BUG?
With the bible search set to search all text in NT in NASB with match all word forms when I do a search on the word guilt I only get the word guilt showing in my hits but not guilty. I tried with the word full and did get other hits such as fullness. Shouldn't I get all forms of the word guilt showing? When I put in the word guilty it does give me hits on the word guilty.
Comments
-
Hi Tim
This doesn't seem to be specific to the NASB - I get the same results in the ESV and NIV.
Graham
0 -
TimEngwer said:
With the bible search set to search all text in NT in NASB with match all word forms when I do a search on the word guilt I only get the word guilt showing in my hits but not guilty. I tried with the word full and did get other hits such as fullness. Shouldn't I get all forms of the word guilt showing? When I put in the word guilty it does give me hits on the word guilty.
I'm not sure that "Match all word forms" recognizes endings that convert nouns into adjectives (though it does appear to go the other way around if it finds "fullness" from "full"). I tried searching for all word forms of "fool" and it didn't find "foolish" and for "god" in The Message, and it didn't find godlike.
I don't know if there's any documentation anywhere of what types of "word forms" this feature is able to find. Certainly the basic noun ones (plural -s) and verb ones (-ed, -ing), and probably many verb-to-noun ones (-ment: govern --> government), but not all (govern --> governor doesn't get found).
These are either bugs or "we never intended to do that" ("by design") or "it's too difficult" kinds of things. Someone from Logos will need to weigh in.
0 -
where do I select "Match all word forms"?
0 -
0
-
Thank you, Rosie, you are always wonderfully helpful
0 -
I suspect it may be by design but then again I guess we won't know until logos staff comments.
0 -
i am running 4.0d.
When searching for guilt i don't pick up guilty.
But if you look at the original language for the word translated guilt and the word guilty you will find that they are different. So i would suspect that is why we don't see them under match all forms. Match all forms possibly looks for all variations of the lemma for that word.
Also notice that the lemma for guilty is the only word that is translated from this greek word (not guilt, so we see the difference in our language from the greek) was wrong about this [:$]
0 -
If you want guilt and guilty you could search for "guilt?" OR "guilt*" (no quotes)
Prov. 15:23
0 -
steve clark said:
When searching for guilt i don't pick up guilty.
You get the same result in any resource because the algorithm for matching English word forms does not think they have the same stem/root. It is a different version of the stemming algorithm used in L3 and it is fallible! Note that stemming is not applied to languages for which lemmas are defined eg. Greek.
Dave
===Windows 11 & Android 13
0 -
Hi Dave
I'm not sure I follow.
You say that stemming is not applied to languages where lemmas are defined, such as Greek, so does this not mean that it doesn't apply to searches in the New Testament?
Presumably lemmas can't be used in languages which don't have RIs - I am assuming the required tagging is not done. In this case, does it revert back to use of the stemming algorithm?
Searching for "love" in John in the ESV brings back answers for words which come from both φιλέω and ἀγαπάω which are clearly different 'roots' so I don't see how the algorithm can work on the basis of starting from the same one.
I'm confused!
Graham
0 -
I would say it is bug since it is not consistent behavior with other words. When you write guilt* it does find all word forms. When you try other similar words like "sin" it gives you all words forms (without the star sign). The function "Match all word forms" works with the word "sin" well, however with the word "guilt" it does nothing. I think we should report it as a bug.
P.S. It has nothing to do with Greek RI lemma usage. When you try the same search in the NET Bible (no RI yet) it works exactly the same way.
Bohuslav
0 -
Interesting thing is that there are other words which give us the same results.
"front" gives just that form of the word.
"front*" gives all forms, including frontier, frontlets etc.
So, I think we need to hear from Logos on that matter.
Bohuslav
0 -
Bohuslav Wojnar said:
I would say it is bug since it is not consistent behavior with other words. When you write guilt* it does find all word forms. When you try other similar words like "sin" it gives you all words forms (without the star sign). The function "Match all word forms" works with the word "sin" well, however with the word "guilt" it does nothing. I think we should report it as a bug.
Using the asterisk does not mean the same as "all word forms." It could be coincidental that a word begins with the same few letters, but it isn't related to the same root. For example "sin*" would find sins, sinner, sinning, sinned, but it would also find single (completely unrelated word), however you wouldn't want "all word forms" to do that. Also, one might expect "all word forms" to include prefixed versions of words ("make" --> makes, maker, remaking), but as far as I can tell it doesn't.
0 -
Bohuslav Wojnar said:
Interesting thing is that there are other words which give us the same results.
"front" gives just that form of the word.
"front*" gives all forms, including frontier, frontlets etc.
So, I think we need to hear from Logos on that matter.
I think one could argue that frontier and frontlets are different words derived from the same root, not different inflections of the word front. I would expect "front*" to find them, but not "all word forms" which I think is for inflections, not different words derived from the same root.
0 -
Graham Criddle said:
You say that stemming is not applied to languages where lemmas are defined, such as Greek, so does this not mean that it doesn't apply to searches in the New Testament?
Stemming will apply when searching for English words in any
resource when you select "Match all word forms". The stems are derived
from a rule & pattern-matching algorithm but, as mentioned above,
"guilt" doesn't pick up "guilty" (and vice-versa) even though one would
think that "guilt" is the stem. Stemming is an alternative to performing a dictionary lookup of words (lemmas).Graham Criddle said:Searching for "love" in John in the ESV brings back answers for words which come from both φιλέω and ἀγαπάω which are clearly different 'roots' so I don't see how the algorithm can work on the basis of starting from the same one.
"love" will pick up "loves", "loved", "lovely", "loving" - which could
be based on the stem "lov" (it doesn't have to be a real word!).Dave
===Windows 11 & Android 13
0 -
Bohuslav Wojnar said:
I would say it is bug since it is not consistent behavior with other words. When you write guilt* it does find all word forms. When you try other similar words like "sin" it gives you all words forms (without the star sign). The function "Match all word forms" works with the word "sin" well, however with the word "guilt" it does nothing. I think we should report it as a bug.
It's not a bug. It's a function of the Porter stemming algorithm used in both L3 and L4 (with some customisation by Logos). To see this type "guilt guilty" in the top box at http://qaa.ath.cx/porter_js_demo.html - the "Stemmed" box shows the different stems!
Type "sinning sinned sinful sinless" to see that sinless is not stemmed by "sin".
Dave
===Windows 11 & Android 13
0 -
Thanks Dave
That clarifies things
Graham
0 -
Rosie Perera said:
I think one could argue that frontier and frontlets are different words derived from the same root, not different inflections of the word front. I would expect "front*" to find them, but not "all word forms" which I think is for inflections, not different words derived from the same root.
Because of its algorithmic nature that logic does not apply - "frontier" and "frontlet" have different stems & neither is stemmed from "front" - use the Porter demo to see this.
Dave
===Windows 11 & Android 13
0 -
Dave,
I guess the problem is that while this is behaving as per the algorithm it results in uncertainty as to what is actually going to be produced for any given search.
I assume this algorithm is "non-trivial" which would explain some "edge" behaviour but it does make me question in what areas this matching capability can be used with confidence when people are not familiar with the stems which are being used.
Graham
0 -
Dave Hooton said:
It's not a bug. It's a function of the Porter stemming algorithm used in both L3 and L4 (with some customisation by Logos). To see this type "guilt guilty" in the top box at http://qaa.ath.cx/porter_js_demo.html - the "Stemmed" box shows the different stems!
Type "sinning sinned sinful sinless" to see that sinless is not stemmed by "sin".
If that's the case, it explains the matter. However I agree than with the precious poster that we are not able to be sure what result we can expect (if we do not know exactly the stem used for each word. For example, what is the logic for using different stem for guilt and guilty? Sin and sinless? May be I do not understand English enough to try to grasp that.
Bohuslav
0 -
Graham Criddle said:
I assume this algorithm is "non-trivial" which would explain some "edge" behaviour but it does make me question in what areas this matching capability can be used with confidence when people are not familiar with the stems which are being used.
The Porter algorithm provides a very good (not perfect) idea of the variant forms of an English word. So if you want to be more precise switch off "Match all word forms" and use wild cards and/or specific forms of the word. Matters would not be more perfect with a dictionary lookup as different morphologies demonstrate that people have different ideas about the lemma for a word.
BTW "guilt" and "guilty" are different dictionary entries in COED & Merriam-Webster.
Dave
===Windows 11 & Android 13
0 -
Bohuslav Wojnar said:
For example, what is the logic for using different stem for guilt and guilty? Sin and sinless?
They have separate entries in an English dictionary!!
Dave
===Windows 11 & Android 13
0 -
Dave Hooton said:Bohuslav Wojnar said:
For example, what is the logic for using different stem for guilt and guilty? Sin and sinless?
They have separate entries in an English dictionary!!
Excellent answer! Here you can see what all the word forms are for each dictionary definition:
These definitions are from Collins 8th ed. (in Logos).
0 -
so searching in the NIV (NT) we can see using a wildcard character produces all manner of words unrelated to sin. And if we search for sin without the wildcard we need to be aware of possible missing forms.
So we need to understand that match all forms may miss some forms.
But according to the English dictionary, we still cannot trust how it differentiates the words: e.g. your second definition shows sins, sinning, sinned, but the match all forms found sins, sinning, sinned but missed sinner.
Search:
sin*Search: sin sin sin sin’s sin’s Sinai since sincere sincerely sinews sinful sinful sing singing single sink sinned sinned sinner sinners sinning sinning sins sins So this is a little scary. I sometimes do searches to bring up in sunday school to emphasize how a word is used or how often. But now i need to search for everything (since i may not remember every form of a word) and sort out unrelated words. Bummer [:(]
Edit: does this incompleteness of searching also pertain to lemma searches?
0 -
Maybe we should consider looking up words http://www.logos.com/ebooks/details/mwthes or here first, then plug that into the Logos4 Search?
Before anyone rushes to buy it, its already included in most of the Logos Packages.
0 -
steve clark said:
Edit: does this incompleteness of searching also pertain to lemma searches?
I trust lemma searches in Logos Reversed Interlinears the most. Practically I do not do any other searches for the word study. I use regular search only to find the particular scripture I am looking for.
I think that thread has quite important information for users to realize the fact how the search of English words function in Logos. What it actually means to do search with "Match all word forms" switched on. The difference between that and the simple use of a wildcard asterisk etc.
It helped me very much. Thanks to all of you for that.
Bohuslav
0 -
steve clark said:
Edit: does this incompleteness of searching also pertain to lemma searches?
If you search (lemmas) for sin ἁμαρτία, ἁμαρτάνω you will not get sinner ἁμαρτωλός; sinful comes from ἁμαρτία, ἁμαρτάνω & ἁμαρτωλός in the derived sense of being full of sin; sinless appears in the form of no sin οὐ ἁμαρτία, οὐ ἁμαρτάνω or without sin ἀναμάρτητος (once only).
I'm no Greek expert but I've gleaned information from Louw-Nida domain 88.289-318 (sin, wrongdoing, guilt) and from lemma searches. So your searches will be incomplete if you don't understand the original language information, and/or don't make proper use of the information, you get from an RI or lexicon.
Dave
===Windows 11 & Android 13
0 -
Dave Hooton said:steve clark said:
Edit: does this incompleteness of searching also pertain to lemma searches?
If you search (lemmas) for sin ἁμαρτία, ἁμαρτάνω you will not get sinner ἁμαρτωλός; sinful comes from ἁμαρτία, ἁμαρτάνω & ἁμαρτωλός in the derived sense of being full of sin; sinless appears in the form of no sin οὐ ἁμαρτία, οὐ ἁμαρτάνω or without sin ἀναμάρτητος (once only).
Yes. That's why I think it is better to find out the relations of the words in the original language than in any translation. That is the value of the embedded Reverse Interlinears in the Logos Bibles that enables us to do searches in the original languages and see the relations of the words etc.
Just one example: recently I studied the word "steadfast love" in Hebrew which is the word Chesed of course. It is for instance so interesting to see the connection of Chesed (steadfast love) and Chasid (holy one or faithful one). In our language that two things are not connected the way we see it in the Hebrew. It made me think a lot about that connection.
But I know I am now (as we say in the Czech) "bringing wood to the forest." [:)]
Bohuslav
0 -
Bohuslav Wojnar said:
Yes. That's why I think it is better to find out the relations of the words in the original language than in any translation. That is the value of the embedded Reverse Interlinears in the Logos Bibles that enables us to do searches in the original languages and see the relations of the words etc.
Just one example: recently I studied the word "steadfast love" in Hebrew which is the word Chesed of course. It is for instance so interesting to see the connection of Chesed (steadfast love) and Chasid (holy one or faithful one). In our language that two things are not connected the way we see it in the Hebrew. It made me think a lot about that connection.
YES.
I am sure there must be many word-plays, rhymes, poetry, "sayings" and all kinds of things; that would be more clear in the original languages. In general, much of this is lost in translation. Sometimes, there will be an attempt to bring some of that into English, or maybe a footnote. Just as Psalm 119 might have some attempt to render the acrostic in a footnote, or maybe headings before each section by letter (NIV and ESV).
0 -
Bohuslav Wojnar said:
But I know I am now (as we say in the Czech) "bringing wood to the forest."
Which means I won't be able to "see the forest for the trees" (as we say in English)[:D]
Dave
===Windows 11 & Android 13
0 -
Bohuslav Wojnar said:
I trust lemma searches in Logos Reversed
Interlinears the most. Practically I do not do any other searches for
the word study. I use regular search only to find the particular
scripture I am looking for.Thanks Bohuslav, i mostly use the original language lemmas too. So i will start using them exclusively when using occurrences as an example in class. (unless its just a FWIW point relative to a translation; and then point that out as well)
Dave Hooton said:I'm no Greek expert but I've gleaned information from Louw-Nida domain 88.289-318 (sin, wrongdoing, guilt) and from lemma searches. So your searches will be incomplete if you don't understand the original language information, and/or don't make proper use of the information, you get from an RI or lexicon.
Thanks Dave, hopefully i have not misled those in my class very often (BTW i'm not the teacher; except occasionally sub when they are gone). i more often than not use the the lemma anyway. But as you point out i will need to investigate more thoroughly in the future. Thanks for the tip on the LN and using the RI and lexicons in my lemma searches.
0 -
Dave Hooton said:Bohuslav Wojnar said:
But I know I am now (as we say in the Czech) "bringing wood to the forest."
Which means I won't be able to "see the forest for the trees" (as we say in English)
Our (Czech) saying actually means that you realize you are speaking about something in a situation most of the people present know more about the matter than you... [:)] We say then: "It is bringing the wood to the forest..." [:)]
Bohuslav
0 -
Bohuslav Wojnar said:
Our (Czech) saying actually means that you realize you are speaking about something in a situation most of the people present know more about the matter than you...
We say then: "It is bringing the wood to the forest..." There's a close analogy. If you only look at the trees you may not see/understand the big picture (the forest).
Dave
===Windows 11 & Android 13
0 -
Bohuslav Wojnar said:Dave Hooton said:Bohuslav Wojnar said:
But I know I am now (as we say in the Czech) "bringing wood to the forest."
Which means I won't be able to "see the forest for the trees" (as we say in English)
Our (Czech) saying actually means that you realize you are speaking about something in a situation most of the people present know more about the matter than you...
We say then: "It is bringing the wood to the forest..." For that, we say in English "bringing coals to Newcastle" -- bringing something to a situation or place that already has more of it than you; Newcastle in England is (or was?) known for its coal industry. Even in the US we still use this phrase which came from our British ancestors...
0 -
Rosie Perera said:
Even in the US we still use this phrase which came from our British ancestors...
Well, I'm blowed! Never knew that.
Dave
===Windows 11 & Android 13
0 -
Dave Hooton said:Rosie Perera said:
Even in the US we still use this phrase which came from our British ancestors...
Well, I'm blowed! Never knew that.
Is this something like our American expression "I'm blown away?"
Differing idioms can lead to some rather humorous situations. Soon after our Australian friend arrived here, he visited Wal Mart and asked several people where he might find a thong (called flip-flops around here). He was quite embarrassed to learn that in this part of the world a thong was a feminine undergarment.
0 -
Rosie Perera said:
For that, we say in English "bringing coals to Newcastle" -- bringing something to a situation or place that already has more of it than you; Newcastle in England is (or was?) known for its coal industry. Even in the US we still use this phrase which came from our British ancestors...
Wow, now that's interesting to me because our region is known for centuries as the black coal center in our part of Europe (Ostrava together with Katowice region, which is Silesia or Schlesien). This saying would work well in our region [:)]
Bohuslav
0 -
Jack Caviness said:
Is this something like our American expression "I'm blown away?"
Yes (as I understand it).
Dave
===Windows 11 & Android 13
0
