Re: Verb Tense Usage Statistics

From: Wes Williams (
Date: Wed Mar 04 1998 - 17:43:54 EST

> Morphology Asst. -- Verb, perfect tense, any mood, any voice, any person,
> any number = 857
> Command line -- *@v?r?--?? gave 857
>From acCordance statistics -- Perfect, 1,571Here is an answer from the
Bibleworks list. I omitted the search technical criteria, but kept the
explanations of the different tagging systems that underlie Friberg and
Accordance, which may be of interest to list members.

> Thanks for your note. I'll respond as best I can to a complex
> question. I general you have to realize that Friberg sometimes
> prefixes or suffixes forms with a '+' or '-' to indicate contextual
> considerations. Also, you have to be pretty familiar with the way he
> codes things to get what you want. Also he will sometimes give a
> person, gender or number where you don't expect it because it gives
> additional info about usage in context. For example some of his
> participles are assigned a person because they are not usually
> translated as a participle. <snip>

> Also remember that participles, though they are verbs,
> have a different code structure than regular verbs, of course. You
> have to keep this in mind. On to your examples:
> > Morphology Asst. -- Verb, perfect tense, any mood, any voice, any person,
> > any number = 857
> > Command line -- *@v?r?--?? gave 857
> > >From acCordance statistics -- Perfect, 1,571
> If you want to final all perfect tenses the search to do would be
> ..*@v?r*
> or
> '*@v?r*
> Termination a morphological construction will pick up all forms of
> any length. In your example here you would miss all the participles.
> I get 1501 and 1489 hits for this. It's different from Accordance
> but that doesn't bother me. There are a number of cases where it's a
> subjective call. The BW string and AND searches give slightly
> differing results because the AND search counts a multiple parsings
> twice if your are doing an and search. For gathering statistics you
> should use the string search method. See the manual, which, I think
> discusses the situation.
> > Morphology Asst. -- Verb, present tense, any mood, any voice, any person,
> > any number = 7,031
> > Command line -- .*@v?p?--? gave 6,932
> > >From acCordance statistics -- Present, 11,583
> Again, use
> '*@v?p* which gives 11,328 hits.
> And so on with your other examples. Do not expect Accordance and
> Friberg to agree completely. The differences arise from the fact that
> in some cases the codes are judgement calls and in other cases
> because the coding schemes and the way they classify parts of speech
> are quite different. Friberg's classification system is more "usage"
> ariented, while Accordance is more morphological. Both schemes are
> useful and necessary. Friberg is especially goo when you are looking
> for grammatical usage patterns. His scheme takes some getting used
> to, but you can do some things that are not possible with other
> coding schemes.
> By the way, the next release will have multiple morphological
> systems. Friberg will still be there, but we will also have a scheme
> that is much closer to the methods used by Accordance/Gramcord. Both
> systems are useful, and for serious work, necessary.

This archive was generated by hypermail 2.1.4 : Sat Apr 20 2002 - 15:39:08 EDT