Recursive syntax trees for Hebrew

[jonathanrobie]

We need syntax trees for Hebrew that include nested clauses.

Two possible options (there may be others):

1. Negotiate treebanks from an existing source
2. Parse Eep Taalstra data using Open Hebrew morphology

What are the advantages of each approach? What ramifications would each have for cost? quality? other considerations?

[cmerwich]

For the trees, you may want to take a look here:

https://github.com/ETCBC/lingo/tree/master/trees https://github.com/ETCBC/lingo/tree/master/trees

2018-03-30 13:27 GMT+02:00 Jonathan Robie notifications@github.com:

We need syntax trees for Hebrew that include nested clauses.

Two possible options (there may be others):

1. Negotiate treebanks from an existing source
2. Parse Eep Taalstra data using Open Hebrew morphology

What are the advantages of each approach? What ramifications would each have for cost? quality? other considerations?

— You are receiving this because you are subscribed to this thread. Reply to this email directly, view it on GitHub https://github.com/biblicalhumanities/copenhagen-alliance/issues/10, or mute the thread https://github.com/notifications/unsubscribe-auth/AHD4qct3jhvbT4gOAr9e_v3m4FSXex7bks5tjha2gaJpZM4TBmDD .

[rdeblois1960]

Thank you, Christiaan. My problem with the trees that I get through Shebanq is that it is impossible to have a CLAUSE embedded in a PHRASE. Are you saying that the TextFabric trees are any different?

[johanovic]

Out of interest, could you give an example of where you have a clause embedded in a phrase, @rdeblois1960 ?

[rdeblois1960]

Sure, take Lv 14:19: וְעָשָׂה הַכֹּהֵן אֶת־הַחַטָּאת וְכִפֶּר עַל־הַמִּטַּהֵר מִטֻּמְאָתוֹ וְאַחַר יִשְׁחַט אֶת־הָעֹלָה׃ The second clause has a prepositional phrase that contains another clause "the one who is to be cleansed from his uncleanness". ETCBC treats cases like this in a very complicated and linguistically awkward way.

[cmerwich]

@Reinier de Blois: I've enclosed a screenshot (and copied them into this message) of the parse trees for the three sentences within Lev 14:19 .

Leviticus 14:19 node=1181037 bSlot=60664 0 1 2 3 4 5 6

(S(C(CP(cj "וְ"))(VP(vb "עָשָׂ֤ה"))(NP(dt "הַ")(n "כֹּהֵן֙"))(PP(pp "אֶת")(dt "הַ֣")(n "חַטָּ֔את"))))

Leviticus 14:19 node=1181038 bSlot=60671 0 1 2 3 4 5 6

(S(C(CP(cj "וְ"))(VP(vb "כִפֶּ֕ר"))(Cadju(CP(pp "עַל")(dt "הַ"))(VP(vb "מִּטַּהֵ֖ר"))(PP(pp "מִ")(n "טֻּמְאָתֹ֑ו")))))

Leviticus 14:19 node=1181039 bSlot=60678 0 1 2 3 4 5

(S(C(CP(cj "וְ"))(AdvP(n "אַחַ֖ר"))(VP(vb "יִשְׁחַ֥ט"))(PP(pp "אֶת")(dt "הָ")(n "עֹלָֽה"))))

I haven't looked at them yet, it's Easter, a diner is waiting for me. Maybe they are an answer to your question.

2018-03-31 16:27 GMT+02:00 Reinier de Blois notifications@github.com:

Sure, take Lv 14:19: וְעָשָׂה הַכֹּהֵן אֶת־הַחַטָּאת וְכִפֶּר עַל־הַמִּטַּהֵר מִטֻּמְאָתוֹ וְאַחַר יִשְׁחַט אֶת־הָעֹלָה׃ The second clause has a prepositional phrase that contains another clause "the one who is to be cleansed from his uncleanness". ETCBC treats cases like this in a very complicated and linguistically awkward way.

— You are receiving this because you commented. Reply to this email directly, view it on GitHub https://github.com/biblicalhumanities/copenhagen-alliance/issues/10#issuecomment-377696746, or mute the thread https://github.com/notifications/unsubscribe-auth/AHD4qdqpAVbBAUe46rE5nRCtPHdWShjeks5tj5I5gaJpZM4TBmDD .

[rdeblois1960]

Thanks and enjoy your dinner! Unfortunately the problem is still there:

Leviticus 14:19 node=1181038 bSlot=60671 0 1 2 3 4 5 6

(S(C(CP(cj "וְ"))(VP(vb "כִפֶּ֕ר"))(Cadju(CP(pp "עַל")(dt "הַ"))(VP(vb "מִּטַּהֵ֖ר"))(PP(pp "מִ")(n "טֻּמְאָתֹ֑ו"))))) This sentence is analyzed as a sequence of a regular Clause and this strange animal that is called an Adjunct Clause. In a recursive system this second "clause" would be treated as a prepositional phrase of the first clause, with an embedded clause. Verse 20 contains a clause with the same verb, which is linguistically very similar to the one in verse 19, but analyzed differently.