Jump to content

Stop treating prefixes and suffixes as separate words in Hebrew


TYA

Recommended Posts

Please bear with my explanation here as I don't have Acc in front of me to do convenient screenshots.

 

The problem is the scope of the search. This is represented by the icon of the little paragraph to the right of the search box before the - in the circle. Check my screenshot and you will see that my icon there is slightly different from yours. Click once or twice on you icon there to change it so it matches mine and rerun the search.

 

Thx

D

Link to comment
Share on other sites

I think it is important to recall that every single search is a totally new search, will "undo" your previous search, and will therefore, always start you at the top of the document. This is normal Accordance behavior.

 

There isn't a "two step" search process other than the one that Daniel and I screenshotted - that is, using two search elements together.

 

As you noted, Accordance has stated they intend to add a find on page search.

 

However... I'll also add, that if you can get your head around this new way of search you'll be really pleased, I think, at how much more flexible and powerful it is.

Edited by jarcher
Link to comment
Share on other sites

Dear Daniel, what am I doing wrong?  Your process seems like it would accomplish what I'm looking for, but when I do what you say, it still throws me up to the top of the document (i.e. first occurrence of "common").

 

TYA, are you sure you didn't use an "OR" search in this screenshot? I'm not able to reproduce this result until I change the condition to OR instead of AND.

Link to comment
Share on other sites

Daniel, success!  You were right: I had to adjust that icon to "Search within each main article," and it then gave me what I was looking for--viz., the next occurrence of "common" in that part of the document.  Thank you for helping.

 

Jarcher, thank you also for the good suggestions.

 

To your remark about me wrapping my head around this process, let me humbly say, I couldn't fathom the benefit of *not* being able to drill down a search.  Thankfully, Daniel has provided a tentative way to do this for now, but the "Find on Page" function will be even quicker and more effective in Accordance.

 

One of the things that drew me to Accordance the most was the large (and growing) library of content.  Being able to drill down through a two-stage process (at least in regards to English searches, although I've already lodged my request / complaint in this thread regarding Hebrew searches, which still remains unresolved), is absolutely necessary.

 

Other programs I use that have the "find on page" utility are extremely valuable and efficient in this regard.  But thank you again, and thanks to all.

 

It is still the Hebrew search (for which I started this thread) that I'm most concerned about.  Would someone please confirm if my concern is true--that Accordance will somehow require double the work to find a Hebrew word, simply dependent on whether that Hebrew word happens to be in a text that is tagged versus untagged text.

 

Again, I'm fully aware that different software's do things *differently*.  That's clearly not the concern here.  I'm only concerned if Accordance has no way of accomplishing (what I think are) fairly basic tasks common to lexicographic research, in a way even close to the efficiency available in BW.

 

Based on what Accordance' Support has confirmed during my many email exchanges, I'm not missing something simple, and I just want to say in regards to why I've opened so many Feature Requests after coming over to Accordance...

 

I can 1) search Hebrew / Aramaic texts with more across-the-board consistency (reliability, per se) in BW, and 2) then review those search results within original language versions roughly *5 to 10 times* faster than in Accordance.

 

I could understand if it was maybe twice as fast in BW, but without exaggeration, for me to work through a list of 100 search results, analyzing them quickly in various Bible versions, when it comes to the number of clicks--the number of windows I have to open, and the need (in Acc.) to repeat the exact same process over and over again, I would have to spend roughly 5 to 10 times more in Acc. to accomplish the same thing.

 

I wouldn't gripe if this was something I only do every now and then, but I perform lexicographic analyses like this almost every week, sometimes requiring several hours because of the numerous versions / languages that I want to dig into, cross-reference, etc.  Sorry if it sounds too much like kvetching, but my sincere hope is that I'll find a better way to bridge the gap in Accordance.

 

Shalom all

post-35231-0-66416700-1536341259_thumb.jpg

Edited by TYA
Link to comment
Share on other sites

Someone confirm that my fear isn't true--that Accordance isn't going to make me have to double my work to find a Hebrew word, simply dependant on whether that Hebrew word happens to be in a text that is tagged versus untagged.

 

Can you dive into this a bit more on exactly what you mean by search results being different in tagged vs. untagged texts (which I think you mentioned previously as well)? I'm sure many of us are happy to help as we're able.

Link to comment
Share on other sites

Can you dive into this a bit more on exactly what you mean by search results being different in tagged vs. untagged texts (which I think you mentioned previously as well)? I'm sure many of us are happy to help as we're able.

 

Technically, no, because I as I said, it is only a "concern."  That concern actually arose from the Help menu when I was trying to follow Helen's suggestion to do an inflected search instead (something I still can't figure out how to do, and would appreciate help).  That Help page caught my attention when it said...

 

To find an inflected form in a tagged text, enclose it in plain quotation marks.

 

This immediately caused me to think that Accordance is implying that it has different search processes for tagged texts versus untagged texts.  I was asking if someone could confirm whether this was true or not, because I'm already frustrated as it is.  I don't want another layer of frustration.

 

All that being said, I want to return to what I can clearly and definitively point to as my frustration / challenge.  Please see the two screenshots attached.  Thank you

post-35231-0-00616800-1536346694_thumb.jpg

post-35231-0-02514800-1536346699_thumb.jpg

Edited by TYA
Link to comment
Share on other sites

TYA, in the second screenshot I cannot see the content behind the instant details box. What was searched for exactly there?

Link to comment
Share on other sites

Nothing.  I'm not sure how that string of characters got in there, but the left window pane was the search.

Link to comment
Share on other sites

Nothing.  I'm not sure how that string of characters got in there, but the left window pane was the search.

 

Maybe that is part of the issue. You must have search criteria in a research search pane.

Link to comment
Share on other sites

Accordance does indeed use different kinds of searches in tagged texts. In tagged texts you can search for the lexeme (no symbols) , the root (+ [plus sign]), or the inflected form (character string) (double quotes before and after the string). To find a specific conjugation or declension, search for a lexeme and use a grammatical tag. Accordance can only search untagged texts for the character string, since tagging information isn't available for the other two kinds of searches.

 

It sounds to me like you are trying to use Accordance without using any of our training materials. That is going to be frustrating, as you will keep trying to search as you once did in BW. They are two different programs. They don't work the same way. Try this podcast for a basic introduction to Hebrew searching: http://www.accordancefiles2.com/podcasts/p147_basichebrewsearches.mp4

  • Like 2
Link to comment
Share on other sites

Dear Dr. J,

 

Yes, I indeed need the training, and am excited to look further into that.  But one major problem was that I wasn't able to figure out how to do an inflected search, even using the Accordance Help menu.  However, jarcher very generously walked me through doing that, and I was able to perform this gratifying inflected search (attached).  I say "gratifying" because it returned search results with both prefixes and affixes of various types, across a more broad spectrum of literature than I've accomplished up to this point, AND including texts that are both tagged (e.g. BHS) and untagged (e.g. Delitzsch and Modern Hebrew NT).

 

Help me understand, if you say: "Accordance can only search untagged texts for the character string, since tagging information isn't available for the other two kinds of searches," then why does this search (attached) show texts that are both tagged and untagged?  (And of course, I obviously want the search to function that way).

post-35231-0-59423600-1536359194_thumb.jpg

  • Like 2
Link to comment
Share on other sites

Hi TYA,

 

  I was going to suggest you try the search this way but Jeremy beat me to it. Good to see it works.

 

  I'm pretty sure the reasoning is as follows but others may correct me. I don't work for Oaktree and the below is based on observation.

 

  The thing to understand here is the default handling of the search in different cases. If tagging (this term has a variety of meanings actually but suffice to say here it means, the lexeme for the inflected word is known) is present then unquoted searches search on the basis of the entered word being a lexeme. (You will notice searching say HMT-W4 or BHS that if you do a word search you will see a pull down list of possible lexemes as you type you search word.) If you quote the search word then Accordance assumes you know you want that specific inflected form of the word and searches for that regardless of the text having tagging or not. Now for texts and modules which do not have tagging the only choice is that the search is for what you enter.

 

  Research queries have the difficulty that they can go across tagged and untagged modules, but they really just run a bunch of regular searches over the modules in the Research group you are searching. (At least that is my assumption about how they actually function. I haven't seen the code.) As a consequence if you do not quote the string some of the searches may be run as though they were looking for lexemes rather than inflected words and others may not. Where the string you entered is a lexeme that's all good and you get hits. Where it is not you won't. This is why the unquoted results hit entries in the untagged texts only. Likewise lexica and other tools are not tagged. So they also show hits in unquoted searches.

 

Thx

D

Edited by דָנִיאֶל
  • Like 1
Link to comment
Share on other sites

Thank you Daniel, and thank you all for this extensive help.  I don't fully wrap my ahead around all of it (as Daniel said, not knowing the code, or exact process), but you've all helped point me in the right direction.  Thanks again for that.

Link to comment
Share on other sites

  • 11 months later...

But as Matt said, they are not prefixes as we talk about them in English. They are words. Separate.

I'd be curious for you to give the example of why you want to do a literal search that spans words.

 

They are treated as separate because they are. Enclitic pronouns and suffixes are stand alone words.

 

I know this is digging up old conversation from when I was brand new to Accordance (1 year ago; Yay! 1-year anniversary everybody!), but I could never understand why someone would consider "bet," "kaf," and "lamed" prepositions in Hebrew to be separate words, unless one means how they are translated into English, like "in the house," "like a king," or "to the city."

 

But in Hebrew they are prefixed to the rest of the word, leaving no whitespaces between them and the primary lemma.  That's why I contended at the beginning of this topic that, no, they aren't separate words, functioning all by themselves.

 

Now, just for fun (or perhaps, healthy debate), I thought I'd post this lesson I came across in Ross Hebrew Grammar today, which asserts that these prepositions aren't separate words, but calls them "inseparable prepositions," and Ross contrasts them with other prepositions that do function as individual words.

 

3.4 Inseparable Prepositions

 
Many Hebrew prepositions are individual words and function in a sentence as English prepositions do. These offer little difficulty to beginning students, because they are treated as individual words. But the prepositions ‏בְּ‎ (in, with, by), ‏כְּ‎ (like, as), and ‏לְ (to, for) are always prefixed directly to the word they govern. Hence, they are called inseparable prepositions. The normal pointing for these forms is simple vocal shewa. (Ross Hebrew Grammar; Pt. 1, Lesson 3)
Edited by TYA
  • Like 1
Link to comment
Share on other sites

For what it's worth, the notion of "inseparable prepositions" isn't really accurate. For a morpho-syntactic perspective, these Hebrew prepositions are clitics since they attach to a variety of word types (nouns, adjectives, participles, even verbs).

 

As for the larger discussion, the notion of "word" is terribly complicated. A word in one language is not the same as the other, and so on. Since BH allows a lot of morpheme fusion (i.e., there are inflectional affixes, prepositional clitics, cliticized pronouns, the clitic article, and so on), the simplest notion of word in BH is the graphic word, i.e,. what's written "without white spaces". But since writing is itself an extra-linguistic, analytical process, graphic word is hardly sufficient. So that leads to prosodic word, which is the items between sound breaks and so forth. But we can't hear BH anymore. So both graphic word and prosodic word are not BH phenomena, but medieval Masoretic phenomena. B. Elan Dresher has written a nice article about this called "The Word in Tiberian Hebrew". Below is a dropbox link to the PDF (hope I'm not transgressing forum rules!)

 

https://www.dropbox.com/s/8yscbvmkspitemm/Dresher_WordTibHebrew_2009.pdf?dl=0

 

Anyway, hopefully the point of my rambling is clear: any notion of "word" has problems when applied to our beloved BH corpus. As long as we get used to one particular convenience (and this thread seems to have been about the clash of two different conventions in different software), we simply have to be ok with it as long as the convenience allows us to access the data.

 

The minor point of it is that the way that introductory textbooks (as well as most of the reference grammars) describe BH linguistic phenomena is greatly lacking in accuracy. But if they weren't, I wouldn't have a job I guess.

Edited by Robert Holmstedt
  • Like 2
Link to comment
Share on other sites

For what it's worth

 

It is greatly appreciated.  Yes, I realized early on that our main contention at the beginning of this thread was over the meaning of "word."  The whole fact that such a concept is nebulous, as you say, is one reason that I have been pushing for the "literal search between the whitespaces" on this forum over the last year.

 

While I appreciate the existing "structures" in the various syntax modules, I don't want anyone dictating to me what they think a "word" is when I choose to search outside those existing bounds--especially across multiple texts in the Research Tab.  Again, I'm not for detracting from existing functionality--only adding to it.

 

Robert, your expertise, and willingness to share it here makes me proud to be an Accordance user, and proud to be a fellow "forum ranter," if I can put it that way :)

Edited by TYA
  • Like 1
Link to comment
Share on other sites

+1 for letters search with other criteria such as commands or symbols

+1 for letters search with other criteria that allows for finding word plays (e.g. finding Gen 6:8 נח...חן) (maybe something like a construct search for just letters? AGREE in letters?)

 

Thanks!

Link to comment
Share on other sites

  • 1 year later...

I came upon this thread looking for more information about the Hebrew “Letters” search. I was actually pondering the question of how to define a Hebrew word for notes that I am writing for a class on Accordance Bible software. I did not take the time to define a Hebrew word but just skirted around the issue by assuming “string of letters bounded by spaces” to be the definition of a Hebrew “word” and by trying to be more specific when discussing well, when discussing “words”, such as by saying: “the noun in the word וּבַחֲמָתְךָ is parsed with the following tags:”. At the time, I had no idea where to go look for discussion about what is a Hebrew word and the grammars I had access to did not touch upon it anywhere. So thanks for sharing, Robert.

It seems to me that the originally requested feature should not be something too difficult to code for. The “Letters” search already exists in tagged Hebrew texts and a search without quotation marks in an untagged Hebrew text does the same job. If you could somehow select for a “Letters” search in the Research tab, it could search for “Letters” in tagged texts and run a basic search without quotation marks or equal symbols in untagged Hebrew texts. Since the “Letters” search option has already been developed for the tagged texts I don't think it would require an entire re-hauling of Accordance to accomplish, and the original question shows that this would have its use.

That said, the “Letters” search seems to be run a pretty complicated search. I say that because it takes much longer than other types of searches in a tagged Hebrew text. 

Link to comment
Share on other sites

If you could somehow select for a “Letters” search in the Research tab, it could search for “Letters” in tagged texts and run a basic search without quotation marks or equal symbols in untagged Hebrew texts. Since the “Letters” search option has already been developed for the tagged texts I don't think it would require an entire re-hauling of Accordance to accomplish, and the original question shows that this would have its use.

 

I still hold to my claims / complaints. Just saying... :)

Edited by TYA
  • Like 1
Link to comment
Share on other sites

  • 3 weeks later...

I just want to point out that if, in the “Reading/Research” category of the preferences, you set the “Hebrew Searching:” option to “Literal Search”, you can run a “research” across tagged and untagged texts for a string of consonants and Accordance will find results in both grammatically tagged and untagged Hebrew texts. For some reason, I thought that this was not possible when I first read through this thread, but it is.

What is still not possible is what you pointed out in your first post: the literal search does not allow the use of search symbols.

Link to comment
Share on other sites

Thanks Iconoclaste, but I'll still point out that, search symbols (like wildcards) aside, the white spaces between letters still aren't being taken into account in the Research Tab--at least not in tagged texts.  See this screenshot showing that I get the same results in certain modules between מ עצת and מעצת.  The space in the first of these makes absolutely no difference in the results I get for the tagged texts (BHS, WMT-4, Qumran), even setting the Preferences to "Literal," as you suggested.

 

By the way, I went back and read my original posts on this thread from 2 years ago, and I just want to apologize to anyone if I came across as rude.  Accordance really is a terrific company, and the people who post here are absolutely wonderful--a pleasure to be in contact with.  I was just very frustrated with what I was discovering in Accordance at that time.  Don't get me wrong: I'm no happier about the situation, but I'm just not as caught off guard now, if that is okay to say.

post-35231-0-81846600-1613116970_thumb.jpg

Edited by TYA
Link to comment
Share on other sites

Please sign in to comment

You will be able to leave a comment after signing in



Sign In Now
×
×
  • Create New...