First, a technical note. You're using depth wrong. The search I created with depth =0 was a strictly "tight" search to find "someone" as the syntactic subject of a clause that has a direct speech (i.e., quote) as the verbal complement. By increasing the depth to 2 (not even just 1), you'll get hits further down in the clause that may not have that "someone" as the subject. That is why your search resulted in "shame".
Second, a deeper problem is that when there is a sequence of verbs like "Peter answered
[quote]", my tight search won't identify these at hits. And this is absolutely correct, since it very strict syntactic terms "Peter" is not the syntactic subject of "said". Rather, the syntactic subject is a null anaphor connecting back to Peter as the discourse agent. (At least, this is how it works in the Hebrew database; I know Marco made some principled decisions to depart from our pattern early on and looking at the Greek NT database, it appears that he does not include null subjects but rather connects "answered" and "said" in the same predicate). Due to the "double" nature of such predicates in the Greek database, I ran my search and simply dropped the Pred=verb and the results were all the proper hits. You also listed 16:16 and 22, but neither of those are correct, since in v. 22 there is not speech verb, but rather a follow up quotation introduced by "saying..." and in v. 16 Peter is not the subject, Simon is (and Peter is tagged as an appositive -- not sure I agree with this approach, but that's neither here nor there). Finally, 26:69 is irrelevant.
*edit: by the way, simply adding Simon along with Peter as the lexical item in the subject column catches a few more, incl. 16:16.*
The point of much of this is that the user of the syntax databases must learn to think syntactically. Whether you're using ours, or ETCBC or the Andersen-Forbes database (which I hope will come out in Accordance this year), one cannot simply jump in without figuring out the backbone of the design and expect to get perfect results.
Edited by Robert Holmstedt, 05 April 2017 - 02:02 PM.