332 Chapter 10 Modifications to argument structure
verbs occur in the object voice, thus indicating that an important if not principal
function of -e and -agi is, in fact, in service of identifying the role of the subject.
Viewed this way, -e and -agi appear to be part of the same morphological sys-
tem as actor voice and object voice. And the function of them all combined
parallels the role of voice morphology in western Austronesian languages, iden-
tifying the subject (or topic or focus) of a clause.
It is also the case that the distribution of Madurese voice (+ applicative)
morphology can be quite similar to that of Philippine languages. A comparison
of the relative occurrence of actor voice, object voice, and object voice + appli-
cative morphology reported in Davies 2005 demonstrated that the pattern of
distribution was quite comparable to that reported for Cebuano in Bell 1988.
The comparison is given in (171). (The terminology for Cebuano voice catego-
ries is that used in Bell 1988. It differs somewhat from Schachter’s 1976 termi-
nology for Tagalog, which is the terminology most widely used in the literature
on Philippine languages.)
(171) Cebuano voice distribution
(Bell 1988)Madurese voice distribution
(Davies 2005)Agent-topic 50.7% Actor voice 48.0%Object-topic 31.6% Object voice 38.3%Reference-topic 12.4% -e 9.7%Instrument-topic 5.0% -agi 4.0%Total goal-topic 49.3% Total object voice 52.0%In (171), the object voice number (38.3%) reported for Madurese excludes ob-
ject voice used with -e and -agi, and the -e and -agi categories represent those
instances of -e and -agi combined with object voice. The ‘total object voice’
category combines object voice, -e and -agi, which is comparable to Bell’s cov-
er term ‘goal-topic’, the category comprised of all the non-agent-topic instances
of voice morphology. As is clear from this table, as well as the figures for the
Madurese texts given in (170), the combined effect of object voice and the ap-
plicative affixes is quite similar to Bell’s tallies for Cebuano. Clearly, the statis-
tics and their comparison is neither rigorous nor definitive. However, the re-
markably similarity across the two languages is suggestive. The applicative
morphology together with the voice morphology give a result quite like that of
the richer voice systems found in some other Western Austronesian languages.
Despite this suggestion, it is clear that treating the -agi suffix as a part of
the voice system of Madurese is no closer to providing a unified analysis than