Saturday, March 24, 2018

Finding Hebrew words translated by a Greek word in the LXX: Accordance, BibleWorks, Logos

A question arose on the Accordance forum about finding the underlying Hebrew word/s which were rendered with a given Greek word. It got me to thinking about how I would accomplish this task in the major Bible software programs: Accordance, BibleWorks, and Logos. For this exercise, I am just going to see what Hebrew words are being rendered with a form of the verb ἱλάσκομαι in the LXX. It's a word that has a range of meaning including "be merciful, propitiate, expiate, make atonement." Cognate forms are important in the NT, since they are used to describe Jesus. (E.g., ἱλαστήριον in Rom 3.25 and ἱλασμός in 1 John 2.2.) In most cases, the software is making use of some edition of Emmanuel Tov's (and Frank Polack's) Parallel Aligned Hebrew-Aramaic and Greek Texts of Jewish Scripture. (The only exception might be Logos' Lexham Greek-English Interlinear Septuagint.)


I'm still learning Accordance, so others may correct me or suggest a better way, but this is what I think you need to do. 1) Open the LXX; 2) Search for ἱλάσκομαι; 3) Open a new tab with the MT-LXX Parallel; 4) Use the MERGE command to link this tab with the LXX tab. As the graphic above shows, the Greek word is highlighted, but a person would need to go through all the hits and pick out the Hebrew. 
UPDATE: I had initially lamented there was no way to see the Hebrew roots in Accordacnce. Thanks to Joel Brown from Accordance, I need to update this entry, because there is an easy way to see the Hebrew roots underlying the Greek.
The trick is to use the Analysis Pie Chart and choose MTLXX Lex or to use the standard Analysis and customize the display by dragging in the MTLXX Lex category. The results are visible above. Since Accordance can also do Greek root searches, you can get a broader picture of usage quickly. It's also quite easy to copy/paste the results (here displaying part of the root search results):

ἱλάσκομαι (ἵλεως) to be merciful, pardon, be propitious = 12
            אתה סלח = 1
            כפר = 3
            לְ סלח = 1
            נחם = 1
            סלח = 4
            (No direct translation) = 2
ἱλασμός (ἵλεως) propitiation, means of forgiveness = 6
            כִּפֻּרִים = 2
            סְלִיחָה = 1
            (No direct translation) = 3
ἱλαστήριον (ἵλεως) propitiation, place of forgiveness = 28
            הַ כַּפֹּרֶת = 2
            כַּפֹּרֶת = 17
            כַּפְתּוֹר־1 = 1
            עֲזָרָה = 5
            (No direct translation) = 3

This is the ideal way for getting the results desired in the initiating question.
(As an aside, both Accordance and BibleWorks found ἐξιλάσκομαι in Deut 21.8 as a hit for ἱλάσκομαι. That appears to be an incorrect morph tagging that Logos corrects.)
Accordance can perform some other nifty tricks by combining MERGEs in order to see where a specific Hebrew word is translated by a specific Greek word. Cf. David Lang's guide HERE.


BibleWorks has a Parallel Hebrew and LXX module that makes good use of the Tov alignment. After opening the module (using the icon in the button bar or Resources > Parallel Hebrew and LXX), a window opens and one can the select Search > Search for Hebrew-LXX Equivalents (or click on the א=α binocular icon). This opens the smaller window shown in the graphic above, and one can type in either a Hebrew lemma and find all the ways it's translated into Greek or a Greek lemma and find all the Hebrew words it's translating. The results need a bit of checking due to the way Tov had to correlate the Hebrew and Greek. (Mainly it's separating out Hebrew prefixes like ל or ו.) By clicking on any given result, you can see that it calls up the verses where it occurs. I can see the actual Hebrew and Greek texts on top, and I can view lexical information at the bottom. If I want to see an English translation of a result verse, double-clicking will open it up in BW's Browse Window. From my results window, I also have the option of selecting and copying the results, so that I can generate this:
1   ~xn     ilaskomai
4 rpk    ilaskomai
6 xls    ilaskomai

One drawback of BW is that it cannot do Greek root searches, so if I wanted to look at cognate forms of the Greek word, I would need to search each one separately.


Logos offers the powerful options for getting the results I want, and I can get at it in a couple ways. I can start with a simple lemma search of the LXX, and when the results are returned, I can choose the Verses view to compare any number of translations. As shown above, I have my LXX hits highlighted, I've added the Hebrew, and I've also the NRSV which has the sympathetic highlighting. (UPDATE: In the comments, Mark Barnes noted that by choosing the Lexham Hebrew Bible to display, you can also see the Hebrew highlighted.) In that results window, I can also choose the analysis view, and then I get this (and this is only a partial view):
I have chosen to organize the analyis by the Hebrew lemma, and I'm provided with a clear view Greek and Hebrew lemmas and a full analysis of each. That is extremely helpful, but... I cannot find an easy way to copy/paste out the information. Logos does allow root searches, so I could also generate results like this for all cognates.
But there is another way to answer my original question. I can do a lemma search for ilaskomai, right click on one of the results, select the lemma, and then select Bible Word Study. This generates a lot of information, but what I want to see is this:
Here is a clear illustration of the Hebrew words underlying the Greek, and clicking on any one of the segments will display all the places where it occurs. Again, extremely helpful, and I can right-click to copy the graphic, but I dont see any way to generate a simple listing. 
UPDATE: Mark Barnes in the comments noted workarounds for getting the results exported using Excel or using the Bible Word Study. (Read his explanation HERE.) He also noted an even easier way to get the results most quickly. Use the Lexham Analytical Lexicon of the Septugint. It's a simple matter of looking up the word which displays results like this:
There you have it. A few things to note. 
  • When it says the סלה is the word 4x, it actually is referring to the number of verses, not the number of times. (The word occurs twice in 4Kgs (2Kings) 5.18, so it's actually 5x.) 
  • The English Gloss is there because those are instances where the LXX differs from the MT, so there is no underlying Hebrew. 
  • Fine print! If you've been counting, the results displayed miss 2Chron 6.30 which showed up in my other searches. Why? It's a text critical issue. In that verse, Rahlf's edition of the LXX has the word ἱλάσῃ from ἱλάσκομαι. In Swete's edition of the LXX, however, the word is ἰάσῃ, apparently from ἰάομαι. I checked a print copy of Swete, and that's indeed what he has. Is it a typographical error in the print edition? I think the sense is: "You will hear from heaven... and you will heal (instead of expiate) and give to a man according to his ways." I.e, the Lexham Analytical Lexicon apparently is using Swete rather than Rahlf's. Oddly, the Lexham English Septuagint tries to render ἰάσῃ as a name, Jahaziel.


Accordance probably does the best job of addressing the task of finding all the underlying Hebrew words translated by a specific Greek word in the LXX. It also the advantage of being able to do a Greek root search for a broader scope. Logos is a very capable program, and using the Lexham Analytical Lexicon of the Septuagint is the easiest way of all. It also offers the most presentation and hyperlinked options. BibleWorks does accomplish the task in its own way and can generate the list of Hebrew words simply. I certainly don't think a person is going to buy one of these programs simply based on this exercise, but I hope it gives an idea of what each is capable.

To see another example of how to use the MT-LXX parallel to conduct other tasks, you'll want to read this earlier post where I walk through some examples. 

If I've missed some better or easier way to address the originating question, please let me know in the comments.


  1. The Accordance associate writes, "However, by [MERGE]ing in results from a text, Accordance is able to blend the inflected forms and locations and give pretty accurate results."

    As an academic, "pretty good" results are under par. I think your post shows which is the more academically qualified application to use for this type of research.

  2. Thanks, Mark. It's interesting to see the different approaches in the different applications.

    Two things to add.

    First, in your first Logos screenshot, if you use the Lexham Hebrew Bible (rather than the BHS), then the Hebrew words will be highlighted alongside the Greek.

    Second, in Logos, it's possible to get the list of Hebrew lemmas via an Excel export. Whilst not that simple, it is very powerful. But even easier is to use the Lexham Analytical Lexicon of the Septuagint.

    As space is limited here, and I can't add images, I've put a bit more info on a post on the Logos forums.

    1. Thanks, Mark! You'll see that I updated my post.

  3. Thanks for the post Mark! I thought I'd point you to a response I put together this evening over here using parabible (a website I built using open(ish) data). I'd love to hear your thoughts!

    1. Thanks, James. I'm not familiar with Parabible. From looking at your post, it appears that it does a great job of matching the Greek lexeme used for translating a Hebrew word. Can it, however, find all the Hebrew lexemes behind a specific Greek word?

    2. Right now you can't summarise the results but you can search only on the "parallel lexeme" which will find matches regardless of the Hebrew lexeme.

      If you had to wade through a large number of the standard option and so were trying to filter out a particular Hebrew root, you could add that root as another search term and "invert" it (in the dropdown) - inverting a search term only gives results that don't have that term.

  4. Mark, Accordance has an Analytics tool that lets you sort and view results in various ways. You can even create a pie graph, similar to the one Logos has.

    1. Thanks, A.D.! I couldn't figure out how to accomplish what you say is possible... until Joel Brown posted on the Accordance forum. You'll see that I updated my post above, and Accordance moved to the top of my list for accomplishing the task at hand.