Trying to make Sense of the new Bible Sense Search

Page 2 of 2 (34 items) < Previous 1 2
This post has 33 Replies | 2 Followers

Posts 4652
Forum MVP
Fred Chapman | Forum Activity | Replied: Wed, Jul 1 2015 5:17 PM

It is probably best for me to back out as well. I am not a semantic higharky (just kidding I know how to spell it) expert either.Indifferent

Posts 2131
LogosEmployee

Doc B:

Andrew Batishko (Faithlife):
I'll back out of the conversation here. Smile

Understood. It would be great that, as you back away, you shove one of your colleagues who put together the BSL (politely, of course) into your spot at the rail. Big Smile

Jeremy is much more competent to answer than I am.

Posts 113
LogosEmployee

Fredc:

Jeremy, Thanks for your reply, but I think you missed my point. The subject of prayer was just an example to illustrate what I see as a shortcoming in the BSL. There are many topics / subjects where that would also apply. My point is that if someone wanted to do a search for all the _____ (prayers was just an example) they would have to know every sense associated with ______ (again prayers are just an example) before they could perform a complete search. There should be a way to identify a broad category of the sense and then drill down if desired. In other words if I want to identify all the prayers (just and example) I can search <sense to pray> and it will return every prayer. I can drill down from there to find only those prayers that are considered begging. 

Fred:

I think I understand better what you are getting at. My last reply was to the effect that the initial goal of the BSL was to find specific concepts and not broad conceptual categories. With that said, though, do you (and others here as well) think it would be useful to have a search like: <Sense CONTAINS pray>? This could essentially find all occurrences of senses where the string "pray" occurs in the label for the sense. No promises, but if you think that would let you do what you are talking about here, I can pass along the suggestion. Or even better, you or someone else here might suggest it in the Suggestions area of the forum to see if it garners other user support: https://community.logos.com/forums/28.aspx . I'm not sure if I'm technically supposed to post there, though I'm always glad to see any more features added that would get more people using the BSL Smile

Posts 4652
Forum MVP
Fred Chapman | Forum Activity | Replied: Thu, Jul 2 2015 5:51 AM

Thanks for your reply Jeremy. I would like to see a the ability to search for <Sense CONTAINS ______>. It would be great if when that is entered we saw the drop down (that we currently see) with all sense options (i.e. to pray, prayer (act), etc.) This would give the user the option to search for everything that contains pray or drill down to something more specific.

I will create a suggestion in that forum.

Posts 1053
William Gabriel | Forum Activity | Replied: Thu, Jul 2 2015 7:57 AM

Jeremy Thompson:

 With that said, though, do you (and others here as well) think it would be useful to have a search like: <Sense CONTAINS pray>? This could essentially find all occurrences of senses where the string "pray" occurs in the label for the sense. 

I think this suggestion would be incredibly helpful. I'm going to reply to Andrew in a minute with another suggestion.

Posts 4652
Forum MVP
Fred Chapman | Forum Activity | Replied: Thu, Jul 2 2015 8:00 AM

I created a suggestion here

Posts 1053
William Gabriel | Forum Activity | Replied: Thu, Jul 2 2015 8:01 AM

Andrew Batishko (Faithlife):

 It sounds to me that your desired hierarchical structure is different than that chosen by the people that put together the BSL. They were attempting to represent certain structures while you are desiring different ones. This is a really common problem with any kind of semantic hierarchy.

I believe the semantic tree that was created has a good structure. I think we need another dimension, another tree that organizes along senses that make sense to group together. So you might have a new branch "Pray" that points to all the leaves that are in disparate branches of the current tree. You may have a new branch "House" that points to all the varied uses of House, even though they might be radically different in semantic meaning.

I know this would be another large undertaking, both in coding and tagging, but I think average Bible study users (raising my hand here) would get tremendous benefit from having datasets that track with common thinking.

Posts 26988
Forum MVP
MJ. Smith | Forum Activity | Replied: Thu, Jul 2 2015 11:10 AM

William Gabriel:

I believe the semantic tree that was created has a good structure. I think we need another dimension, another tree that organizes along senses that make sense to group together. So you might have a new branch "Pray" that points to all the leaves that are in disparate branches of the current tree. You may have a new branch "House" that points to all the varied uses of House, even though they might be radically different in semantic meaning.

I know this would be another large undertaking, both in coding and tagging, but I think average Bible study users (raising my hand here) would get tremendous benefit from having datasets that track with common thinking.

What is necessary for this to work is for you to define the additional semantic relationships you want captured. Otherwise it ceases to be a semantic lexicon.  Or do you mean expanding the currently relationships to allow "multiple inheritance" i.e. removing the hierarchy restriction and allowing networks of meaning? Or put another way "senses that make sense to group together" needs to be defined as to why it makes sense to group them together otherwise it becomes private. By this I mean I group together a Christmas cactus, a Dresden china tea cup and a white porcelain angel - everyone in my immediate family would instantly recognize the category as "things I was given before age 10 by important unrelated women in my life". That is a meaningful group but it is neither semantic nor public.

That said, I believe that the Bible sense lexicon should be a network not a hierarchy and would like to see the kinds of semantic relationships represented expanded.

Orthodox Bishop Hilarion Alfeyev: "To be a theologian means to have experience of a personal encounter with God through prayer and worship."

Posts 8075
LogosEmployee

MJ. Smith:
That said, I believe that the Bible sense lexicon should be a network not a hierarchy and would like to see the kinds of semantic relationships represented expanded.

It is a network: for example, "cognitive content" relates to "internal feature" via two different paths.

logos4:Senses;KeyId=ws.cognitive_content.n.01

Posts 26988
Forum MVP
MJ. Smith | Forum Activity | Replied: Thu, Jul 2 2015 1:13 PM

Thank you very much - a previous Faithlife post had led me to believe it was strictly hierarchy. I am much relieved.

Orthodox Bishop Hilarion Alfeyev: "To be a theologian means to have experience of a personal encounter with God through prayer and worship."

Posts 1053
William Gabriel | Forum Activity | Replied: Thu, Jul 2 2015 1:43 PM

MJ. Smith:

What is necessary for this to work is for you to define the additional semantic relationships you want captured. Otherwise it ceases to be a semantic lexicon.  Or do you mean expanding the currently relationships to allow "multiple inheritance" i.e. removing the hierarchy restriction and allowing networks of meaning? Or put another way "senses that make sense to group together" needs to be defined as to why it makes sense to group them together otherwise it becomes private. By this I mean I group together a Christmas cactus, a Dresden china tea cup and a white porcelain angel - everyone in my immediate family would instantly recognize the category as "things I was given before age 10 by important unrelated women in my life". That is a meaningful group but it is neither semantic nor public.

That said, I believe that the Bible sense lexicon should be a network not a hierarchy and would like to see the kinds of semantic relationships represented expanded.

I am asking for extra definitions and semantic relationships. Perhaps what I'm asking has too large a scope since we could argue of any number of potential relationships--and we'd probably squabble about them too--but it seems like some groupings make for common sense.

Based on Bradley's response, it sounds like a particular sense can have multiple inheritance / networking. So they'd at least be capable of creating trees with multiple origins.

We currently have:

activity -> worship -> prayer (act)

communication -> message (content) -> prayer (contents)

request -> petition (formal) -> prayer (petition) -> prayer vow

writing (product) -> sacred text -> prayer (text)

These are good senses, I don't disagree with them one bit. But when a typical Christian thinks of Prayer, all of the above are included in that sense, so it seems like you need a prayer root that contains all of them but also keeps the established tree intact.

What comes before it? I'm not sure--maybe communication still.

Or, perhaps each one of these little-p prayers get's a big P Prayer before it and it becomes an abstract root that works with <sense Prayer>

activity -> worship -> Prayer -> prayer (act)

communication -> message (content) -> Prayer -> prayer (contents)

request -> petition (formal) -> Prayer -> prayer (petition) -> prayer vow

writing (product) -> sacred text -> Prayer -> prayer (text)

Posts 26988
Forum MVP
MJ. Smith | Forum Activity | Replied: Thu, Jul 2 2015 5:19 PM

Okay, I understand better what you want - the problem nicknamed the tennis problem ... how to gather together the unrelated elements - ball, net, racquet, backboard, court ... that come together in a tennis games but are not related by any of the standard semantic relationships ... that is the problem that semantic domains are designed to resolve. Certainly the sense lexicon can be categorized into semantic domains and I can see how that would be a useful tool. But it is something very different than what the BSL is designed to achieve. The closest analogy I can think of is that the BSL is designed like a grocery store - produce here, meat there, baking supplies on the 4th aisle ... And you want it organized by recipe ... everything needed for chili con carne here, for potato salad there etc. The for a recipe approach is quite common for "foreign" recipes requiring exotic ingredients. Bothe are useful but they are two different things..

Orthodox Bishop Hilarion Alfeyev: "To be a theologian means to have experience of a personal encounter with God through prayer and worship."

Posts 570
Schumitinu | Forum Activity | Replied: Fri, Jul 3 2015 8:44 PM

William Gabriel:

Andrew Batishko (Faithlife):

 It sounds to me that your desired hierarchical structure is different than that chosen by the people that put together the BSL. They were attempting to represent certain structures while you are desiring different ones. This is a really common problem with any kind of semantic hierarchy.

I believe the semantic tree that was created has a good structure. I think we need another dimension, another tree that organizes along senses that make sense to group together. So you might have a new branch "Pray" that points to all the leaves that are in disparate branches of the current tree. You may have a new branch "House" that points to all the varied uses of House, even though they might be radically different in semantic meaning.

I know this would be another large undertaking, both in coding and tagging, but I think average Bible study users (raising my hand here) would get tremendous benefit from having datasets that track with common thinking.

The other dimension that we are missing is what is called "contextual domains". What the BSL provides and FL has done an excellent job with are "lexical domains". That is why 'prayer (an act)' and 'to pray' are separated. One is looking at nouns the other at verbs, which has been kept separated. Now, what we are looking for are another set of labels that could be added to the lexical labels which would group these senses together as concepts. In our case "Prayer". I have already suggested a framework for this (see for example for prayer: http://semdom.org/v4/4.9.5.2).

I wasn't aware that the Lexham Theological Wordbook summarizes the senses for that concept. That is already a step in the right direction. And I also found the same to be true in the cultural concept sections. Very helpful. This however cannot be searched easily. And I think that is our main point. How to make these concepts searchable. Would one way to solve this be to add a search query to those entries in the Lexham Theological Wordbook and the Cultural Concept section?

The search ability that has been suggested (CONTAINS) might work too. Whatever FL comes up with I whish that the search can be used in Visual Filters. Please consider that.

Correction: the cultural concept sections do not contain the word senses. This is a separate section in the Factbook.

Posts 26988
Forum MVP
MJ. Smith | Forum Activity | Replied: Fri, Jul 3 2015 8:54 PM

Please read the Tip of the day for a summary of what each of a number of these lexical tools are doing. I'm still not sure that we have a solid definition of what we want unless we want fruit punch ... a mixture of apples, oranges, currants and durians.Wink

Orthodox Bishop Hilarion Alfeyev: "To be a theologian means to have experience of a personal encounter with God through prayer and worship."

Page 2 of 2 (34 items) < Previous 1 2 | RSS