitted that Malayali people always have problems with breathy-voiced plosives and
oftentimes confuse them with their voiceless aspirated counterparts. This anecdote might
suggest breathy-voiced plosives are probably not part of the original Malayalam phoneme
inventory and that their existence in contemporary Malalayalm is due to contact with Hindi or
other Indo-Aryan languages in India (e.g. Marathi, which is spoken to the north of Kerala).
Finally, an interesting thing about Malayalam consonants is that the voiceless aspirated
bilabial plosive /pʰ/ alternates with [f], at least in some words. For instance, the word for “fruit;
result” is either [pʰalam] or [falam], and the word for “bear fruit” is either [pʰalikːuka] or
[falikːuka]. Moreover, the breathy-voiced bilabial plosive /bʱ/ sometimes alternates with [f] as
well. For example, the word for “earth” is either [bʱumi] or [fumi] and the word for “fear” is
either [bʱajam] or [fajam]. I believe this is because Malayalam speakers tend to confuse /bʱ/
with /pʰ/. As a result, both /bʱ/ and /pʰ/ alternate with [f] in some cases. Also, [f] is most often
found in loan words.
17
2.1.2. Vowels
Malayalam has 11 monothongs and 5 diphthongs, as shown in Table 5 and 6 respectively.
Like in some consonants, length is phonemic in all monophthongs except for the schwa, whose
status is rather vague at this point. A diphthong is falling if it starts with a vowel of higher
prominence and ends in a semivowel with less prominence (e.g. /ai/), and is rising if the case is
the other way around (e.g. /ia/). Also, a diphthong is closing if it starts with a more open
element and ends in a more close element (e.g. /au/), and is opening if the case is the other way
around (e.g. /ua/). As shown in Table 6, all falling diphthongs are closing and all rising ones are
opening, which is in accord with the typological tendency.
18
(28) /a/-/ə/ muʈːa ‘egg’ muʈːə ‘knee; joint’
(29) /a/-/u/ palːə ‘tooth’ pulːə ‘grass’
puɻa ‘river’ puɻu ‘worm’
(30) /a/-/aː/ ʋaɾʲu ‘come’ ʋaːɾʲu ‘pick’
ʋanam ‘forest’ ʋaːnam ‘sky’
paɾʲa ‘rice barn’ paːɾʲa ‘rock’
(31) /a/-/ai/ kajə ‘fruit’ kaijə ‘hand’
(32) /i/-/e/ iʋiʈe ‘here’ eʋiʈe ‘where’
ipːoːɭ ‘now’ epːoːɭ ‘when’
iŋəne ‘like this’ eŋəne ‘how’
(33) /e/-/ə/ n̪ aːɭe ‘tomorrow’ n̪ aːɭə ‘day’
(34) /u/-/o/ upːə ‘salt’ opːə ‘signature’
Pair (26)~(29) show that /a/ is different from /i/, /e/, /u/, and /ə/. I believe /a/ is also different
from /o/ although I have not found any minimal pair to show this. There is, however, some
indirect evidence. Since Pair (32) illustrates the contrast between /i/ and /e/, and Pair (34) the
contrast between /u/ and /o/, and that /a/ is different from /i/, /e/, and /u/ for sure, it makes more
sense if /a/ is also distinct from /o/. In addition, Pair (30) shows that the difference between
short /a/ and long /aː/ is phonemic, and I believe length is also phonemic in other
monophthongs (except for the schwa) although for now no minimal pairs have been found to
illustrate this. However, my belief is based on two grounds: (a) the consultant is very sensitive
to the difference between short and long vowels; (b) a vowel system with only one pair of short
and long vowels would be highly implausible. Finally, Pair (31) shows the contrast between the
monophthong /a/ and the diphthong /ai/, which confirms that Malayalam has diphthongs as a
separate category that is phonologically distinct from monophthongs.
2.1.2.2. Distribution
Table 8 shows the distribution of each proposed vowel phoneme.
19
Table 8: Examples of Malayalam vowel phonemes
W-Initial W-Medial W-Final
/a/ at̪ ə ‘that’ ɕaɾʲiɾʲam ‘body’ para ‘rice barn’
aʋaɭ ‘she’ muɭakə ‘chili’ puɻa ‘river’
aʋaːd̪ aɭam ‘failure’ maraʋi ‘amnesia’ pat̪ a ‘bubble’
arijuka ‘know’ paɻajad̪ ə ‘old’ mula ‘breast’
agn̪ i ‘fire’ auɕat̪ am ‘medicine’ muɹʧːa ‘sharp’
/aː/ aːɳə ‘male’ maːd̪ aʋə ‘mother’
aːɾʲə ‘who’ naːkːə ‘tongue’
aːrə ‘six; river’ aːhaːɾʲam ‘food’
aːna ‘elephant’ baːlika ‘girl’
aːpat̪ ːə ‘danger’ elːaːm ‘all’
/i/ ila ‘leaf’ arijuka ‘know’ iraʧːi ‘meat; flesh’
ipːoːɭ ‘now’ aʈiʧːu ‘beat’ joːni ‘skin’
iɾʲikːuka ‘sit’ ʤiɾʲakəm ‘cumin’ agn̪ i ‘fire’
iʈad̪ ə ‘leftside’ eʋiʈe ‘where’ bʱaŋi ‘beauty’
iʋiʈe ‘here’ mat̪ il ‘barrier’ d̪ aʈi ‘fat; heavy’
/iː/ ʤiːʋi ‘creature’ st̪ riː ‘female’
miːn ‘fish’
niːn̪ d̪ a ‘long’
ʋiːd̪ i ‘broad’
ʋiːʈə ‘house’
/u/ uɭːi ‘onion’ kuɲːə ‘child’ aʈiʧːu ‘beat’
upːə ‘salt’ aʈut̪ ːə ‘near’ t̪ aɾʲu ‘give’
/u/ und ̪ ̪ uka ‘push’ d̪ ukʰəm ‘sadness’ epoɻu ‘always’
ut̪ ːamam ‘perfect’ guɳam ‘virtue’ kon̪ ːu ‘killed’
uraŋuka ‘sleep’ kaɻut̪ ːə ‘neck’ ʋiːɳu ‘fall’
/uː/ uːɳə ‘food’ guːʈːam ‘crowd’ puː ‘flower’
muːkːə ‘nose’
ruːpa ‘rupee’
t̪ uːkːuka ‘wipe’
/e/ erukːuka ‘take’ peɳːə ‘female’ aŋəne ‘like that’
elːaːm ‘all’ kaseɾʲa ‘chair’ d̪ aɻe ‘below; under’
epːam ‘when’ meliɲːa ‘thin’ iʋiʈe ‘here’
eʋiʈe ‘where’ ʧeʋi ‘ear’ n̪ eːɾʲe ‘straight’
epoɻu ‘always’ ʋeɳʈa ‘okra plant’ paːle ‘milk’
20
/eː/ eːɳi ‘ladder’ keːɭkːuka ‘hear’
eːt̪ ə ‘which’ meːkʰam ‘cloud’
n̪ eːɾʲe ‘straight’
peːʈi ‘fear’
ʋeːʈːa ‘hunt’
/o/ on̪ ːə ‘one’ kolːuka ‘kill’ kaɻikːuno ‘eating’
opːə ‘signature’ epoɻu ‘always’ kuɻikːuno ‘digging’
oɾʲu ‘one (thing)’ goːt̪ ambə ‘wheat’ kuɾʲakːuno ‘barking’
oɻukːuno ‘flowing; floating’ ʧorə ‘rice’
oɾʲupaːʈə ‘many’ t̪ oli ‘skin’
/oː/ oːla ‘palm leaf’ ipːoːɭ ‘now’
apːoːɭ ‘then’
epːoːɭ ‘when’
joːni ‘skin’
st̪ oːt̪ ram ‘thanks’
/ə/ iŋəne ‘like this’ ʋajarə ‘belly’
aŋəne ‘like that’ t̪ aiɾʲə ‘yogurt’
eŋəne ‘how’ t̪ eŋə ‘coconut tree’
purət̪ ə ‘outside’ ʧirakə ‘wing’
ʧand ̪ ̪ rən ‘moon’ ʧuʈə ‘warm’
/ai/ aikːam ‘unity’ kaijə ‘hand’
ʋaikːuka ‘play (instruments)’
laikːuka ‘dissolve’
/au/ auɕat̪ am ‘medicine’ jauʋaɳam ‘youth’
/ei/ d̪ eiʋam ‘god’
n̪ eijə ‘fat; butter’
/ia/ uɳaŋia ‘dry’
maŋia ‘dim’
naɾʲaŋia ‘lemon’
/ua/ suasəm ‘breath’
suaɾʲam ‘voice’
1. Unlike geminate consonants, which only occur at the word-medial position, geminate
vowels can occur at all positions, as evidenced by the distribution of /uː/.
2. Unlike monophthongs, diphthongs have a rather limited distribution. For instance, while
21
/ei/ and /ua/ occur only at the word-medial position, /ia/ occurs only at the word-final
position.
3. The schwa never occurs at the word-initial position.
These unreleased allophones are in fact recognized in the Malayalam writing system, where
released and unreleased consonants are represented by two different symbols.
22
syllables are reversed.
Additionally, the palatalization phenomenon is not restricted to the suffix /-kːuka/ even
though the examples above all end with it. In fact, the same palatalization pattern also takes
place in monomorphemic words. Compare for instance [aikːʲam̚ ] ‘unity’ with [t̪ akːaːɭi]
‘tomato’ and [muːkːə] ‘nose’. Once again, the palatalized geminate occurs immediately after
2
Contrary to the expectation, the intervocalic [k] in [abakaʈam̚ ] ‘accident’ is not voiced. Since the apa- here a
prefix meaning “bad”, I suspect there might be a constraint that prevents voicing from happening across
morpheme boundaries, though more data are needed in order to confirm this constraint.
3
According to Piotr, the /u/ vowel is fronted to [ʉ] when it immediately follows a palatalized geminate. Since
more acoustic evidence is needed in order to verify this, I transcribed all instances of /u/ as [u], which would be
sufficient for the current purpose.
23
high front vowels, but not after other vowels.
The maximal structure of an onset is “st̪ r” (as in /st̪ riː/ ‘female’), which is possibly the only
three-consonant onset type in Malayalam. The order of its consonant types is, from the nucleus
outwards, liquids, plosives, and fricatives. This order violates the Sonority Sequencing
Principle, which states that sonority progressively decreases towards the onset and coda from
24
the nucleus, since /s/ (or fricatives) has higher sonority than /t̪ / (or plosives) but is further away
from the nucleus than /t̪ /.
2.3.1. Onset
Most consonants in Table 1 may occur in the onset position, as shown in (2).
25
/ɲ/ ɲaːn ‘I’
/ɲː/ ma.ɲːa ‘yellow’
/ŋ/ u.ra.ŋu.ka ‘sleep’
/r/ ra.ɳi ‘queen’
/ɾʲ/ ʈu.ɾʲe ‘far’
/s/ sim.ham ‘lion’
/ɕ/ ɕa.ɾʲi ‘correct’
/h/ hi.mam ‘snow’
/ʧ/ ʧe.ʋi ‘ear’
/ʧː/ i.ra.ʧːi ‘flesh’
/ʤ/ ʤiː.ʋi ‘creature’
/ʋ/ ʋa.ʈi ‘stick’
/ɻ/ pa.ɻam ‘banana’
/j/ ma.jil ‘peacock’
/l/ mu.la ‘breast’
/lː/ ma.lːi ‘coriander’
/ɭ/ mu.ɭa.kə ‘chili’
/ɭː/ u.ɭːi ‘onion’
Consonant clusters are restricted to the type of plosives plus the trill (e.g. /ɾʲa.t̪ ːri/ ‘night’), with
a third optional /s/ at the beginning of the cluster (e.g. /st̪ riː/ ‘female’).
2.3.2. Nucleus
The nucleus can be any monophthong in Table 5 or any diphthong in Table 6, as shown in
(3).
26
/oː/ oː.la ‘palm leaf’
/ai/ ai.kːam ‘unity’
/au/ au.ɕa.t̪ am ‘medicine’
/ei/ d̪ ei.ʋam ‘god’
/ia/ ma.ŋia ‘dim’
/ua/ sua.səm ‘breath’
2.3.3. Coda
Candidates for the coda are much more restricted than those for the onset in terms of both
number and type. While the onset may allow up to three consonants in a row, the coda may not
consist of more than one consonant. With regard to type, on the other hand, only a handful of
consonant types are allowed in the coda whereas all but one consonant can occur in the onset.
Consonants that may occur in the coda are shown in (4).4
They roughly fall into two natural classes, namely, stops (orals and nasals) and liquids (laterals
and rhotics). However, not every member of these classes can occur in the coda. For instance,
as a member of liquids, /ɻ/ is not allowed in the coda.
4
Although the consultant seems to think of the sequence [nd ̪ ̪ ] in /ʧan̪ d̪ rən/ ‘moon’ as a unit, she made a pause
between [ʧan̪ ] and [d̪ rən] when Casey asked her to tap for the stress of this word. I took this as the justification for
syllabifying /ʧan̪ d̪ rən/ into [ʧan̪ .d̪ rən], rather than [ʧa.n̪ d̪ rən]. Also, it makes sense if /n̪ / can occur in the coda since
most nasals are allowed in the coda.
27
2.4. Typological Status
On the whole, the proposed phoneme inventory is quite plausible and well-balanced, except
for three relatively unusual matters. First, as the sole member of alveolar plosives, /tʰʲ/ is rather
unusual because it is both aspirated and palatalized, with no other counterparts in the alveolar
plosive category. For now, I have no extra comments on this particularity.
Second, the palatalized tap/flap /ɾʲ/ is a bit uncommon too. I suspect the tap/flap in
Malayalam is palatalized so as to increase its perceptual difference from the trill /r/ considering
the fact that they are two different phonemes but perceptually similar. Despite this, the
palatalized tap/flap is not completely bizarre anyway because Akamatsu (1997) for instance
argues that in Japanese /ɾʲ/ is a phoneme distinct from /ɾ/ (e.g. shoryaku [ɕoɾʲagɯ]
‘abbreviation’), though it occurs mostly in Sino-Japanese lexical items.
Third, while there is a voiceless alveolar fricative /s/, there is no voiced alveolar fricative /z/,
a phoneme quite common in many languages. However, lack of voicing contrasts in fricatives
is actually not rare at all. Maddieson’s research (2005) shows that about two-thirds of the
world's languages lack voicing contrasts in fricatives.
3. Morphology
To facilitate the inputting of data, I adopted a practical orthography of Malayalam for the
discussions of morphology and syntax instead of using IPA. The correspondence between IPA
and practical orthography is given in Appendix II.
The purpose of this section is not to present an exhaustive account of Malayalam
morphology, but to provide some examples for three major morphological processes, namely,
inflectional, derivational, and compounding. Morevoer, in terms of morphological typology,
Malayalam is more an agglutinating language than a fusional one since words tend to consist
of multiple morphemes and that when they do the morpheme boundaries are usually clear, as
shown by the multimorphemic words in (5) below.
28
(6) Gender: -an ‘masculine’; -i ‘feminine’
a. kaLL-an ‘thief (m.)’; kaLL-i ‘thief (f.)’
b. kaary-an ‘doer (m.)’; kaary-i ‘doer (f.)’
(7) Number: -ø ‘singular’; -kaL ‘plural’
a. puuca ‘cat’; puuca-kaL ‘cats’
b. paaTE ‘song’; paaTu-kaL ‘songs’
(8) Case: -ø ‘nominative’; -(y)il ‘locative’
a. muri ‘room’; muri-yil ‘in the room’
b. viiTE ‘house; home’; viiT-il ‘in the house; at home’
(9) Tense: -unnu ‘present’; -um ‘future’
a. paray-unnu ‘say (prs.)’; paray-um ‘will say’
b. samsaaryikky-unnu ‘talk (prs.)’; samsaaryikky-um ‘will talk’
(10) Mood: -u ‘imperative’; -aal ‘conditional’5
a. kaRikky-u ‘eat (imp.)’; kaRicc-aal ‘eat (cond.)’
b. coodikky-u ‘ask (imp.)’; coodicc-aal ‘ask (cond.)’
(11) Voice: -ø ‘active’; -appeT ‘passive’
a. koll-uka ‘to kill’; koll-appeT-uka ‘to be killed’
b. kariyaakk-uka ‘to tease’; kariyaakk-appeT-uka ‘to be teased’
5
The -aal conditional is suffixed to the past tense verb root.
29
Based on Asher & Kumari (1997) and Moag & Moag (1967), the third person demonstrative/
pronominal paradigm is given in Table 13.
By the same token, nominals can also be derived from verbs suffixed by the singular neuter
pronoun -tE when they first derive into participles (which are adjectival in nature) via -a
suffixation. The two-step derivational process is illustrated in Table 14.
30
Presumably, the adverb nannaayi ‘well’ is also the outcome of this derivational process.
4. Syntax
The organization of this section basically follows that of Thomas Payne’s A Sample
Grammatical Sketch of English 6 , which is intended as a sample grammatical sketch for
students who are expected to write about the descriptive morphosyntax of an unknown
language for a one-year course. Due to the paucity of my current data, however, some sections
in Payne’s sample are not included here, such as tense/aspect/mode, clause combinations, and
pragmatically marked structures. Accordingly, that leaves the topics covered in the following
sections to word classes, basic constituent orders, predicate nominals, existential/locational/
possessive sentences, expression of grammatical relations, causatives, passives, reflexives,
reciprocals, questions, imperatives, and finally negation.
4.1.1. Nouns
Malayalam nouns share four morphosyntactic properties. First, they inflect for number.
While singular is unmarked, plural is marked by -kaL, which can be suffixed to human,
nonhuman animate, and inanimate nouns, as respectively shown in (15) through (17).
6
Payne’s Sample is available online at http://www.uoregon.edu/~tpayne/engram.htm.
31
(15) anyjE aaLu-kaL oryutaryaayiTTE vannu.
five person-PL one.after.another come.PST
‘Five people came one after another.’ (RECIP.011)
(16) ryaNTE puucca-kaL-e vaangng-aam.
two cat-PL-ACC buy-HORT
‘Buy two cats.’ (NP.008)
(17) ii ryaNTE kaaru-kaL-il eet=aaNE ninakkE kuuTutal iSTam?
this two car-PL-LOC which=COP1.PRS 2SG.DAT more liking
‘Which of these two cars do you like more?’ (INTERROG.019)
Although -kaL marks plurality for most Nouns, there is a special plural marker -maary for
some kinship terms. For instance, amma is “mother” and its plural form is ammamaary (see
Onam.001 for the context it occurs in) instead of *ammakaL.
In addition to number, Malayalam nouns also inflect for case. In (18), for instance, each of
the five Nouns is marked by a different morphological case.
Third, Malayalam nouns do not inflect for gender. The only exceptions are third person
singular human pronouns and some human nouns. While “he” is avan (the familiar form) or
addeeham (the honorific form), “she” is avaL. Some human nouns may refer to males or
females, depending on the nominal ending. For example, a male singer is paattukaaryan and
a female one paattukaaryi. By the same token, a male sibling (i.e. brother) is sahooteryan and
a female one (i.e. sister) sahooteryi. It seems that -an marks masculine nouns and -i feminine
ones. But it is unclear at this point what kinds of nouns these gender markers may be affixed
to.
Finally, Malayalam nouns immediately precede postpositions and form a constituent with
them, as illustrated in (19).
4.1.2. Verbs
Like verbs in many other languages, Malayalam verbs inflect for tense, aspect, and mode
(or TAM for short). The complete range of grammatical TAM distinctions is not clear at this
point. For now, only some instances from each category can be given for illustration.
Tense. The citation form of Malayalam verbs ends with -uka, such as verukkuka ‘to hate’
and toRikkyuka ‘to kick’, and the citation form without -uka is the verb root. While present
tense is marked by -unnu suffixed to the verb root, future tense is marked by -um, as shown in
32
(20) through (23).
In past tense, a different verb root is used, which is unpredictable from the citation form. For
instance, the past tense verb stem of kaRikkyuka ‘to eat’ is kaRiccu and that of uNTaakkuka
‘to make’ is uNTaakki, as in (24) and (25) respectively.
Aspect. The citation form (with a vowel change at the end) combined with the copula aaNE
(present)/aayiryunnu (past) gives rise to progressive aspect, as shown in (26) and (27), where
the infinitive ending -uka changes to -uke.
Moreover, the past tense verb root (with deletion of the final vowel) combined with iryunnu
seems to indicate present perfect, as seen in (28) and (29).
33
Mode. The suffix -aNam, when suffixed to the (nonpast) verb stem, is associated with
deontic modality. In (30), for instance, varyaNam means “should come” while the citation
form of the verb (i.e. varyuka) simply means “to come”.
Finally, as can be seen from the examples above, there is no number or gender agreement
between a Malayalam verb and any of its arguments.
4.1.3. Adjectives
Malayalam adjectives do not inflect and can directly modify nouns, as nalla ‘good’ in (31)
shows.
One characteristic of Malayalam adjectives is that a great number of them end with -a,
which can be analyzed as an attributive morpheme, as shown in (32).
Another distinctive feature of Malayalam adjectives is that the attributive use and
predicative use of adjectives differ in form. Specifically, an adjective can modify a noun in
the attributive context without any change, but it has to be nominalized by the third person
neutral pronoun -tE (or its allmorph -t) if used in the predicative context. Examples in (33)
34
and (34) illustrate this contrast.
4.1.4. Adverbs
Malayalam adverbs do not inflect and they modify verbs, as illustrated in (35) and (36).
4.1.5. Postpositions
Malayalam postpositions are invariant and they take nouns as their complements, as
mumpE ‘before’ in (37) and varye ‘until’ in (38) illustrate.
4.1.6. Conjunctions
Like postpositions, Malayalam conjunctions are also invariant, but unlike postpositions
they link a whole clause, rather than a nominal phrase, to the main clause. This is shown by
pakshe ‘but’ and kaaryaNam ‘because’ in (39) and (40) respectively.
35
(39) joon sahaayikky-um, pakshe anna sahaayikky-attilla.
John help-FUT but Anna help-FUT.NEG
‘John will help, but Anna won’t.’ (COORD.004)
(40) amma uNTaakki-tary-unn-a bhakshaNatt-inE
mom made-give-PRS-ATTR food-DAT
ryucci kuuTutal aayiryikkyum,
taste more COP1.FUT
kaaryaNam sneeham ennE paray-unn-a masaala
because love QUOT call-PRS-ATTR spice
kuuTi at-il-E ceerkk-um.
also that-LOC-DAT add-FUT
‘Foods made and given by the mother will be more delicious because a spice called love
will also be added (to them).’ (Onam.001&002)
(41) S V
joon=um anna=yum samsaaryiccu.
John.NOM=COORD PN.NOM=COORD talk.PST
‘John and Anna talked.’ (T&A.005)
(42) S O V
aa kuTTi avaNTe sahooteryi-ye iTiccu.
that child.NOM 3SG.M.GEN sister-ACC hit.PST
‘The boy hit his sister.’ (NEG.005)
Despite the basic S(O)V word order, Malayalam may have a relatively free word order
system since each argument in a clause is always marked by case and its grammatical relation
to the predicate is thus unambiguous. Example (43) shows a case where four grammatical
relations occur in the same clause, and it might just be one of the many word order
possibilities. However, more naturally occurring data are needed in order to find out to what
degree the word order is “free”.
(43) S IO OBL DO V
enre amma enikkyE koci-yil ninnE oryu kattE-ø ayaccu.
1SG.GEN mom.NOM 1SG.DAT Cochin-LOC from one letter-ACC send.PST
‘My mom sent me a letter from Cochin.’ (Case.009)
36
(44) anyjE aaLu-kaL oryutaryaayiTTE vannu.
five person-PL one.after.another come.PST
‘Five people came one after another.’ (RECIP.011)
(45) [aa kuTTi] [avaNTe sahooteryi-ye] iTiccu.
that child.NOM 3SG.M.GEN sister-ACC hit.PST
‘The boy hit his sister.’ (NEG.005)
(46) ni oryu nalla paaTTukaaryan aayiryunnu.
2SG.NOM one good singer.M COP1.PST
‘You used to be a good singer.’ (T&A.024)
(47) amma uNTaakki tary-unn-a bhakshaNatt-inE
mom make.CVB give-PRS-ATTR food-DAT
ryucci kuuTutal aayiryikkyum.
taste more COP1.FUT
‘Foods made and given by Mom will be more delicious.’ (Onam.001)
Importantly, not every postposition requires its nominal complement to be marked in the
same grammatical case. For instance, while nominals headed by vaRi ‘via’ and varye ‘until’
are in nominative, those headed by nunnum ‘from’ and iTakkyE ‘between’ should be marked
by locative and genitive respectively, as in (48) and (49).
37
(51) ni oryu nalla paaTTukaaryan aayiryunnu.
2SG.NOM one good singer.M COP1.PST
‘You used to be a good singer.’ (T&A.024)
Interestingly, if the subject NP is the one and only thing that exists, a special construction
is used. Specifically, the predicate becomes uLLu and the subject NP is marked by -ee, which
seems to express emphatic meaning of some sort. Example (53) is illustrative.
In locational clauses, the subject NP precedes the locative phrase and the same existential
uNTE serves as the predicate at the end, as in (54).
The existential predicate uNTE is also used in possessive clauses, where the possessor is
marked in dative case and the possessum in nominative. The possessum used in this
construction can be body parts or kinship terms, as respectively shown in (55) and (56).
Like the case in existential/locational clauses, the same “-ee plus uLLu” construction is
used if the possessum is the only entity that is possessed, as illustrated in (57) and (58) below.
38
(57) avannE oryu kai=yee uLLu.
1SG.DAT one hand.NOM=EMPH EX.PRS
‘He has only one hand.’ (NP.023)
(58) enikkyE oryu sahooteryan=ee uLLu.
1SG.DAT one brother.NOM=EMPH EX.PRS
‘I have only one brother.’ (NP.027)
However, if the object is inanimate, the accusative marking can be zero, as in (60), where the
grammatical relation is indicated by word order.
When there are two objects in the same clause, direct object is marked by accusative and
indirect object by dative, as in (61).
In addition to a recipient, the dative case also marks goal of motion, which can be
alternatively marked by -ooTE, whose function is still unclear. Example (62) illustrates this
point.
Location, whether static or dynamic, is marked by locative case -il, as in (63) and (64) below.
39
(63) kuTTi muri-yil ninnE puratt-ekkyE ooTi.
child.NOM room-LOC from outside-DAT run.PST
‘The child ran out of a room.’ (Case.008)
(64) frij-il oryu kooRi uNTE.
fridge-LOC one chicken.NOM EX
‘There is a chicken in the fridge.’ (NP.016)
4.6. Causatives
In terms of transitivity, Malayalam seems to have three verb stems, namely, intransitive,
transitive, and causative stems, as shown in (65) through (67) respectively.
Comparing (66) with (67), we see that the causative stem is derived from the transitive stem
infixed by the causative morpheme <ppi>. As (67) illustrates, the argument structure in
Malayalam causative constructions is marked by nominative for the causer, accusative for the
patient of the caused event, and the posposition koNTE ‘by’ (which “governs” accusative) for
the causee.
Interestingly, as an alternative to the causative stem, the transitive stem can also be used in
the causative context (where there are causer, causee, and patient of the caused event). This
alternative, however, is lexically restricted, as the contrast in (68) shows.
While the transitive stem poTTiccu and its synomym uTaccu are interchangeable in the
transitive context (as in (66)), only poTTiccu, but not uTaccu, is acceptable in the causative
context.
In spite of the functional overlap between the transitive and the causative stem (such as
40
poTTiccu and poTTippiccu), the two stems should be distinguished as they denote different
event structures. This difference in event structure is especially clear when the causee is
absent, as the minimal pair in (69) shows. The transitive stem poTTiccu can only have the
transitive interpretation when the caussee is absent whereas the causative stem poTTippiccu
always has the causative interpretation even if the caussee is absent.
4.7. Passives
Malayalam passive verbs are derived from the present tense root encliticized by the passive
auxiliary appeT-uka ‘PASS-INF1’, whose past tense form is appeTTu ‘PASS.PST’. For instance,
while the active past tense form of koll-uka ‘kill-INF1’ is konnu, its passive counterpart is
koll=appeTTu, as the contrast in (70) and (71) show.
Aside from the difference in verb forms, the passive construction also differs from the
active construction in terms of case marking. In actives, the agent and patient are marked by
nominative and accusative respectively; in passives, the agent is marked by instrumental and
the patient by nominative. More examples of the passive construction are given in (72) and
(73), where the active infinitival verb forms are aTikky-uka ‘hit-INF1’ and kattikky-uka
‘burn-INF1’ respectively.
41
In addition, the agent in passives is optional, as is often the case in passives of many
languages. Example (74) is illustrative.
Syntactically, both (75) and (76) should be considered “active” since they only differ in word
order. But pragamatically, (76) is more “passive-like” than (75), for the patient is profiled and
the agent backgrounded in (76).
4.8. Reflexives
Malayalam makes use of the reflexive adverb tanne to express reflexivity, as shown in (77).
Without tanne, (77) would be ambiguous over two readings, as illustrated in (78).
4.9. Reciprocals
There are two ways to express reciprocality. One is to use the adverb tammil, which
42
literally means ‘among them’ and is probably the locative form of some word on account of
the -il ending in it. Example (79) is illustrative.
It is unclear what conditions the use of one way or the other, but the demonstrative adverb
strategy seems to be quite productive, as seen in (81) through (83).
4.10.Questions
4.10.1. Polar questions
Polar questions in Malayalam are formed by adding the interrogative clitic =oo to the
predicate of an affirmative clause, as in (84) and (85), which without =oo would have been a
statement rather than a question.
In the case of negative polar questions, the interrogative clitic =ee is attached to the
predicate of a negative clause, as in (86) and (87), which without =ee would have been a
negative statement rather than a question.
43
(86) samaaharyam uNTaakk=ill=ee?
collection make=NEG=NEG.Q
‘Don’t (people) make collections?’ (Onam.37)
(87) pinne namukkE ottiryi paaTTu-kaL=ill=ee?
and.then 1PL.INC.DAT many song-PL=NEG.EX.PRS=NEG.Q
‘And then don’t we have many songs?’ (Onam.33)
For content questions which involve a verb that is not the copula, a special cleft-like
construction is used, whereby a nominalized verb co-occurs with the copula, as in (90)
through (92).
4.11.Imperatives
Imperative mood in Malayalam is expressed by some suffixes attached to the nonpast Verb
stem, and the addressee is unexpressed but implied. There are several suffixes that can carry
out this job. First, the citation form ending -uka:
44
Second, the suffix -u, which might be a shortened form of -uka:
I call this suffix “deontic” because it adds a deontic meaning to the verb it is attached to, as in
(96). So the literal meaning of (95) might be “(You) should buy (some) milk.”
The evidence of -aam being a hortative suffix comes from sentences like (98), where the
suffix adds a “let’s” meaning to the verb it is attached to.
Finally, the prohibitive suffix -aNTa, which expresses a negative command or request:
4.12.Negation
Negation is formed by using the negative existential predicate illa, whose positive
counterpart is uNTE. Example (100) and (101) show a minimal contrast between these two
existential predicates.
45
(100) frij-il oryu kooRi uNTE.
fridge-LOC one chicken.NOM EX
‘There is a chicken in the fridge.’ (NP.016)
(101) frij-il kooRi illa.
fridge-LOC chicken.NOM NEG.EX.PRS
‘There is no chicken in the fridge.’ (NP.018)
The negative illa is by default in the present tense, so if it is used in other tenses the copula
is required so as to indicate tense, as in (102).
In clauses where the main verb is not a copula, the negative illa is suffixed to the verb. In
past tense, it is suffixed to the past tense verb root (with deletion of the final vowel), as in
(103).
In present tense, the verb is negated by -aarilla, where illa obviously indicates negation but
the function of aar is not clear. This is illustrated in (104).
In future tense, the verb is negated by -attilla, which again has a mysterious morpheme in it
(i.e. att), as shown in (105) below.
46
Appendix I: Glossary
47
ɖamaɾʲu ‘a small drum’ gaɳam ‘group’
d̪ aɻe ‘below; under’ garʋə ‘pride’
d̪ aʈi ‘fat; heavy’ gʱanam ‘heaviness’ (= kanam)
d̪ eiʋam ‘god’ gʱaʈːam ‘landing place’
ʤenanam ‘birth’ gʱaʈikaɾʲam ‘clock’
ɖʱakːa ‘large drum’ gʱoːɾʲam ‘frightful’
ɖʱamaːnam ‘kettle’ gʱoːɕam ‘noise’
d̪ ʱanam ‘money’ goːt̪ ambə ‘wheat’
ʤiːʋi ‘creature’ guːʈːam ‘crowd’
ʤiːʋikːuka ‘live’ guɳam ‘virtue’
ʤiɾʲakəm ‘cumin’ guɳanam ‘multiplication’
d̪ iɕa ‘direction’ guɾʲu ‘guru’
d̪ iʋasam ‘day’ haːnikaɾʲam ‘dangerous’
ɖokʈəɹ ‘doctor’ hərd̪ ajam ‘heart’
d̪ ukʰəm ‘sadness’ himam ‘snow’
d̪ uɹ- ‘bad’ iː ‘this (adnominal)’
eːɳi ‘ladder’ id̪ ːeːham ‘he (formal; proximal)’
eːt̪ ə ‘which’ -il ‘in; at’
elːaːm ‘all’ ilːa ‘no’
elːə ‘bone’ ila ‘leaf’
end
̪ ̪ ə ‘what’ iŋəne ‘like this’
eŋəne ‘how’ ipːoːɭ ‘now’
epːam ‘when’ iraʧːi ‘meat; flesh’
epːoːɭ ‘when’ iɾʲikːuka ‘sit’
epoɻu ‘always’ iʂʈam ‘liking’
erijuka ‘throw’ iʈad̪ ə ‘leftside’
erukːuka ‘take’ it̪ ə ‘this (pronominal)’
eʋiʈe ‘where’ iʈikːuka ‘hit’
fajam ‘fear’ iʋiʈe ‘here’
palaga ‘(wood) plank’ jauʋaɳam ‘youth’
falagam ‘(wood) plank’ joːni ‘skin’
falam ‘fruit; result; product’ (=pʰalam) -kːja ‘fruit’
faɳam ‘head of a serpent’ kaːkːa ‘crow’
faɹgunam ‘Aquarius’ kaːlə ‘leg’
48
kaːtʰʲə ‘wind’ kiʈakːuka ‘lie’
kai(jə) ‘hand’ kolːuka ‘kill’
kajaɹ ‘a piece of rope’ kolusːə ‘anklet’
kajə ‘fruit’ kombə ‘cattle horn’
-kaɭ ‘PL marker’ kon̪ ːu ‘killed’
kaɭːan ‘thief’ korukːuka ‘give’
kalːə ‘stone’ koʈum ‘strong (wind)’
kaɭːi ‘female thief’ kuɲːə ‘child’
kaɭi ‘game’ kuɾʲakːuno ‘barking’
kaɭikːuka ‘play’ kuraʧːi ‘few’
kaɳːə ‘eye’ kuɻikːuno ‘digging’
kanam ‘heaviness’ (=gʱanam) kuɾʲumuɭakə ‘pepper’
kaɾʲaɭ ‘liver’ kuʈːi ‘child’
kaɾʲanam ‘because’ kuʈal ‘intestines’
kaɻikːuka ‘eat’ kuʈikːuka ‘drink’
kaɻikːuno ‘eating’ kuʈil ‘hut’
kaɻiʋə ‘skill’ kuʈumbam ‘family’
kaɻukːuka ‘wash’ laikːuka ‘dissolve’
karupːə ‘black’ liːna ‘person name’
kaɻut̪ ːə ‘neck’ maːd̪ aʋə ‘mother’
kaɻut̪ a ‘donkey’ maːmsəm ‘meat; flesh’
kaseɾʲa ‘chair’ maːnam ‘sky’
kat̪ ːi ‘knife’ maːsika ‘magazine’
kaʈːi ‘thick’ maːtʰʲə ‘mat’
kaʈal ‘ocean’ maha- ‘great; big’
kaʈi ‘bite’ mahaːn ‘expert’
keːɭkːuka ‘hear’ majil ‘peacock’
keʈːuka ‘tie (v.)’ majuːɾʲam ‘peacock’
kʰananam ‘digging’ malːi ‘coriander’
kʰani ‘(gold) mine’ mala ‘mountain’
kʰaɹʧːu ‘expense’ manːa ‘manna’
kʰeːd̪ am ‘distress; sorrow’ maɲːa ‘yellow’
kiːɾʲi ‘mongoose’ maɲːə ‘snow; ice’
kiɭi ‘bird’ maɳːə ‘soil; sand’
49
maɳam ‘smell’ ɲaːn ‘I’
manasːə ‘mind’ n̪ aːtʰʲəm ‘foul smell’
manuɕen ‘person, human’ nad̪ i ‘lake’
maŋia ‘dim’ nagʱam ‘claw; nail’
maɻa ‘rain’ nalːat̪ ə ‘good’
mara ‘shade’ namːaɭ ‘we’
maɾʲad̪ a ‘manner’ n̪ anaɲːa ‘wet’
maɾʲam ‘tree’ n̪ anaʋə ‘wetness’
maraʋi ‘amnesia’ ɲaŋaɭ ‘us’
maɕi ‘ink’ naɾʲaŋia ‘lemon’
mat̪ i ‘enough’ naʈakːuka ‘walk’
maʈi ‘lap; laziness’ naʈi ‘actress’
mat̪ il ‘barrier’ n̪ eːɾʲe ‘straight’
meːgʱam ‘cloud’ n̪ ei(jə) ‘fat; butter’
-eːl ‘on’ neɲʤə ‘chest’
meliɲːa ‘thin’ nəɹd̪ ːajam ‘merciless’
mərkam ‘animal’ n̪ i- ‘un-; without’
miːn ‘fish’ ni(ŋaɭ) ‘you (casual)’
muːkːə ‘nose’ niːn̪ d̪ a ‘long’
mukʰam ‘face’ nin̪ ːu ‘stood’
mulːa ‘jasmin’ niɻaɲːə ‘full’
mula ‘breast’ n̪ isaːɾʲam ‘unimportant’
muɭakə ‘chili’ n̪ iɕabd̪ am ‘silent’
mun̪ ːə ‘three’ oːla ‘palm leaf’
muri ‘room’ on̪ ːə ‘one’
murikːuka ‘split’ opːə ‘signature’
murʧːa ‘sharp’ oɾʲu ‘one (thing)’
muʈːa ‘egg’ oɻukːuno ‘flowing; floating’
muʈːə ‘knee; joint’ oɾʲupaːʈə ‘many’
muʈi ‘hair’ paːd̪ aːɭam ‘underground’
naːkːə ‘tongue’ paːd̪ əm ‘foot’
n̪ aːɭə ‘day’ paːle ‘milk’
n̪ aːlə ‘four’ paːləm ‘bridge’
n̪ aːɭe ‘tomorrow’ paːra ‘rock’
50
paːt̪ a ‘road’ ɾʲaʤaʋə ‘king’
pajarə ‘beans’ ɾʲakːt̪ əm ‘blood’
palːə ‘tooth’ ramsaːn ‘Ramsaan Festival’
palaga ‘a piece of wood’ raɳi ‘queen’
panːa ‘bad’ ɾʲaɳʈə ‘two’
pana ‘palm tree’ ɾʲat̪ ːri ‘night’
paɳam ‘money’ raʋa ‘Rava flower’
panʤasaːɾʲa ‘sugar’ ɾʲaʋile ‘morning’
para ‘rice barn’ ruːpa ‘rupee’
paɻajad̪ ə ‘old’ ɕaːɾʲiɾʲam ‘beauty’
parakːuno ‘flying’ safalam ‘bear fruit; come true’
paɻam ‘banana’ samajam ‘time’
paʧːa ‘green’ saɳʈoɕam ‘happiness’
paʧːapːə ‘greenery’ ɕard̪ ikːuka ‘vomit’
paʈːi ‘dog’ ɕaɾʲi ‘correct’
pat̪ a ‘bubble’ ɕaɾʲiɾʲam ‘body’
paʈikːuka ‘study’ simham ‘lion’
peːn ‘louse’ sinima ‘cinema’
peːɾʲə ‘name’ st̪ oːt̪ ram ‘thanks’
peːʈi ‘fear’ st̪ riː ‘female’
peɳːə ‘female’ suaɾʲam ‘voice’
peɳkuʈːi ‘girl’ suasəm ‘breath’
pʰalam ‘fruit; result; product’ (= falam) sukʰəm ‘good; health’
pʰalikːuka ‘work out; bear fruit’ surt̪ i ‘rumor’
pinːe ‘then; later’ t̪ aːŋkəɭ ‘you (formal)’
piɻijuka ‘squeeze’ ʧaːɾʲam ‘ash’
piʈikːuka ‘hold’ t̪ aːraʋə ‘duck’
puː(ʋə) ‘flower’ t̪ aiɾʲə ‘yogurt’
puɻa ‘creek; pond’ t̪ akːaːɭi ‘tomato’
puram ‘back of the body’ t̪ ala ‘head’
purət̪ ə ‘outside’ ʧand
̪ ̪ rən ‘moon’
puɻu ‘worm’ t̪ ara ‘floor’
puɾʲuɕen ‘male’ t̪ aɾʲakəm ‘star’
raːn̪ ːi ‘place name’ t̪ aɾʲu ‘give’
51
ʧaʈːi ‘pot; vase’ ut̪ ːamam ‘perfect’
ʈatʰʲu ‘tatoo’ ʋa(jə) ‘mouth’
t̪ aʈi ‘wood’ ʋaːnam ‘sky’
ʈauɳ ‘town’ ʋaːɾʲu ‘pick’
ʧaʋi ‘key’ ʋaikːuka ‘play (instruments)’
t̪ eːn ‘honey’ ʋajaɹ ‘wire’
t̪ eŋə ‘coconut tree’ ʋajarə ‘belly’
ʧerija ‘small’ ʋalad̪ ə ‘rightside’
ʧeʈi ‘plant’ ʋalija ‘big’
ʧeʋi ‘ear’ ʋalikːuka ‘pull’
ʧiːpːə ‘comb (n.)’ ʋalud̪ ə ‘big’
ʧiːt̪ ːa ‘dirty’ ʋaɳːam ‘fat’
ʧin̪ d̪ ikːuka ‘think’ ʋan̪ ːu ‘came’
ʧirakə ‘wing’ ʋanam ‘forest’
t̪ iɾʲijuka ‘turn’ ʋaɾʲa ‘line’
ʧiɾʲikːuka ‘laugh’ ʋarakːuka ‘fry’
t̪ oli ‘skin’ ʋaɾʲaɭʧːa ‘dry’
ʧoɾʲa ‘blood’ ʋaɻi ‘road; way’
ʧorə ‘rice’ ʋaɹɕam ‘year’
t̪ uːkːuka ‘wipe’ ʋaɾʲu ‘come’
ʈuɾʲe ‘far’ ʋaɻut̪ anaŋia ‘eggplant’
t̪ uʈaikːuka ‘wipe’ ʋaɕi ‘stubbornness’
ʧuʈə ‘warm’ ʋatʰʲalamuɭakə ‘paprika chili’
t̪ uʋal ‘feather’ ʋaʈi ‘stick’
ʧuʋan̪ ːa ‘red’ ʋeːɳʈa ‘not’
uːɳə ‘food’ ʋeːʈːa ‘hunt’
uɭːa ‘have’ ʋeɭːam ‘water’
uɭːi ‘onion’ ʋeɭut̪ ːa ‘white’
uɳaŋia ‘dry’ ʋeɳʈa ‘okra plant’
und
̪ ̪ uka ‘push’ ʋere ‘other’
upːə ‘salt’ ʋərɳam ‘wound’
uraŋuka ‘sleep’ ʋiːd̪ i ‘broad’
ut̪ ːamabʱaɾʲa ‘ideal wife’ ʋiːd̪ ijulːa ‘wide’
ut̪ ːamabʱaɹt̪ aːʋə ‘ideal husband’ ʋiːd̪ ikuraɲːa ‘narrow’
52
ʋiːɳu ‘fall’
ʋiːʈə ‘house’
ʋinu ‘person name’
ʋiɕajam ‘subject; topic’
ʋiɕam ‘poison’
ʋiɕamam ‘sadness’
ʋiɕeɕam ‘news’
ʋit̪ ːə ‘seed’
53
Appendix II: A Practical Orthogrphy of Malayalam
Malayalam consonants
Alveolo-palatal
Labiodental
Retroflex
Alveolar
Bilabial
Glottal
Palatal
Dental
Voice
Velar
/p/=p /t/=t /tɕ/=c /ʈ/=T /k/=k
-
/pʰ/=ph /tʰ/=th /tɕʰ/=ch /ʈʰ/=Th /kʰ/=kh
Plosive
/b/=b /d/=d /ʤ/=j /ɖ/=D /g/=g
+
/bʱ/=bh /dʱ/=dh /ʤʱ/=jh /ɖʱ/=Dh /gʱ/=gh
Nasal + /m/=m /n/=n /ɲ/=ny /ɳ/=N /ŋ/=ng
Trill + /r/=r
Tap/flap + /ɾ/=ry
Fricative - /f/=f /s/=s /ɕ/=sh /ʂ/ =S /h/=h
Cen.
+ /ʋ/=v /ɻ/=R /j/=y
Approx.
Lat.
+ /l/=l /ɭ/=L
Approx.
Malayalam vowels
Front Central Back
/i/=i /u/=u
High
/iː/=ii /uː/=uu
/e/=e /o/=o
Mid /ə/=E
/eː/=ee /oː/=oo
/a/=a
Low
/aː/=aa
54
Part II: A Text of Malayalam
1. Introduction
The goal of this second part is to provide an interlinear glossed text of spoken Malayalam
and then to analyze the grammatical features as seen in the text. The chosen discourse is a TV
interview on Onam Festival7, which is the biggest festival in the state of Kerala, India, where
Malayalam is spoken on the daily basis. Onam Festival falls in the first Malayalam month (i.e.
between August and September), and is intricately linked to many aspects of Malayalam
culture, including songs, dancing, cuisine, dresses, flower decorations, and Snake Boat racing,
etc.
The interview took place between a female host and a male guest. The host, whose name is
Meera Krishnan, is a famous film actress in her early thirties, and the guest, whose name is
Padmasree Mohanlal (called Lal in the text), is a famous film actor and producer in his early
forties. Throughout the interview, Krishan asked Mohanlal a series of questions that are more
or less related to Onam Festival. While the whole interview lasted over sixteen minutes, only
about three minutes of it were transcribed due to the time constraint. In transcribing the
interview, I was assisted by a Malayalam consultant, whose name is Sona Joseph, a 26-year-old
female who speaks Malayalam natively and English fluently. She originally came from Kochi
(formerly Cochin), Kerala, India, and is currently a graduate student at Rice University.
The procedure of the transcription work goes as follows. I first played back the whole
interview to Sona, who then explained to me the general ideas of it. After that, we discussed
which part of it was most worth transcribing in terms of both speech clarity and cultural
relevance. Once the agreement was reached, I played back a small chunk of the interview at a
time and Sona repeated to me what she had heard on the recordings at a slower speed. After two
to three small chunks, we would stop transcribing, she would translate the recently transcribed
chunks, and I would ask her relevant questions in order to correctly parse and gloss the
transcriptions.
In what follows, I will first present a bird’s-eye view of the text, that is, a view of the text
from top down. In this view, the text, consisting of parallel Malayalam transcriptions and
English free translations, is organized in terms of conversational turns so that the readers would
have an overall picture of the text contents. Then I will present a worm’s-eye view of the text,
that is, a view of the text from bottom up. In this view, the text, this time consisting of three
interlinear tiers (i.e. text tier, morpheme-by-morpheme tier, and glossing tier), is chunked into
syntactic units roughly at the level of clauses or sentences so that the readers would have
7
The interview is available at http://www.youtube.com/watch?v=kXcD7PFE9aE&feature=channel.
55
detailed ideas about the morphosyntax, phonology, and morphophonology.
‘Our moms, foods made and given by the mother will be more delicious [01], because a spice
called love will also be added (to them) [02]. Do you like foods served by the mother more, or
foods served by the wife more [03]?’
‘What we call “mother” is something that cannot be replaced in any other way [04]. Whoever
serves foods well, serves (them) with love, they (i.e. foods) will have a good taste [05]. A
concept called “wife” can become like “mother”, can’t it [06]? (It) can become like “sister”
[07]. The one called “wife” is a person who has many reflections [08]. What those called
“mother” have is a position above all those (i.e. reflections) [09]. Foods served by the wife
have one good taste; those served by the mother have another good taste [10]. I can only say
56
like that [11]. (I) like, I like whoever serves and gives foods well [12]. In those (i.e. foods) there
isn’t any difference between the mother and the wife [13].’
‘Well, what is called “taste” is really a thing we certainly developed (over time), isn’t it [15]?
I, what I have interest in is (usually) only foods that I first used to eat at home, (whether) in my
taste buds or in my memory [16]. When we think quickly, we (usually) like cooked rice or
many things made of rice, don’t we [17]? For instance, for breakfast (we) might like eating the
idli or dosa, and all things like that [18]. Maybe when (we) go out, to have a change, or a
“maarram” if you will, (we) might eat (something else) for one or two days [19]. There is
nothing more than that [20]. What we like more is things which get along (well) with our
(digestive) system, which are called “aryi”, which have rice (in them), (and) which have
something like that [21]. And then I like eating food with many dishes, don’t I [22]? Whether
that is non-vegetarian or (that) is all vegetarian [23]. But in the (past) few days our interest
towards what is called vegetarian (food) has increased [24]. (We) are in an attempt to reduce
non-vegetarian (food) [25]. What I am interested in are things that have more rice-based
(stuff) [26].
57
Interviewer: malayaLikkyE ennum priyappeTTataaNE ooNapaaTTukaL [27]. atEppooLe,
laaleeTTanE ooNam ennE manassil varyumppooL, eerram aadyame oorymma varyunna
ooNapaaTTeetaaNE [28]?
‘To the Malayali, Onam songs are what is preferred every day [27]. Similarly, when you think
of Onam, which Onam song comes to your mind first [28]?’
‘(I) know many songs I sang in the cinema [29]. In the song “ooNapuuve puuve puuve” there
are many lyrics [30]. (It has) many of my things [31]. And then on various stages I have sung
that song [32]. And then don’t we have many songs [33]? Many (songs) like “kuTTanaaTn
kunjayile” have been called Onam songs [34]. Now as this Onam is coming closer, now there is
a big, a destruct- collection of Onam songs [35]. What needs to be said is actually
“samhaaryam” [36]. Don’t (people) make collections [37]? Many people start writing songs
about Onam [38]. After Onam, (people) will forget all those (songs) [39]. Then another festival
came [40]. People wrote songs about that [41]. They are doing things (like that) [42]. Then that
so-called festival becomes like a big business, (such as) Onam markets (and) things for the
front porch (i.e. Onam flowers) [43]. But among Onam songs what suddenly comes to my mind
is the song “ooNapuuve”, which was sung by Brother Das [44].
58
phonemic transcription helps to distinguish minimal pairs like mantan ‘foolish one’ and
mandan ‘slow one’; otherwise both would be realized as mandan (i.e. [mand̪ ̪ an]) under a
phonetic transcription (see Mohanan & Mohanan (1984: 599) for details).
Similarly, although both the alveolar plosive [t] and alveolar nasal [n] can be found at the
phonetic level, their transcriptions are rendered at the phonemic level. According to Mohanan
& Mohanan (1984: 581), [t] only occurs as a geminate or is preceded by the alveolar nasal.
Moreover, /r/ is realized as [tt] when undergoing gemination (e.g. aarE ‘river’ vs. aattil ‘in the
river’). Also important is the fact that there are no morpheme-internal [rr] sequences at the
phonetic level. Accordingly, Mohanan & Mohanan (ibid.) consider the underlying form of [tt]
to be /rr/ and thus that of its singleton counterpart [t] to be /r/. Following their analysis, I
transcribed words like [eːttaʋum] as eerravum ‘most’ and [ende] as enre ‘my’. Note that the
sequences of /r/ plus vowels would not be ambiguous. Wherever the sequences follow the
alveolar nasal, /r/ is realized as [d] (“first” as [t] and “then” [d] due to the voicing effect at the
intervocalic position); otherwise it is always realized as [r] elsewhere. As for the alveolar nasal,
Mohanan & Mohanan (ibid.: 582) again point out that the singleton dental and alveolar nasal
are in complementary distribution: [n̪ ] occurs morpheme-initially while [n] morpheme-finally
and intervocalically. Thus, I used the symbol n to transcribe both [n̪ ] and [n]. Most importantly,
subsuming [t] under /r/ and [n] under /n/ give rise to an elegant phonemic inventory which
would otherwise contain two “aberrant” alveolar stops (i.e. the plosive [t] and the nasal [n]).
I said the transcriptions are “by and large” phonemic because in some cases they are rather
phonetic when they could have been more phonemic (i.e. more abstract) given the general
phonological rules in Malayalam. To achieve a maximal descriptive economy, Mohanan &
Mohanan (1984: 596) propose to subsume [n̪ ], [ɲ], and [ŋ] all under /N/, whose place of
articulation is unspecified and determined by the following stop. For instance, the underlying
form of [aɲjə] ‘five’ could be postulated as /aNc/ (where /c/ determines /N/ is realized as [ɲ])
and that of [paŋgajam] ‘lotus’ as /paNkajam/ (where /k/ determines /N/ is realized as [ŋ]).
Elegant as this proposal is, I chose to specify in my transcriptions the place of articulation for
nasals (except for the contrast between dentals and alveolars for the reasons stated above). The
decision was motivated by two reasons. For one, I would like to strike an (admittedly
somewhat arbitrary) balance between descriptive economy and the level of abstractions. For
the other, keeping nasals like [ɲ] and [ŋ] at the phonemic level leads to a balanced phonemic
inventory in which where there is a plosive there is a corresponding nasal at the same place of
articulation.
Another aspect where my transcriptions are not strictly phonemic has to with palatals stops.
Mohanan & Mohanan (1984: 585) propose a palatalization rule where all surface palatal stops
(such as [kʲ], [kʲʰ], [gʲ], [gʲʰ], and [ŋʲ]) derive from velar ones followed by front vowels. So the
dative suffix as in [magaɭkkə] ‘to the daughter’ and [kuɖɖikʲkʲə] ‘to the child’ can both be
rendered as /-kkə/ at the phonemic level. However, the fact is much more complicated than the
59
palatalization rule reveals. As Mohanan & Mohanan admit, “[p]alatalization applies in
monomorphemic words, with numerous lexical exceptions. It applies across affixal junctures in
causatives, verbalizers, and datives, but not in plurals and across compound junctures” (ibid.:
588; italics mine). On top of this, palatalization is also subject to “a number of idiosyncratic
dialectal and idiolectal variations” (ibid.: note 23). Given these numerous exceptions and
idiosyncracies that block the application of the palatalization rule8, I found it more desirable to
specify the palatalization at the lexical level. Thus, in my transcriptions palatal stops are
rendered by their corresponding velar stops followed by the symbol y. For instance,
[kaɻikʲkʲuga] ‘to eat’ is transcribed as kaRikkyuka, where “kky” represents [kʲkʲ].
8
To account for the complicated cases where the application rule does not apply, Mohanan & Mohanan (1984)
have to posit four strata in which different phonological rules are applied under certain circumstances.
9
That is if “rice” and “rice-based”, “taste” and “taste buds”, “vegetarian” and “non-vegetarian” are all counted as
“different words”.
60
are marked for locative and dative case respectively. The marking of plurality is a bit tricky
since it is sometimes expressed by English -s (e.g. ryiflekshEns in [8]) and sometimes by
Malayalam -kaL (e.g. steejukaLil in [32]). More data are needed in order to find out why some
English nouns are borrowed as plural forms (i.e. those having English -s) while others are
borrowed as singular forms and then undergo native pluralization (i.e. those having Malayalam
-kaL).
10
Aside from a velar assimilated to the manner of articulation of its preceding nasal, a nasal is also assimilated to
the place of articulation of its following stop in this example.
11
In other words, the past tense form and converbal form can be identical. However, the /E/ ending is preferred
61
nouns is the /E/ ending whereas that for finite verbs is the /u/ ending, it is hard to decide
whether /E/ or /u/ is the “underlying” form (if there is one). Thus, in cases like this the citation
form is given on morphophonemic level, that is, /E/ for nouns and /u/ for verbs.
After presenting the text, I offer some justifications for the glossing and then a linguistic
analysis of the morphosyntax observed in the text.
62
63
64
65
66
67
68
69
3.2. Justifications for the glossing
As is the case for transcriptions, it can be a thorny issue as to what levels of abstraction
should be arrived at for the glossing. There are two extreme approaches. One is to completely
go by forms, namely, to gloss a morpheme the same way regardless of its meanings or
functions in the context where it occurs. This form-based approach is often adopted when it
comes to glossing case markers, which crosslinguistically tend to assume multiple functions
(see Haspelmath 2009 for discussions on the glossing of case markers). The other extreme is to
completely go by functions, namely, to gloss a morpheme based on the meanings or functions
it assumes in the context where it occurs. This function-based approach can be useful in
glossing morphemes which have too general functions that their meanings would seem opaque
if glossed the same way across the board. Obviously, both approaches have drawbacks of their
own. While the form-based approach can sometimes veil the meanings of a morpheme in
certain context (e.g. what does dative on temporal or abstract nouns mean?), the function-based
approach results into enumeration of seemingly unrelated glosses that could have gone by the
same name.
Considering the pros and cons of these two extreme approaches, I adopted a mixture of the
two, with probably more weight put on the form-based approach. The purpose here is to
provide glosses that are as fine-grained as possible. To this end, a phonological unit on the text
tier was decomposed, to the best of my knowledge, on the morpheme-by-morpheme tier to the
degree that further decomposition would make no sense. For instance, oryupakshe ‘maybe’ in
Onam.19 could have been decomposed into oryu ‘one’ and pakshe ‘but’, but it seems difficult
(at least to me) to derive the meaning “maybe” from “one” and “but”. So I left it as it is. On the
other hand, allengkil again from Onam.19 could have been glossed simply as “or” (which is the
meaning given by the consultant) instead of being broken into alla ‘NEG.COP1’ and -engkil
‘COND2’. However, I opted for the decomposition route because I found the literal
interpretation “if not” fits as equally well in Onam.19 as “or”, at least logically speaking.
70
In addition, some remarks are necessary regarding the glossing of the imperfective finite past
tense form (which always ends with /u/) and the converbal form (which ends with either /u/ or
/E/). Jayaseelan (2005) convincingly shows that the converbal form should be treated as
functionally distinct from the imperfective finite past tense form, in spite of their formal
identity, because the former can also co-occur with finite future and past tense in the serial verb
construction. Due to this, I adopted the function-based approach and glossed some forms as
CVB for converbs and others as PST for imperfective finite past tense, depending on the context.
Furthermore, there is evidence that some postpostions in Malayalam historically evolve
from converbs, thus sharing the same form with converbs. So to make this historical
development clear, I glossed all postposition-like forms as converbs wherever possible. For
instance, kuriccE in Onam.41, which could have been glossed simply as “about”, was glossed
as “concern.CVB” for it shows a structural identity with converbs in general. Note that the
distinction between postpostions and converbs is not discrete and clear-cut, but instead is a
matter of degree on a continuum. Some postpostions/converbs may have their finite
counterparts, but this is not always the case. While kaRinynyE ‘pass.CVB’ (or simply ‘after’)
and its finite present tense counterpart kaRiyunnu ‘pass.PRS’ are both acceptable forms,
kurikkyunnu, the potential finite present tense counterpart of kuriccE “concern.CVB” (or simply
“about”), is not acceptable.
Finally, the morpheme -iTTE is quite common in the text and its gloss “PFV” might not seem
straightforward from the context, so a justification is needed. Admittedly, there is little (if at all
any) perfective meaning in bisnisaayiTTE of Onam.43. In fact, Onam.43 would still be
acceptable and mean roughly the same if -iTTE had not been used, as is the case in
ooNaccanrakaLaayi and kaaryangngaLaayi from the same clause. However, the evidence for
-iTTE indicating a perfective meaning comes from minimal pairs like (1) and (2).
While the events of counting and crying happen simultaneously in (1), the event of counting
precedes that of crying in (2), with the only difference being the presence or absence of -iTTE.
Nevertheless, it seems that this particular morpheme has grammaticalized to some extent since
there are cases where its presence or absence does not make much difference in meanings, such
as (3).
71
(3) avan nannaayi(TTE) paaTum
avan nannaayi(-iTTE) paaT-um
3SG.NOM well(-PFV) sing-HAB
‘He sings well.’ (Elicited)
That is to say, the morpheme -iTTE synchronically may or may not mark perfectivity,
depending on the context where it occurs. But to specify its role as a special morpheme and
maintain the same gloss across the board, I glossed it as “PFV” irrespective of whether it in fact
carries perfective meanings or not.
12
This -a morpheme could have been glossed as PTCP for “participle” instead of ATTR for “attributive”. However,
this morpheme is also found in many adjectives (e.g. nalla ‘good’, putiya ‘new’, valiya ‘big’, paRaya ‘old’, etc.),
and it can also be suffixed to non-adjectival and non-verbal words like ennE ‘QUOT’, giving rise to enna, as in
Onam.44. Thus, I found the term “attributive” better describes its general attributive function than “participle”
does.
13
In cases where the finite past verb ends with /i/, it is -ya, an allomorph of -a, that is actually suffixed to it “on the
surface”.
72
can be turned into a participle and then function as an adjective (hence the term “adjectival
participles”). In Onam.03, for instance, two adjacent verbs modify the noun bhakshaNam
‘food’. In cases like this, only the last verb has to be in the participial form, and the rest of the
verb series (if there are more) remain in the converbal form, just as if they were in finite clauses.
Moreover, not only internally there can be multiple verbs in a participial phrase, but externally
there can be multiple participial phrases that occur in a row to modify the same nominal. An
amazing example comes from Onam.21, where four participial phrases (headed by
ceerynniryikkyunna, parayunna, aayiTTuLLa, and uLLa) are used to modify the noun
kaaryangngaL ‘things’.
Finally, a last feature quite characteristic of Malayalam morphosyntax is the so-called
“dative construction”. Jayaseelan (2004) classifies instances of this construction in Malayalam
into four “rough semantic rubrics”, namely, possession, experiencer, certain modals, and
“know”-class verbs. All these types can be found in the text, except for the last one. In the
possession type, the possessor is marked by dative and the possessum by nominative, as in
Onam.33, where namukkE ‘1PL.INC.DAT’ is the possessor. The “possessor” in this type can
be rather general and does not even have to be animate, as bhakshaNattinE ‘food.DAT’ in
Onam.01 shows. In the experiencer type, the experiencer of some mental or physical state is
marked by dative while the state is coded as a noun in nominative. As Onam.22 illustrates,
enikkyE ‘1SG.DAT’ is the experiencer that undergoes the state of liking something, which is
reified by the noun iSTam ‘liking’. In the “certain modal” type, the verb has some kind of
modal meanings and the “notional subject” is marked by dative. For instance, in Onam.11 the
“notional subject” is enikkyE ‘1SG.DAT’ and the modal verb is parru ‘can.CVB’.
Among the many modals in Malayalam, -aam is of particular interest. According to
Jayaseelan (2004: 233), “the modal -aam ‘may’ occurs in the dative construction when it has
the meaning of ‘permission’, and in the nominative construction when it has the meaning of
‘possibility’.” The examples he used to illustrate this point are given in (4).
However, there is an example in the text that does not seem to accord with Jayaseelan’s
generalization above. In Onam.06, where the verb suffixed with the modal is maaraam
‘become.POT’, koNseptE ‘concept.DAT’ is marked in dative, but the “possibility” interpretation
73
appears more plausible than the “permission” interpretation, if the latter is even applicable in
this context14. This suggests that the semantics of -aam is actually quite complex. As a matter
of fact, the very same morpheme can also have a cohortative and imperative interpretation as
illustrated in (5) and (6) respectively15. Apparently, more data and work is needed in order to
find out how the (possible/permissive) modal, cohortative, and imperative meanings are all
instantiated by the morpheme -aam and what other “potential” (no pun intended) functions it
may undertake.
14
I found the term “modal” too general for -aam, for it conveys only some special kinds of modal meanings. Thus,
to incorporate the “possibility” and “permission” interpretation, I glossed -aam as POT for potential.
15
Morphemes are assumed to be “the same” if they demonstrate the same distribution patterns.
74
References
Akamatsu, Tsutomu. 1997. Japanese phonetics: Theory and practice. München: Lincom
Europa.
Asher, R. E. and T. C. Kumari. 1997. Malayalam. London: Routledge.
Comrie, Bernard and Norval Smith. 1977. Linguistic descriptive studies: questionnaire. Lingua
42.1-72.
Geethakumary, V. 2002. A Contrastive Analysis of Hindi and Malayalam. Ph.D. Dissertation.
Gundert, Rev. H. 1970. A Malayalam and English Dictionary. Reprint of the edition
1871-1872. Osnabrück: Biblio Verlag.
Haspelmath, Martin. 2009. Terminology of case. In The Oxford Handbook of Case, eds. by
Andrej Malchukov and Andrew Spencer. Oxford: Oxford University Press.
Jayaseelan, K. A. 2005. The serial verb construction in Malayalam. In Clause Structure in
South Asian Languages, eds. by Veneeta Dayal and Anoop Mahajan, pp. 67–91.
Dordrecht: Kluwer.
Ladefoged, Peter and Ian Maddieson. 1996. The Sounds of the World's Languages. Oxford:
Blackwell.
Ladefoged, Peter. 2005. Vowels and Consonants (Second ed.). Blackwell.
Lewis, M. Paul (ed.), 2009. Ethnologue: Languages of the World, Sixteenth edition. Dallas,
Tex.: SIL International. Online version: http://www.ethnologue.com.
M.T., Hany Babu. 2006. Genericity, quantification, and modality: the many faces of -um and
-unnu in Malayalam. CIEFL Occasional Papers in Linguistics 12.
Maddieson, Ian. "Voicing in Plosives and Fricatives", in Martin Haspelmath et al. (eds.) The
World Atlas of Language Structures, pp. 26–29. Oxford: Oxford University Press,
2005.
Moag, Rachel and Rodney Moag. 1967. A Course in Colloquial Malayalam. Milwaukee,
Wisconsin: University of Wisconsin.
Mohanan, K. P. and Tara Mohanan. 1984. Lexical phonology of the consonant system in
Malayalam. Linguistic Inquiry 15(4): 575-602.
Mohanan, Tara and K. P. Mohanan. 1999. Two forms of “BE” in Malayalam. In Proceedings
of the LFG99 Conference, eds. by Miriam Butt and Tracy Holloway King.
75