Clause participant tagging question

Page 1 of 1 (20 items)
This post has 19 Replies | 1 Follower

Posts 21915
Forum MVP
Graham Criddle | Forum Activity | Posted: Sat, Jun 27 2015 3:22 AM

The BWS for φαίνω has entries in the Subject subsection of "Clause Participant" section for "Angel of the Lord" and "God" but they both show the same results.

The word "Lord" in Matthew 2:13 is tagged with both "Angel of the Lord" and "God"

Should the tagging be more granular here?

Thanks, Graham 

Posts 791
LogosEmployee
Eli Evans (Faithlife) | Forum Activity | Replied: Mon, Jun 29 2015 3:52 PM

Hi, Graham. The entire phrase "Angel of the Lord" is tagged with <Person Angel of the Lord> and only the smaller sub-phrase "of the Lord" is tagged with <Person God>. (Including "of" and "the" because of the Greek genitive kuriou.) So what you're seeing is the overlap of the two tags. 

Posts 21915
Forum MVP
Graham Criddle | Forum Activity | Replied: Mon, Jun 29 2015 11:54 PM

Hi Eli

Eli Evans:
The entire phrase "Angel of the Lord" is tagged with <Person Angel of the Lord> and only the smaller sub-phrase "of the Lord" is tagged with <Person God>. (Including "of" and "the" because of the Greek genitive kuriou.) So what you're seeing is the overlap of the two tags. 

Thanks for the explanation - its helpful

But I think it also demonstrates the issue.

The BWS indicates that God is the subject of the verb "appeared" in Matt 1:20 but it is actually the angel who appeared - the κυριος in genitive is giving us more insight into whose the angel is. 

The clause search generated from that section is below - showing that God is the subject but He is not

I think the tagging is right but should this section of the BWS exclude genitives and just allow nominatives? And I am aware that I don't know the full implications of this suggestionSmile

Graham

Posts 2467
Lee | Forum Activity | Replied: Tue, Jun 30 2015 2:17 AM

I fully agree with Graham.

Posts 26492
Forum MVP
MJ. Smith | Forum Activity | Replied: Tue, Jun 30 2015 11:09 AM

If you look at the syntactic tree, the classification makes sense as it is part of the noun phrase that serves as the subject. This is in contrast to the NP that is usually within a VP and serves as the object. Confusing yes; understandable yes.

Orthodox Bishop Hilarion Alfeyev: "To be a theologian means to have experience of a personal encounter with God through prayer and worship."

Posts 2467
Lee | Forum Activity | Replied: Tue, Jun 30 2015 1:52 PM

In the sentence "an angel of the Lord appeared to him" which is the subject?

On all tests the answer would be "(an) angel of the Lord".

So either the tagging is faulty (which it isn't), or the algorithm needs to be clarified or refined.

Posts 21915
Forum MVP
Graham Criddle | Forum Activity | Replied: Tue, Jun 30 2015 11:21 PM

MJ. Smith:
If you look at the syntactic tree, the classification makes sense as it is part of the noun phrase that serves as the subject. This is in contrast to the NP that is usually within a VP and serves as the object. Confusing yes; understandable yes

Hi MJ - that is a compelling argument (and I appreciate you looking further into this at https://community.logos.com/forums/p/106932/739633.aspx#739633 - I assume that is related to this)

I agree with your analysis of confusing and understandable but wonder about "useful". As it stands, the BWS section and the associated search seem to be showing the words included within the Subject part of the Clause - but, as Lee suggests, it isn't showing us what most people would regard as the Subject.

I understand the problem I am raising in that "useful" is rather subjective whereas a strict algorithm based on clausal structure is deterministic and may well be all that we can reasonably expect.

Posts 26492
Forum MVP
MJ. Smith | Forum Activity | Replied: Tue, Jun 30 2015 11:51 PM

Faithlife needs to make clear the target audience and the linguistic underpinnings of some of their coding if they want to maximize usefulness. I understand why they take a its like N modified to meet our needs but without knowing what forced the modifications and how those modifications are defined ... but then I am into precision .... sometimes to a fault.

Orthodox Bishop Hilarion Alfeyev: "To be a theologian means to have experience of a personal encounter with God through prayer and worship."

Posts 21915
Forum MVP
Graham Criddle | Forum Activity | Replied: Wed, Jul 1 2015 12:12 AM

MJ. Smith:
Faithlife needs to make clear the target audience and the linguistic underpinnings of some of their coding if they want to maximize usefulness

Agreed

MJ. Smith:
I understand why they take a its like N modified to meet our needs but without knowing what forced the modifications and how those modifications are defined

Sorry I don't understand this - please clarify

Posts 10031
Forum MVP
NB.Mick | Forum Activity | Replied: Wed, Jul 1 2015 12:16 AM

Lee:

In the sentence "an angel of the Lord appeared to him" which is the subject?

On all tests the answer would be "(an) angel of the Lord".

So either the tagging is faulty (which it isn't), or the algorithm needs to be clarified or refined.

My understanding is that the tagging exactly reflects your standard answer (you yourself say so: it isn't faulty).

The problem is that the tagging and search is on the word level.

So the question is: which grammatical function does "Lord" have in the sentence?

Answer: part of the subject 

The user seeing it as hit for "Subject:Lord" may be confused. Maybe a future refinement could differentiate - however, since the Greek word for Lord is not in the nominative case, users could see the difference today and create respective searches.

Running Logos 8 latest beta version on Win 10

Posts 21915
Forum MVP
Graham Criddle | Forum Activity | Replied: Wed, Jul 1 2015 12:57 AM

NB.Mick:
The problem is that the tagging and search is on the word level.

Or that the tagging is at the phrase level

NB.Mick:

So the question is: which grammatical function does "Lord" have in the sentence?

Answer: part of the subject 

Yes - in that it is part of the noun phrase which serves as subject (as MJ points out) but as the screenshot in my original post shows the BWS doesn't refer to  "part of the subject" it refers to "subject"

NB.Mick:
however, since the Greek word for Lord is not in the nominative case, users could see the difference today and create respective searches.

I don't think you can do that in a clause search - which is what get triggered from the BWS.

But you are correct in that you can do it in other ways - and that is really helpful

But you need to go through the process of recognising the issue, working out what is going on and coming up with some refined searches. And maybe that is how it needs to be and this discussion has certainly been useful in clarifying some of this stuff for me.

AppreciatedBig Smile

Posts 2467
Lee | Forum Activity | Replied: Wed, Jul 1 2015 3:21 AM

NB.Mick:

The problem is that the tagging and search is on the word level.

1. The tagging is at the word or phrase level, depending on the subject matter.

2. Why should the search be on the word level, if we're talking about a "clause participant" search with such refinements as grammatical roles and semantic roles? To me, either we admit it's broken, or the documentation needs to reflect how the search really functions.

Posts 21915
Forum MVP
Graham Criddle | Forum Activity | Replied: Wed, Jul 1 2015 4:01 AM

Lee:
2. Why should the search be on the word level, if we're talking about a "clause participant" search with such refinements as grammatical roles and semantic roles?

Interesting that you mention Semantic Roles. Looking at the Agent for this verb it has both Angel of the Lord and God - which raises similar questions.

And this may be related to the issue discussed at https://community.logos.com/forums/t/104154.aspx - but it is not identical.

Posts 10031
Forum MVP
NB.Mick | Forum Activity | Replied: Wed, Jul 1 2015 5:06 AM

I personally am not sure about the sematic roles, but it looks wrong.

Additionally I checked "Son of Man" which to me seems a grammatically similar construction to "Angel of Lord", but it is treated differently: in the BWSs I ran, Man is not shown as Agent or as Subject when Son of Man is.

Running Logos 8 latest beta version on Win 10

Posts 21915
Forum MVP
Graham Criddle | Forum Activity | Replied: Wed, Jul 1 2015 5:41 AM

NB.Mick:
in the BWSs I ran, Man is not shown as Agent or as Subject when Son of Man is.

I think that's due to tagging - with the phrase being tagged "Son of Man" - i.e. a title

Posts 2467
Lee | Forum Activity | Replied: Wed, Jul 1 2015 2:04 PM

Here we'd be looking at the constituent "the sign of the Son of Man" τὸ σημεῖον τοῦ υἱοῦ τοῦ ἀνθρώπου.

Hence, the search function correctly omits the verse as a hit for "man".

Posts 791
LogosEmployee
Eli Evans (Faithlife) | Forum Activity | Replied: Tue, Jul 7 2015 9:36 AM

I will forward this discussion to the editorial team. Maybe someone who knows more than I do can weigh in.

Posts 131
LogosEmployee
Peter Venable | Forum Activity | Replied: Tue, Jul 7 2015 10:04 AM

While the tagging is correct, Clause Search has a bug that causes queries such as "subject:God" to include results for which God is inside the subject. It should only include these results when the query says "related:God" or "person:God"

I reported the bug so it will be fixed in a future release.

Posts 21915
Forum MVP
Graham Criddle | Forum Activity | Replied: Tue, Jul 7 2015 11:00 AM

Thanks Peter

That sounds really good

Graham

Posts 2467
Lee | Forum Activity | Replied: Tue, Jul 7 2015 5:18 PM

Peter Venable:

It should only include these results when the query says "related:God" or "person:God"

when the query says "related:God": I think I agree.

"person:God": I don't quite understand. How would that work with a phrase like like "Angel of the Lord" or "the sign of the Son of Man"?

(The complexity of the search function arises from analysis of semantic categories rather than purely mechanical grammatical categories.)

Page 1 of 1 (20 items) | RSS