7.3 Intr
ansitive, t
ransitive and causative
stems
279
Ko
.
n
.
da ra-
ʔ
a 2sg, ra-du 2pl. A verb in Dravidian is inflected for tense/aspect/mood and
carries a verbal base as its nucleus. A verbal base in Dravidian may be simple, complex
or compound. A simple base is identical to the monosyllabic verb root (C
1
)
˘
¯
V(C
2
),ora
disyllabic one extended with a short vowel (C
1
)V
1
C
2
-V
2
in which -V
2
does not contribute
to the root meaning, e.g.
∗
¯a ‘to be, become’ [333],
∗
key ‘to make’ [1957],
∗
cal ‘to go’
[2781],
∗
man ‘to be’ [4778],
∗
w¯a
.
z ‘to flourish’ [5372],
∗
par-a ‘to spread’ [3949],
∗
i
.
z-i
‘to descend’ [502]; a complex base has a root and a formative suffix, encoding voice,
transitivity or causation, e.g.
∗
a
.
t-a-nku ‘to be subdued, hidden’:
∗
a
.
t-a-nkk ‘to control,
hide’ [63],
∗
key-pi- ‘to cause one to do’; a compound base has more than one root with
the final constituent as a verb, e.g.
∗
akam ‘inside’: Ta. aka-ppa
.
tu ‘to be included’, Te.
aga-pa
.
du ‘to be seen, to f
all in the visual
field’ [7].
Morphologically a verb may be finite or non-finite. A finite verb has the structure
stem + tense-mode + (g)np (gender–number–person) marker, which normally agrees
with the head of the subject noun phrase (NP), Ta. n¯a
ncey-t-¯en ‘I did’, Ko
.
n
.
da v¯anru ki-
t-an ‘he did’. Historically some descendant languages have lost the agreement features,
either partially or fully, like Modern Malay¯a
.
lam, or neutralized all gender–number con-
trasts in the third person,
like Toda and Brahui. A
non-
finite verb has two components,
the verb base + tense/aspect, e.g. Ta
.
cey-tu ‘having done’, Ko
.
n
.
da ki-zi id., perfective
participle or gerund in both the languages; syntactically, it heads a subordinate clause. In
unmarked word order the verb, finite or non-finite, occupies the end position of the clause.
7.3 Intransitive, transitive and causative stems
A simple verb may be inherently intransitive (
∗
¯a- ‘to be’) or transitive (
∗
ciy- ‘to give’)
depending on its meaning and its relationship with the complement phrases in a given
clause. I suggested in section 5.3 (earlier TVB: §2.38, pp. 145–6) that ‘sonorant suffixes
of the R type (l,
.
l, r ,
.
z, w, y) were added to (C)
˘
¯
V- or (C)VC-V-stems to form ex-
tended intransitive/middle-voice stems’. Synchronically, a transitive verb is changed to
intransitive/middle voice in Ku
.
rux and Malto by adding -r, e.g. Ku
.
r. kam- ‘to make’:
kam-r- ‘to be made’, Malt. ey- ‘to bind’: ey-r- ‘to bind oneself’. This seems to be a relic
of a Proto-Dravidian usage, since it is not found in any of the neighbouring Indo-Aryan
and Munda languages. Note that most verbs ending in formative -(V) l/-(V) r in South
Dravidian I and South Dravidian II tend to be intransitive.
Three complementary modes of
forming transitive
–causative stems were quite
an-
cient: (1) by the addition of
∗
-tt to monosyllabic roots that end in an apical stop or nasal
/t
n
.
t
.
n/, and of
∗
-pp to roots ending in final
∗
-i or
∗
-y (the Proto-Dravidian conditions are
not all recoverable). In Central Dravidian both these suffixes got generalized as causative
markers; (2) by the addition of a causative morph -pi- ∼ -wi- ∼ -ppi- to a transitive verb
stem, simple or complex; (3) a complementary type to these is represented by roots of
(C
1
)
¯
VC
2
type, where C
2
is a liquid sonorant, or a disyllabic root of the type (C) VCV- or
(C) VCV-y. A subset of these stems formed transitives by geminating the final stop of