[scribed by Dan]

For relevant Indonesian data, see web link:

• http://moin.delph-in.net/CapitolHillPassives

Analysis of many of these phenomena are not yet implemented, but planned.

Passives in Indonesian are more frequent than in English. Frequent use in imperatives, for politeness, and note also that in relative clauses, can't relativize non-subject, so always get passive when the lower clause argument is other than (deep) subject.

Two main types of passives, plus some less frequent variants.

Francis: In web notes, ex 3a -- you can't add 'self' have self cleaned, ...

Do we get passive type 2 for ditrans verbs? yes:

• Budi saya belikan buku
• Budi 1SG PASS-give book

Propose to first combine |saya belikan| in unusual complement-head order, and then have the result now able to participate in ordinary head-complement rule, even though |belikan| cannot. Suggest marking 1SG and 2SG pronouns as pre-head, and everything else as post-head, to avoid over-generating with head-compl vs compl-head rules.

Need complement-head rule for passive type 2, but maybe nowhere else? Perhaps constrain this rule to lexical (passive) head only.

When coordinate verbs in type 2 passive, have to have the |ku-| (1SG) on both conjuncts, so conclude that it is an affix, not a clitic.

• Makanan kumasak dan kumakan
• Food was cooked by me and eaten by me

but not

• [ungrammatical] Makanan kumasak dan makan

For ter-passive, similar to type 1 passive, but can occur with `accidently' where can have by-phrase.

So propose

• one ter- for V-to-Adj, and
• one ter- for passive V with optional overt agent, but add intersective modif. for `accidentally'

There is also an interpretation for 'potentially' (which is probably scopal), but maybe leave underspecified for now, even though might be tempted to treat `potentially' reading by adding scopal modifier.

Might also try to derive passives from nouns or adjectives with |ke-...-an|, but perhaps not treat as passive.

• not fully productive; only one word possible
• but somewhat productive: Francis -- |ke-oreore-an| where |oreore| is a recent lexical addition to Indonesian.

Information structure effects?

• not clear whether passive in Indonesian should add an ICONS (analogous to English), so perhaps leave alone for now, especially since use of passive for relative clauses is syntactically motivated, perhaps with no effect on information structure.

Post-coord nya: clitic, not affix, even though written without space

• Budi menasak dan memakannya
• Budi ACT-cook and ACT-eat=2SG

Imperative with passive:

• imperative force
• unusual word order
• demoted subject can't be realized
• demoted subject is 2nd person

Emily suggests an analysis with separate di- passive lexical rule to directly make an imperative passive with the above properties, and where the deep object is still a complement, not on the SUBJ list.

Then to predict the absence of reflexives in these passive imperatives (in contrast to active imperatives), perhaps say that there is a constraint on argument structure in Indonesian that the first item on ARG-ST cannot be reflexive.

Maybe a temptation to also have a syntactic rule for head-subject just for these passive imperatives, but this seems unecessary since already need a lexical rule for these, and it can treat the deep object as still an object.