frequency stats question

[Donald Cobb]

Hi all,

I'm preparing a cours (for tomorrow!) to show students with minimal knowledge of Biblical languages how to use Accordance. I have a couple of questions. 

 

One of the things I would like to show is how frequency stats can help highlight the content of a book. But I'm running into problems. I would like to show what the vocabulary of Rom 5-8 is, appearing between 50 and 10 times. However, when I do the following command it gives some results that I'm not sure I understand: * [COUNT 10-50] [RANGE Rom 5-8]

 

This actually gives me two lists of words, one that gives words that have anywhere between 1 and 137 hits :

 

 

Then scrolling down the analysis list I come to another list that begins with words that occur 50 or fewer times but then continues with words that only occur once: 

 

 

This second list gives me pretty much what I'm looking for, but why does it not include the words in the first list that occur between 50-10 times? What is the logic assigning one word to one list, one to another? Am I doing something wrong?

 

Second question: if I make a selection of the words occurring between 50-10 times to do a search on those words alone, is there any way of doing is that is easier than copying and pasting each word separately with an [OR] between them? 

 

Thanks for any help you can give me on this!

Edited September 9, 2021 by Donald Cobb

[mgvh]

Interesting...

I think there are a bunch of things going on.

1. Are you looking for words that occur 10-50x just in Rom 5-8?
   OR
2. Words that occur 10-50x in the Greek NT that occur in Rom 5-8?
   OR
3. Words that occure 10-50x in Romans that occur only in Rom 5-8?

So depending on what you want, try some of these options:

• Instead of what you have in your command line, try using AND:    [COUNT 10-50] <AND> [RANGE Rom 5-8]
  That will give you results for #2, I think and returns 201 total words with 102 forms
• Use that command line I just provided but also include a Range line for just Romans. (I.e., click on the plus and use/create just Romans as a Range. Cf. graphic)
  That will give you 525 words and 68 forms. I.e., #3 above. (I think this is what is showing up in your second list.)
• Use same command as I gave but create a range in the Range line for just Romans 5-8
  That will give you 856 words and 40 forms. I.e., #1 above.

Edited September 9, 2021 by mgvh

[Donald Cobb]

Thanks Mark. It's the first search you listed, which means—I think!—that I should have à short list of 14 or so words. But I can't get that by any of the options you suggest. So I really can't figure out if I'm missing something or if there's a bug somewhere in there.

[mgvh]

Hi @Donald Cobb! (a fellow BW user of yore)

I really think that there are 40 words in Rom 5-8 that occur 10-50 times. I've attached what my search looks like. Here's my list displaying NUMBER (not frequency which would be the same in this instance) and counting down.

>>>>>>

Total number of verses = 107
  (total number of verses displayed = 107)

[COUNT 10-50]  (856 total words)

Number of different forms = 40:
(Triple-click a form to see its occurrences)

        γάρ    for, since, then, indeed = 48
        καί    and, also, even, and yet, but = 48
        ἐν    (+dat) in, with, by, to = 46
        ἁμαρτία    (ἁμαρτάνω)    sin, sin offering = 42
        διά    (+gen) through; (+acc) because of = 38
        θεός    god, God = 36
        οὐ    not; no = 36
        εἰς    (εἷς_1)    (+acc) into, to, for (prep) = 33
        νόμος    law, principle = 33
        εἰμί    (ἐιμί)    to be, exist (sum) = 32
        ὅς    who, which, what = 30
        αὐτός    he, she, it; self, same; they (when pl) = 27
        ὅτι    (ὅς, τίς)    that, because, since; why? = 24
        Χριστός    (χρίω)    Christ = 24
        πνεῦμα    (πνέω)    wind; breath; spirit, Spirit = 23
        ἀλλά    (ἄλλος)    but, rather, on the contrary = 22
        θάνατος    (θνῄσκω)    (m) death = 21
        σύ    you; you people (when pl) = 19
        οὗτος    this, this one, he = 18
        ἀποθνῄσκω    (ἀπό, θνῄσκω)    to die = 17
        εἰ    if; whether = 17
        σάρξ    (f) flesh, meat, body, sinful nature = 17
        Ἰησοῦς    Jesus; Joshua = 14
        οὖν    therefore, thus = 14
        κατά    (+gen) against, down; (+acc) according to = 13
        τίς    who? what? why? = 13
        ἀπό    (+gen) from, by, since = 12
        εἷς    (εἷς_1)    one = 12
        ζάω    to live = 12
        ζωή    (ζάω)    life, living, way of life = 12
        ἐκ    (+gen) of, out of, from = 11
        μή    not, lest = 11
        χάρις    grace, favor; gratitude; gift = 11
        ἄνθρωπος    man, human = 10
        γίνομαι    to become, be, be born, be created = 10
        ἵνα    that, in order that, so that = 10
        οἶδα    to know = 10
        οὔτε    (οὐ, τέ)    neither = 10
        πᾶς    all, every; the whole = 10
        πολύς    much, many; (adv) more = 10

[mgvh]

I just fired up BibleWorks and ran the search and used the Wordlist Manager. Here are the results. Same 40 words.

γάρ @* 48

καί @* 48

ἐν @* 46

ἁμαρτία @* 42

διά @* 38

θεός @* 36

οὐ @* 36

εἰς @* 33

νόμος @* 33

εἰμί @* 32

ὅς @* 31

αὐτός @* 27

ὅτι @* 24

χριστός @* 24

πνεῦμα @* 23

ἀλλά @* 22

θάνατος @* 21

σύ @* 19

οὗτος @* 18

ἀποθνῄσκω @* 17

εἰ @* 17

σάρξ @* 17

ἰησοῦς @* 14

οὖν @* 14

κατά @* 13

τίς @* 13

ἀπό @* 12

εἷς @* 12

ζάω @* 12

ζωή @* 12

ἐκ @* 11

μή @* 11

χάρις @* 11

ἄνθρωπος @* 10

γίνομαι @* 10

ἵνα @* 10

οἶδα @* 10

οὔτε @* 10

πᾶς @* 10

πολύς @* 10

[Donald Cobb]

Yes Mark, I miss the BW Wordlist manager for things like this! I'm not actually interested in every word in that range, but rather the specific vocabulary that comes out in these chapters (e.g., ἁμαρτία, νόμος, πνεῦμα, θάνατος, ἀποθνέσκω, σάρξ, ζἀω, ζωή...). So, yes, there are more words than what my comments suggested. I should have been clearer.

 

I always found the Wordlist manager helpful for things like this! But it may just be that I was so used to it.

 

Actually, I'm trying to get this worked out, mostly to demonstrate what can be done with Accordance and what a search like that can suggest exegetically. But I appreciate you taking the trouble to look into this! 

Edited September 9, 2021 by Donald Cobb

[mgvh]

This is only an approximate measure, but here's an exercise I've done with Romans.

• Set Range to Romans.
• Run: *@[verb indicative]
• Open Hits Graph > Click on Keep
• Using same tab, run: *@[verb imperative]

Now you will see rather clearly that the imperatives start piling up after ch11. I.e., Paul is being 'friendly' in the first 10 chapters or so (especially since he doesn't know the Roman community), and then he starts giving advice once he's established his credentials.

[mgvh]

Yeah, I miss the Wordlist Manager too...

 

Ok, I just had another idea.

• Set the Range to Romans
• Run this search: (ἁμαρτία, νόμος, πνεῦμα, θάνατος, ἀποθνέσκω, σάρξ, ζἀω, ζωή)
  • (Note that I just copied your list above. You can add whatever other words you want)
• Now show the Hits graph, and you can see the pile up of those words in Romans 5-8

 

(Remove the Romans Range restriction, and search for those words across the whole NT, and it is even more dramatic!)

Edited September 9, 2021 by mgvh

[Donald Cobb]

9 hours ago, mgvh said:

Yeah, I miss the Wordlist Manager too...

 

Ok, I just had another idea.

• Set the Range to Romans
• Run this search: (ἁμαρτία, νόμος, πνεῦμα, θάνατος, ἀποθνέσκω, σάρξ, ζἀω, ζωή)
  • (Note that I just copied your list above. You can add whatever other words you want)
• Now show the Hits graph, and you can see the pile up of those words in Romans 5-8

 

(Remove the Romans Range restriction, and search for those words across the whole NT, and it is even more dramatic!)

 

That's actually one of the things I'll be doing. The idea is to show how the graphs can help give a quick idea of what the text is actually saying through the repetition of key terms. 

 

Interestingly, your comment on doing a search on imperatives gave me the idea of simplifying my search. It struck me as a no-brainer ("why hadn't I thought of this before?") to just search on all verbs and substantives that occur between 10-50 times in this range. But when I do this, the results are totally skewed. I don't know why. Perhaps @Joel Brown  could weigh in here, but it seems to me that there are some conceptual flaws that prevent doing this kind of search, either in terms of ease or precision in results. 

 

Thanks for your thoughts and interaction, Mark !

 

 

 

Edited September 10, 2021 by Donald Cobb

[Λύχνις Δαν]

Hi Donald,

 

  Do you mean something like :  ([verb],[noun])@ [COUNT 10-50] ?

 

Thx

D

Edited September 10, 2021 by Λύχνις Δαν

[mgvh]

Thanks, @Λύχνις Δαν. Now I have to think about how that use of @ works. It does explain @Donald Cobb's initial question about getting two lists of results, since he used an <AND>, and so Accordance was really trying to do two things. The @ limits it to just one search.

 

So, how about this search to get a list of prominent words in Rom 5-8:
([verb],[noun],[adjective]) @ [COUNT +10] <AND> [RANGE rom 5-8]

 

Then, if you want, you can open another NA28 tab and use the HITS command to compare the frequency of those words in Rom 5-8 to other ranges of text.

[Λύχνις Δαν]

I think 's original problem was * [count ..][range...] (there was no explicit AND. At least I didn't see one) which resulted in the * and the count being treated as separate adjacent words in the search, the first being any word and the second being one whose lex occurs 10-50 times. I nearly said as much yesterday but the thread went down another track and it seemed less relevant.

 

Bear in mind the difference in counting with [RANGE] and with the range pull down. The former counts all words in the pulldown selected range and then restricts what it shows to the [RANGE], where the other constrains the search itself to the pulldown selected range.

 

Thx

D

 

[Donald Cobb]

Thanks guys! Daniel, I just tried this:

 

([VERB], [NOUN])@ [COUNT 10-50] [RANGE Rm 5-8], adding Romans in the pull-down range. It pretty much reproduces your suggestion and seems to give good results on the upper end (first hit, ἁμαρτία, 42 occurrences). However, try as I might, I still get results going down to just one occurrence. Mark, I tried your suggestion, but nothing seems to eliminate both the upper limit and the lower limit. 

 

BTW, I just got back from giving my course and it raised a lot of interest for Accordance and what it can be used for. 🙂

 

HOWEVER, one student did notice that I had to manually copy the words from the analysis tab list into the search box and said: can't you just drag and drop? Good point! That isn't possible but it would be a nice addition!

[Λύχνις Δαν]

Romans in the pull down range means that the COUNT is for the whole of Romans while the hits shown will be confined to the [RANGE] command's Rom 5-8. If you remove the [RANGE] you will see that all hits show 10-50 counts. What you need to do is create a defined range of Rom 5-8 and select that in the pull down. Then of course you also will not need the [RANGE ] command anymore.

 

Thx

D

Edited September 10, 2021 by Λύχνις Δαν

[mgvh]

As Daniel noted, you need the <AND> with the Range command in the command line.

Daniel also identifies the distinction between using Range in the command line and using the extra parameter line and creating a specified range.

Thanks, Dan!

[Λύχνις Δαν]

15 minutes ago, Donald Cobb said:

HOWEVER, one student did notice that I had to manually copy the words from the analysis tab list into the search box and said: can't you just drag and drop? Good point! That isn't possible but it would be a nice addition!

BTW, while the student is right that you cannot drag and drop, you can select a word in the Analysis tab and right click -> Look Up or Search etc.

 

Thx

D

[mgvh]

2 hours ago, Donald Cobb said:

HOWEVER, one student did notice that I had to manually copy the words from the analysis tab list into the search box and said: can't you just drag and drop? Good point! That isn't possible but it would be a nice addition!

It might not be getting at what you want, but here is where opening another tab and using the HITS command can act upon the words that appeared in the analysis list.

[Donald Cobb]

16 hours ago, Λύχνις Δαν said:

Romans in the pull down range means that the COUNT is for the whole of Romans while the hits shown will be confined to the [RANGE] command's Rom 5-8. If you remove the [RANGE] you will see that all hits show 10-50 counts. What you need to do is create a defined range of Rom 5-8 and select that in the pull down. Then of course you also will not need the [RANGE ] command anymore.

 

Thx

D

 

Thanks Daniel! This is the one that gets me pretty much where I was wanting to go: I get a short list of the nouns and verbes in Rm 5-8 that appear between 10-50 times in those chapters. I was wanting to avoid making a category for that specifically in the the drop-down rate, just to keep that from having a bunch of ranges that I only use on specific occasions and that might end up cluttering that drop-down range window (it would be nice to do the same for, e.g., Rom 1-4 ; 9-11; 12-16, etc., etc.). But that's no great problem.