W: I don't know Greek, but others in the room do...

Argument order is syntactically free w/in a clause

Coordination

• XP kai XP kai XP
• Also "both and" versions:
  • te XP kai XP
  • XP te kai XP
• discontinuous: XP (something) kai XP

Corpora

• W went through 2 corpora and found instances of discontinuous coordination
• New testament and Herodotus

E.g.

• "Grace and peace to you": grace.F.NOM.SG you.DAT.PL and peace.F.NOM.SG

Berthold asks about determiners

• W: determiners are often repeated
• Guy: how do you represent the agreement with the discontinuous determiners?
• W: ...
• Olga: is the "grace to you" example something like gapping (e.g. .... and peace (to you)
• W: many examples could be compatible with that theory, but not this example
  • Hebrews 9:4
  • 1st Timothy 1:15
  • Acts 13:16
    • Dag: I've written an article about this one; it's an interesting example; paul is the subject of the participle and the matrix verb, but structurally paul is realized in the participle clause; it's a weird control structure
  • Acts 5:31
    • w: coordinated nouns, but interrupter is the conjuncts and interrupter argument of something else
  • Revelation 18:24
  • Matthew 11:19

EMB: what about extraposition; how far can you get by saying the first bit causes extraposition

W: ... also see (golden) (have) (altar) -> golden altar

Guy: perhaps there are two phenomena happening together

EMB: it's not always going to the right of the clause...

Dan: when it's left it could be topicalization

W: that's likely, but doesn't help me much

Dag: the part starting with kai is sentence final right?

W: yes, especially when it's short

Dag: your translation is quite accurate, e.g. the complementizer is translated as a colon, which is quite accurate

W: In all these examples you could say things go all the way to the right of the clause

Guy: couldn't the second conjunct be extraposed?

Dag: yeah you could argue that

Dan: if extraposition, do you have a syntactic of morphsyntactic thing, or is it anaphora; that would be the lazier and easier option

John: do conjuncts have to be the same type?

W: they have to agree with each other generally

Dag: you could coordinate them continuously too

Dan: then we can't just wave our hands about constituency, if there's agreement requirements on coordinands then we need to do something syntactic

Guy: if it's anaphora you could also predict it to go even further away

W: yeah, and they do stay pretty close, but if they cross sentence boundaries then my search would not have found them

Dan: but imagine in that context you say "Paul said he liked to eat fish. Andrew added, 'and bread'"

Guy: but that's quoted speech

Dan: so what? I still want a semantic representation of it

W: you could say "John likes somebody. Andrew said 'yes, himself.'"

Berthold: can you have two dependents overlapping coordination, like interleaved A1 B1 A2 B2?

W: I tried thinking of that the other day, I think it could happen

Dan: ...

Dag: you need a closest conjunct theory; the alternative would be a gapping analysis

Dan: one of the predictions of extraposition is that you'd expect plural marking on the follow (Matthew 9:19 example)

EMB: we'll need a rule for the bottom of the extraposition dep., and say agreement comes from the daughter launching the dependency. Laurie Dermer has an analysis of agreement in coordination and in her reading about closest conjunct agreement she found it's actually more generally distinguished conjunct agreement.

BC: this is dealing with traces then?

Dan: I wasn't thinking that; if you'd say "Jesus and the disciples followed him", it'd be plural

Dag: if it was "the disciples and Jesus", you'd expect singular

Dan: we could go with the prettier analysis or the one that gets us what we want

Dan: in English we could add afterthoughts, "John likes tartar sauce on his fish.. And Mary"

EMB: those are at the end of the utterance, where these are embedded in the sentence

BC: in German we have similar

Dan: and with quirky case you'd get this?

BC: i think so

Dan: a scary thought w/ extraposition is you have to launch candidate sites from every noun phrases

BC: do you want to have multiple dependencies or not?

Dan: how do you do those?

W: you have to have multiple sites at all times?

BC: yes but they mostly go unused; maybe have a recency model...

W: predict you can only do these on the closest, or outermost, etc.

BC: you can have several rules for, e.g., nearest attachment, second nearest, etc... unlikely to have, e.g., 5 removed

BC: look at Tibor Kiss's (NLT 2002?) analysis, for German, ... you want to key them to case, etc.

Dan: coordinated adpositional phrases?

W: I think so, but I can't say for sure

Dan: if you could argue it's semantically empty, you could say that you need the syntax

EMB: I wonder if about Petter's approach is relevant

Petter: ...

W: you need not just the index but the place it was talked about...

P: I don't do coordination well, this is a problem. I grab the first argument, but i don't know how many more there are

Dan: ...

EMB: but this isn't just arguments but also predicates

Dan: the tall and angry cat is not so different from the tall angry cat or the tall cat which is angry

Guy: verbs?

W: yeah, but it's hard to search for these

Dan: is this like "paul bought and sold a goat"? Why is that hard to find?

W: I can't just look at the individual words, I have to look at the whole phrase, e.g. the arguments of all the verbs, etc...

Dag: and you can pro-drop objects and subjects, which complicates things

W: in this case (1st Timothy 1:15), it was found by ...

Dan: so maybe there's not a uniform solution for all cases...

Guy: maybe you just need to extrapose just noun phrases?

W: and live with pro-dropped anaphora for the other ones?

Dan: do you have examples in a control structure (Paul persuaded the gentiles to leave and hebrews)

W: i did not notice any

Dag: it wouldn't be too surprising

Guy: attributive adjectives?

W: yes, but Dag says this (Acts 13:16) isn't an example

Dag: well it's not attributive, it's like a free relative

Dan: or depictive

EMB: so the grace and peace example, it's not the modifier that's distributed?

W: correct

Dan: for the extraposition story, if you're parsing bottom-up, you expose the morphosyntactic properties in some growing list, then when you find a conjunct, you choose one on the list that agrees, so you're not exploding the number of edges

W: but you need a reorder/shuffle operator?

Dan: just have a popping rule that removes things on the list until it finds one

W: but that increases number of edges

Guy: is it expensive to put it on the slash list?

Dan: it won't pack, because one has slash, one doesn't

BC: how often can you do this? Can you have tripartite coordination with things spread all over?

W: probably

BC: that will kill slash, because things need to stay on slash while you're popping them off...

Dan: John ate fish and Mary and bread

Dag: I don't think that will happen

W: actually that sounds like one i found that crossed in some way

W: the summary: some kind of extraposition keeping track of some kind of anchors

Dan: rich enough to keep some case information

BC: like an accusative list, etc..

Dan: only if you're just doing noun phrases

W: yes, so some extraposition for noun phrases, throw in the towel with anaphora for other cases

EMB: but you're gonna get that analysis (the pro-drop one) anyway for these analyses, once you do pro-drop in your grammar.

W: Revelations 18:24, we get this too?

Dan: we get some ambiguity here, do you launch the anchor of prophets, or saints, or all?

EMB: Maybe we can just say that the coordination rule suppresses all anchors beneath it.

Dan: That sounds dangerous.

BC: question for Dan, how to handle "faithful is the word and worthy of acceptance"

Dan: i think that's a depictive marked with an and, "Worthy of acceptance, faithful is the word"

Guy: that sounds like an awkward translations trying to stick close to the greek

...