Clause search help

Page 1 of 1 (20 items)
This post has 19 Replies | 2 Followers

Posts 3680
Francis | Forum Activity | Posted: Fri, Feb 13 2015 6:10 AM

Why do I get this:

In spite of this:

And that:

John 1:12, for instance, is found in both lists.

What am I doing wrong?

Posts 21928
Forum MVP
Graham Criddle | Forum Activity | Replied: Fri, Feb 13 2015 6:18 AM

Francis:
What am I doing wrong?

For some reason it looks as though you need to use parentheses

Posts 3680
Francis | Forum Activity | Replied: Fri, Feb 13 2015 6:26 AM

Graham Criddle:
For some reason it looks as though you need to use parentheses

Good find!

Apparently, it makes no difference which terms the parentheses are applied to:

I wonder if the two occurrences of AND introduce some form of ambiguity which may not be apparent here but could be in other contexts?

Posts 3680
Francis | Forum Activity | Replied: Fri, Feb 13 2015 6:36 AM

Another question:

I don't understand John 7:5 here. The brothers are the subject of the verb. Or is "subject" used in a different sense?

Posts 1507
Forum MVP
Fr Devin Roza | Forum Activity | Replied: Fri, Feb 13 2015 7:42 AM

Francis:

I don't understand John 7:5 here. The brothers are the subject of the verb. Or is "subject" used in a different sense?

I've had the same question before, but with a different subject. In your example, apart from John 7:5, John 2:11 and John 12:44 do not have Jesus as grammatical subject either, and yet those results are also present. Either the "subject" category is not the syntactical subject, or there are a lot of incorrectly classified subjects, enough to warrant a thorough review of the entire database.

Posts 1507
Forum MVP
Fr Devin Roza | Forum Activity | Replied: Fri, Feb 13 2015 7:44 AM

Francis:

Good find!

Apparently, it makes no difference which terms the parentheses are applied to:

Looks like a bug to me. It might be worth reporting in a separate thread if Bradley or one of the developers don't see this one. 

Posts 3680
Francis | Forum Activity | Replied: Fri, Feb 13 2015 7:45 AM

Fr Devin Roza:
John 2:11 and John 12:44 do not have Jesus as grammatical subject either

I did not mention these in my question about these because I noted that Jesus is the subject of another verb in these verses and thought that perhaps this is why they showed up. 

Fr Devin Roza:
Looks like a bug to me

Why do you think it is a bug? In this specific search, I cannot see what difference it could make. If I say I want (Francis AND Graham) AND Devin, or (Graham and Devin) AND Francis, or (Devin AND Francis) AND Graham. I should still get the same results. It would make more of a difference with a different combination of operators (try replace the second AND with OR). In the latter case, parentheses would be critical to make sure the results are correct: is it (Devin AND Graham) OR Francis or is it Devin AND (Graham OR Francis)? Perhaps when there is more than two operators, Logos requires parentheses by default to avoid this kinds of ambiguities?

Posts 21928
Forum MVP
Graham Criddle | Forum Activity | Replied: Fri, Feb 13 2015 8:10 AM

Francis:
In the latter case, parentheses would be critical to make sure the results are correct

Absolutely

Francis:
Perhaps when there is more than two operators, Logos requires parentheses by default to avoid this kinds of ambiguities?

It does seem to require it in the Clause Search - and I think this may be what Fr Devin was suggesting is a bug - but it doesn't in other searches:

Posts 244
LogosEmployee
Ryan Gano (Faithlife) | Forum Activity | Replied: Fri, Feb 13 2015 10:41 AM

Francis:

Why do I get this:

You get that because of a (dare I say it?) ...bug. This will be fixed in an upcoming release. As pointed out, in the meantime parentheses around two of the search terms will resolve the issue.

Ryan

Posts 3680
Francis | Forum Activity | Replied: Fri, Feb 13 2015 11:43 AM

Ryan Gano:
This will be fixed in an upcoming release.

Yes

Ryan, any info on the subject: issue mentioned above as well?

Posts 1507
Forum MVP
Fr Devin Roza | Forum Activity | Replied: Fri, Feb 13 2015 12:01 PM

Francis:

I did not mention these in my question about these because I noted that Jesus is the subject of another verb in these verses and thought that perhaps this is why they showed up. 

But, note how the clauses where Jesus is the subject are not highlighted. That would seem to indicate that Logos is not thinking of that clause.

Ideally, a clause search should work within clauses (note that the search box says "search clauses within..."), regardless of whether they cross verse boundaries or not, or whether they are only part of a verse. I'm not quite sure how Logos clause search works as regards verse boundaries, so the point you bring up could explain some misfires or apparent misfires... but it would also indicate a problem with highlighting in that case.

So, I would think that a search for "subject:Jesus AND verb-lemma:pisteuw" should only show those clauses where both Jesus is the subject and the verb is pisteuw. 

Posts 21928
Forum MVP
Graham Criddle | Forum Activity | Replied: Fri, Feb 13 2015 12:44 PM

Fr Devin Roza:
So, I would think that a search for "subject:Jesus AND verb-lemma:pisteuw" should only show those clauses where both Jesus is the subject and the verb is pisteuw. 

Just been looking further at this and it could be argued that this condition is metSmile

Take 7:5, for example:

Because of the presence of αὐτόν within the Subject Function - and with αὐτόν being tagged as "Person Jesus"

technically the search is returning correct results.

Posts 244
LogosEmployee
Ryan Gano (Faithlife) | Forum Activity | Replied: Fri, Feb 13 2015 12:49 PM

Francis:
Ryan, any info on the subject: issue mentioned above as well?

No info from me but I've passed the question on to those with a better understanding of the underlying data.

Posts 1507
Forum MVP
Fr Devin Roza | Forum Activity | Replied: Fri, Feb 13 2015 1:10 PM

Graham Criddle:

technically the search is returning correct results.

I guess that depends how one defines "correct". Hmm Good find though - that certainly might explains some of the problems, or at least help set expectations as to how to interpret the data if that's the case and the criteria used.

I did a quick check on the verses I had reported for problems with "subject" in the Old Testament, with mixed results. In a couple of cases, the Andersen-Forbes database was incorrect, with "elohim" marked as "subject" when they were actually object (1 Sam 8:8 - the clause "and serving other gods", and 2 Kings 17:7 - the clause "they had worshipped other gods", for example). For other clauses however in the list of examples, in the Syntax database by Andersen-Forbes "elohim" was correctly marked as object, but in the clause database appears as subject.

Posts 3680
Francis | Forum Activity | Replied: Fri, Feb 13 2015 1:13 PM

Fr Devin Roza:
So, I would think that a search for "subject:Jesus AND verb-lemma:pisteuw" should only show those clauses where both Jesus is the subject and the verb is pisteuw.

Good point.

Graham Criddle:
technically the search is returning correct results.

I follow your reasoning but I'm not convinced. The reference to Jesus modifies the actual subject "His" brothers, but is not the subject itself, which is what subject:Jesus would seem to indicate. Jesus is not the subject or part of the subject (He does not perform the action) of pisteuo but its Stimulus. But I am ready to be corrected if the clause search does actually work the way you suggest and returns any element of a subject segment in the clause analysis.

Ryan Gano:
I've passed the question on to those with a better understanding of the underlying data.

Thank you, Ryan.

Posts 21928
Forum MVP
Graham Criddle | Forum Activity | Replied: Fri, Feb 13 2015 1:39 PM

Francis:

Graham Criddle:
technically the search is returning correct results.

I follow your reasoning but I'm not convinced. The reference to Jesus modifies the actual subject "His" brothers, but is not the subject itself, which is what subject:Jesus would seem to indicate. Jesus is not the subject or part of the subject (He does not perform the action) of pisteuo but its Stimulus. But I am ready to be corrected if the clause search does actually work the way you suggest and returns any element of a subject segment in the clause analysis.

I do find the behaviour strange but am speculating it might at least explain what we are seeing!

Will be interested to see what the Faithlife people say

Posts 131
LogosEmployee
Peter Venable | Forum Activity | Replied: Fri, Feb 13 2015 5:02 PM

It is definitely a bug that searching for subject:Jesus brings up John 7:5. The bug causes entities that are not the subject but are mentioned inside the subject to be included in the results. I think the same bug also affects all the other syntactic and semantic categories. It should be fixed soon.

Posts 3680
Francis | Forum Activity | Replied: Fri, Feb 13 2015 10:09 PM

Peter Venable (Logos):
It should be fixed soon.

Yes

Posts 21928
Forum MVP
Graham Criddle | Forum Activity | Replied: Fri, Feb 13 2015 11:13 PM

Peter Venable (Logos):
The bug causes entities that are not the subject but are mentioned inside the subject to be included in the results.

Peter Venable (Logos):
It should be fixed soon.

Thanks Peter

Posts 1507
Forum MVP
Fr Devin Roza | Forum Activity | Replied: Sat, Feb 14 2015 3:26 AM

Great news, thanks!

Page 1 of 1 (20 items) | RSS