[Rod Decker]

Is it possible to search for such things as sentence length in a corpus? (I'm thinking primarily of the Greek texts here--GNT or GNT-T.) A particular example: I've noted that Mark 5:25-28 is very unusual for Mark with a 4 verse sentence with multiple clauses, etc. This is normal for other writers, but exceptional for Mark. What I'd like to be able to do is find all sentences (in Mark) that exceed a single verse, ideally sorted according to length. Any ideas?

[David Lang]

Rod,

This was an interesting challenge, and I hope what I've come up with will he helpful to you. As far as I know, there's no way for Accordance to search for one "field" (sentence) which crosses the boundary of another "field" (verse). However, by searching for the first and last word in a sentence, and specifying the minimum proximity between them, it is possible to get a result which will quickly highlight the long sentences in Mark.

Here's what I did:

1. Open a window containing GNT-T and set the Search within every pop-up to Sentence.
2. Open a construct window and drag the LEX element into the first column. When the dialog appears, click OK without selecting anything. In the LEX element, enter an asterisk.
3. Option-drag that LEX element into the second column to copy it there.
4. Drag the PLACE item underneath the LEX item in the first column and enter 1 in the dialog box. This specifies that this word must be the first in the sentence.
5. Drag a WITHIN above the first two columns and enter a large number like 100.
6. Drag a second WITHIN above the columns and enter a smaller number such as 30.
7. Drag a NOT element over the second (smaller) WITHIN.
8. Switch back to your GNT-T window and insert a WITHIN command after the LINK to the Contruct window. Enter 1 in the dialog and click OK.
9. Enter a FIELD command after the WITHIN. In the dialog box, select End and click OK.

This search will find every place that a word at the end of a sentence is at least 30 but not more than 100 words from the first word in the sentence. Most of the hits in Mark spanned several verses, and this search did find your example of Mark 5:25-28.

Hope this helps.

[Rod Decker]

Thanks! That's one I'd have never figured out on my own. It does just what I needed. Turns out there are about 7 sentences of the length I'd identified (or longer). Length is no marker of complexity, but this gives the raw data to evaluate. The most complex sentence in Mark looks to be 7:2-5 with more than 80 words and 9 clauses, most of which are subordinate. Contrast with 10:32b-34 with 50+ words but 10 clauses, almost all of which are simple και-connected in a very Semitic style.

[Mick Matousek]

David Lang, on 14 November 2008 - 10:28 AM, said:

This was an interesting challenge, and I hope what I've come up with will he helpful to you. As far as I know, there's no way for Accordance to search for one "field" (sentence) which crosses the boundary of another "field" (verse). However, by searching for the first and last word in a sentence, and specifying the minimum proximity between them, it is possible to get a result which will quickly highlight the long sentences in Mark.

David, not to be too trival, but how could one find the shortest sentences ? Your solution finds 83 shortest sentences in 76 verses of the GNT, but they all have one intervening word, so are effectively 3 word sentences. John 1:21 ends with a one Greek word sentence. John 11.35 is 3 Greek word sentence.

[David Lang]

Mick,

This search will find all single word sentences, provided you set the Search within every pop-up to sentence:

[FIELD Begin] <WITHIN 1 Words> ?* <WITHIN 1 Words> [FIELD End]

To find sentences of additional length, just add an additional ?* for each word you want to represent, like this:

[FIELD Begin] <WITHIN 1 Words> ?* ?* <WITHIN 1 Words> [FIELD End]

[FIELD Begin] <WITHIN 1 Words> ?* ?* ?* <WITHIN 1 Words> [FIELD End]

Hope this helps.

[Mick Matousek]

David Lang, on 04 December 2009 - 05:10 PM, said:

Hope this helps.

That was the ticket! Eight one word sentences in GNT. Thanks.

[A. Smith]

I'm trying to understand the logic of accordance search syntax. I understand what the commands in the construct window are doing, but how do the WITHIN and FIELD commands function in this search? Actually, I understand that these commands search for the last word of the sentence. What I don't understand is why they do that. It doesn't make sense to me. Also, why couldn't that have been specified somehow in the construct search? I hope this makes some sense. Thanks

This post has been edited by A. Smith: 07 December 2009 - 08:25 AM

[David Lang]

Anthony,

The FIELD command is a standalone command which represents the beginning or end of the field selected in the Search within every pop-up of the More options. The ?* represents any word. By specifying that this word must be within 1 word of the beginning of the field and within 1 word of its end, we effectively say that we want to find a one-word sentence, clause, etc.

This search could be done Using the Construct window, except that the Construct window does not have a FIELD item that could be placed in a column. To do this using a Construct, you would drag a LEX item into the first column, click OK to dismiss the dialog, and then type an asterisk. To specify that this word must be the first item in the field you could drag the Place item below the LEX item in the first column and give it a place of 1. The problem is that there is no way to specify that this word must also be at the end of the field. You would therefore need to go back to the search window your construct window is linked to, and place <WITHIN 1 Words> [FIELD End] immediately after [LINK Greek Construct].

Hope this makes sense.

[A. Smith]

Ok, so with the Within and Field commands after the construct link, you are specifying that, in addition to the construct, you also want to find the last word of the sentence, clause, whatever. This had to be entered into the search window because the construct dialogue does not allow this particular search. Thanks very much for clearing that up for me. Now one more question. It seems to me that there should be an <AND> between the link and the Within command. When there is, nothing changes; the search works the same with or without it. However, sometimes searches require the <AND>. Why is this search indifferent to the <AND>? Thanks again. Understanding this makes accordance much more enjoyable!

[David Lang]

The AND is essentially implied by the WITHIN command. Accordance will therefore accept <AND> <WITHIN ? Words>, but the <AND> is superfluous.