PDA

View Full Version : rpeated phrase search



almather
04-07-2010, 02:34 PM
This morning I was reading through the book of Hebrews and thought I might have seen a few repeated phrases of say 4 to 5 words. So I thought I would do a quick search of Hebrews to see if I might be correct and if so what these phrases are.

I can not figure out how to do this. BW8 and I was reading ESV.

I thought - ( '* * *) - but that gives me "No match found"

then - ( "#1 #2 #3 ) - but of course that gives me any three words, as I thought the above would as well.

If I plug a word into the first example I can get three word phrases containing that word, but is it possible to search through a book for unknown phrase that might be repeated within that book and if so how, please?

I've tried the phrase matching tool but haven't had success with that, for this purpose either.

Thank you.

Al

Sansom48
04-07-2010, 02:46 PM
Hopefully someone more knowledgeable than me will answer, I do not know how to solve any of your true problems, the phrase matching tool seems to require you to go after a very specific phrase. I would suggest the KWIC/Collocation Table, but that you need to at least have a word to start with, however, it may help you to narrow down your search if you can start with a singular word and see how many time certain words appear before and after it in a particular book. If you use your words list manager to generate a list of the most common words occurring in Hebrews and then do a KWIC table off that, but this would be a long process. You could also try the related verses tool but again you almost need to know what you are searching for. I am sure that you thought of most of these things, sorry I do not know more.

Michael Hanel
04-07-2010, 07:13 PM
This morning I was reading through the book of Hebrews and thought I might have seen a few repeated phrases of say 4 to 5 words. So I thought I would do a quick search of Hebrews to see if I might be correct and if so what these phrases are.

I can not figure out how to do this. BW8 and I was reading ESV.

I thought - ( '* * *) - but that gives me "No match found"

then - ( "#1 #2 #3 ) - but of course that gives me any three words, as I thought the above would as well.

If I plug a word into the first example I can get three word phrases containing that word, but is it possible to search through a book for unknown phrase that might be repeated within that book and if so how, please?

I've tried the phrase matching tool but haven't had success with that, for this purpose either.

Thank you.

Al

I have a few ideas, but could you give a more specific example of what you want to be found because I'm not 100% sure. I think what you're looking for is a repeated set of words (but not in the same verse) in the entire book of Hebrews. So you're not looking for a phrase string "SAMEWORD SAMEWORD SAMEWORD" in the book, but you want to see how many times "SAME PHRASE" is used?

almather
04-07-2010, 08:26 PM
Thanks for the replies. Yes, Michael, that is the idea. I thought about half way through the book or so "I've read that before" but at the time I did not make a note. I thought I would just search for it, and then I thought "are there other phrases?" So I'm looking for word one, followed by word two, by word three, etc, within a larger context, in this case the whole book of Hebrews, that may be repeated again. Certainly if you know one word that is a good starting point, but suppose you don't know any of the words but you want to find repeated phrases.

Al

Adelphos
04-08-2010, 12:10 AM
I don't know how you would do it if all three words were unknown, but if one or more words are known then it's pretty simple. For example, if you want to find all phrases that contain three words and that begin with "they" you would simply type --

'they * *

Or if you wanted to find all phrases that contain three words with "shall" as the second word, then you would type --

'* shall *

and so on.

almather
04-08-2010, 02:13 PM
Thanks for your comments. I guess I'll just have to take notes next time I am reading.

Where is that new search feature announce April 1 when you really need it??:D

Al

Michael Hanel
04-08-2010, 02:25 PM
Thanks for your comments. I guess I'll just have to take notes next time I am reading.

Where is that new search feature announce April 1 when you really need it??:D

Al

I've tried thinking about it too, but I think Adelphos is pretty much right. I was thinking maybe somehow you could use wordlists, but if you want it to pick out and make its own matches, that seems to me pretty complicated.

almather
04-08-2010, 03:55 PM
It seems to me, a person who knows nothing about programming, that using #1 #2 etc ought to be close.

Reason? a search <'#1 #1 #1> does not yield every three word phrase in the Bible, or whatever passage you are searching, but will produce "Holy, Holy, Holy" or "Woe, Woe, Woe", etc. So #1 will pick up different words but will also limit its second and third occurrences to match the first word selected.

Therefore, I would think the phrase search <'#1 #2 #3> would work the same way only yielding a three word phrase with three different words. For each phrase selected #1 would be the same word as would #'s 2 and 3, but as in the "Holy..." and "Woe, ..." examples #1 would pick up different words. I think the complication would come in limiting the search to only those 3 (or however many words plugged in) that are repeated 2 or more times in the text. So maybe if there were a way to ask the search to show results for only those combinations that are found 2 or more times?

Well, onward and upward.

Al

MGVH
04-08-2010, 09:16 PM
Well, this is a start...
I used the Graphical Search Engine, and the trick was adding in an extra word term in the middle that requires at least a one word separation between the two phrases.
The search runs nearly instantaneously, and that's probably what you would usually do. I ran it across the Pauline corpus, and it only took a couple seconds.
One thing to note on the results highlighted. The first hit you will see is the start of the 3 word phrase. The matching 3 word phrase will actually be 4 words long, but the first word is the word I had to add in to separate the phrases.
The problem: this will only find matching 3 word phrases within a verse.
I tried turning on cross-verse searching (in the GSE under Query > Properties > Flags), but it indicates that it wants Range Filters instead. (I don't quite understand those, but I don't think it's what we want.)
I also tried using verse Proximity (in GSE, under a merge box), but no luck either.

Maybe someone can take it another step...

Adelphos
04-08-2010, 10:09 PM
Actually, Mark, that looks like a very good start. It's very straightforward once you actually see it, but so hard to conceptualize beforehand. Good job.

Hopefully it can be refined now that you have given it form.

almather
04-08-2010, 10:15 PM
Thank you, this looks like progress. Something to play with.

Al

almather
04-09-2010, 07:33 AM
Mark, thanks for posting this. It is a start, and I am helped also by seeing it graphically. Now instead of studying and finishing up other things that I hope to get done today I can play with this.:)