compare LXX translation to Hebrew source

[Martin K]

I got some good help from this forum last time, so let me try another one:
Does anyone know how to do a search that will show me all the instances where the LXX has ἐκκλησία for the Hebrew קָהָל.

[Mark Eddy]

You do this kind of search in the Graphic Search Engine (GSE). You need to have two Word boxes: 1) search version BGM (or BLM) search for EKKLHSIA, 2) search version WTM search for QHL. (Both will display in the correct font in the GSE.) Open a Merge box and do an "And" search, link each Word box to the Merge box, and hit "go." This will give a list of verses in which both the Greek EKKLHSIA and the Hebrew QHL are present. This does not guarantee that in each case the Greek word is a translation of the Hebrew word. You have to double check that yourself.
Or, if you are in a hurry, you can open the "Parallel Hebrew/LXX" (the button with a Greek alpha to the left of a Hebrew aleph). Under "Search" choose "Find Hebrew - LXX Equivalents." Type in the "Hebrew Lemma" QHL and "Greek Lemma" EKKLHSIA, and click on the binoculars button. It gives you a list of 70 verses. If you do it this way, you are taking Drs. Tov & Polak's word that these are the only instances. They are usually pretty good, but there is room for argument in some of the more loosely-translated passages in the LXX. You can check them out by doing the GSE search.
I hope this is what you are looking for.
Mark Eddy

[Dale A. Brueggemann]

Does anyone know how to do a search that will show me all the instances where the LXX has ἐκκλησία for the Hebrew קָהָל.

1. Resources menu
2. Parallel Hebrew = LXX
3. Find Hebrew - LXX equivalents (3rd icon on menu bar)
4. Type in ἐκκλησία and קָהָל in the Greek and Hebrew search boxes.
5. It gives you 70 hits, which you can then browse.

[Martin K]

Thanks guys. Appreciated your help.

[Van Parunak]

If you're using the parallel to help explain the NT (an excellent thing to do!), here's another suggestion. Establishing the MT-LXX correspondence can help show the MT background of a NT word, but the argument depends on overall patterns of usage, and the LXX often uses a Greek word once or twice for a given Hebrew word, when in fact the "normal" Greek equivalent is a different word. Or the Greek word may be used for multiple Hebrew words. I find it helpful to ask whether the Greek word in question is both dominant and dedicated for the Hebrew word. Dominant means it's the word most often used in the LXX for the Hebrew word; dedicated means that it's used mostly for this specific Hebrew word rather than being used for others as well.

So while you're in the parallel Hebrew-LXX tool, you might check for other translations of קהל and other Hebrew antecedents of εκκλησια. You do this by listing only the Hebrew word in the Hebrew lemma box in the "search for equivalents" box, leaving the Greek lemma box empty, and finding all the Greek translations, then listing only the Greek lemma, leaving the Hebrew lemma box empty, and finding all the underlying Hebrew terms. As you do this, you may discover other Greek words and other Hebrew words that you need to check. I usually make a table to help me keep track of what's going on.

You'll find lots of chaff in this process. The LXX translation sometimes doesn't align well with the MT, whether because it's following a different Hebrew text, or because the translator didn't know a given Hebrew word or the syntax of the context in which it occurs, or because Tov and Polak got it wrong. Sometimes the chaff includes cognates of either the Hebrew or the Greek word, which you may want to lump with the word you're studying. But if you focus on the most common matches in both directions, you might find something interesting.

This is one case where this happens. Here's a summary of the major correspondences for your pair:

	קהל	עדה	Total Occurrences
εκκλησια	70	0	77
συναγωγη	34	130	199
Total Occurrences	162	172

"Total occurrences" is what a search in Bibleworks returns for WTM and BLM (restricted to OT to avoid the deuterocanonicals, for which Hebrew is only spottily available). It gives a quick idea of how big the counts are compared with the total number of instances.

In this case, εκκλησια is very dedicated to קהל, but it's not very dominant. The LXX felt comfortable in many cases using συναγωγη to translate קהל. So we have to check out συναγωγη as well. When we do, we find that it's used even more often for עדה; it is somewhat dominant for עדה, though not completely dedicated (as its uses for קהל show). Note that the relations are inverse if you start from the other language, and not entirely symmetrical: עדה is dedicated to συναγωγη but not entirely dominant; קהל is weakly dominant for εκκλησια but not very dedicated.

This picture invites a more nuanced understanding of how the LXX aligns with the MT, and how the NT uses them. It's only the starting point: for example, one should go on to explore how the difference in translations may reflect different contexts or idioms in which the Hebrew terms are involved. Is there something special about the קהל-συναγωγη instances that sets them apart from the other two pairings? It might be as simple as the books in which the pairings are used, reflecting different translation conventions. This kind of result can add special depth to the implications of a NT writer's user of the word. (For example, if the difference correlates with book, or with a specific idiom, it suggests possible allusions by the NT writer.) But even at this raw statistical level, it cautions against (in this case) asserting a strict equivalence between εκκλησια and קהל.

If you'd like to see an example where this analysis makes a difference, take a look at http://www.cyber-chapel.org/AtonementInTheNT.pdf . Dominance-dedication analysis shows pretty convincingly (at least to me!) that the use of "atonement" in NT translations (e.g., the AV in Rom 5:11, the NIV in Rom 3:25 and Heb 2:17; 9:5; and related terms in 1 John 2:2; 4:10), sometimes based on casual observations about LXX-MT vocabulary correspondences, is unjustified.

Kudos for the BW team for this excellent tool. Doing this kind of analysis before the tool came out required hours with Hatch-Redpath and the Dos Santos index; now it's almost instantaneous, and the results can give unparalleled insights.

[ISalzman]

If you're using the parallel to help explain the NT (an excellent thing to do!), here's another suggestion. Establishing the MT-LXX correspondence can help show the MT background of a NT word, but the argument depends on overall patterns of usage, and the LXX often uses a Greek word once or twice for a given Hebrew word, when in fact the "normal" Greek equivalent is a different word. Or the Greek word may be used for multiple Hebrew words. I find it helpful to ask whether the Greek word in question is both dominant and dedicated for the Hebrew word. Dominant means it's the word most often used in the LXX for the Hebrew word; dedicated means that it's used mostly for this specific Hebrew word rather than being used for others as well.

So while you're in the parallel Hebrew-LXX tool, you might check for other translations of קהל and other Hebrew antecedents of εκκλησια. You do this by listing only the Hebrew word in the Hebrew lemma box in the "search for equivalents" box, leaving the Greek lemma box empty, and finding all the Greek translations, then listing only the Greek lemma, leaving the Hebrew lemma box empty, and finding all the underlying Hebrew terms. As you do this, you may discover other Greek words and other Hebrew words that you need to check. I usually make a table to help me keep track of what's going on.

You'll find lots of chaff in this process. The LXX translation sometimes doesn't align well with the MT, whether because it's following a different Hebrew text, or because the translator didn't know a given Hebrew word or the syntax of the context in which it occurs, or because Tov and Polak got it wrong. Sometimes the chaff includes cognates of either the Hebrew or the Greek word, which you may want to lump with the word you're studying. But if you focus on the most common matches in both directions, you might find something interesting.

This is one case where this happens. Here's a summary of the major correspondences for your pair:

	קהל	עדה	Total Occurrences
εκκλησια	70	0	77
συναγωγη	34	130	199
Total Occurrences	162	172

"Total occurrences" is what a search in Bibleworks returns for WTM and BLM (restricted to OT to avoid the deuterocanonicals, for which Hebrew is only spottily available). It gives a quick idea of how big the counts are compared with the total number of instances.

In this case, εκκλησια is very dedicated to קהל, but it's not very dominant. The LXX felt comfortable in many cases using συναγωγη to translate קהל. So we have to check out συναγωγη as well. When we do, we find that it's used even more often for עדה; it is somewhat dominant for עדה, though not completely dedicated (as its uses for קהל show). Note that the relations are inverse if you start from the other language, and not entirely symmetrical: עדה is dedicated to συναγωγη but not entirely dominant; קהל is weakly dominant for εκκλησια but not very dedicated.

This picture invites a more nuanced understanding of how the LXX aligns with the MT, and how the NT uses them. It's only the starting point: for example, one should go on to explore how the difference in translations may reflect different contexts or idioms in which the Hebrew terms are involved. Is there something special about the קהל-συναγωγη instances that sets them apart from the other two pairings? It might be as simple as the books in which the pairings are used, reflecting different translation conventions. This kind of result can add special depth to the implications of a NT writer's user of the word. (For example, if the difference correlates with book, or with a specific idiom, it suggests possible allusions by the NT writer.) But even at this raw statistical level, it cautions against (in this case) asserting a strict equivalence between εκκλησια and קהל.

If you'd like to see an example where this analysis makes a difference, take a look at http://www.cyber-chapel.org/AtonementInTheNT.pdf . Dominance-dedication analysis shows pretty convincingly (at least to me!) that the use of "atonement" in NT translations (e.g., the AV in Rom 5:11, the NIV in Rom 3:25 and Heb 2:17; 9:5; and related terms in 1 John 2:2; 4:10), sometimes based on casual observations about LXX-MT vocabulary correspondences, is unjustified.

Kudos for the BW team for this excellent tool. Doing this kind of analysis before the tool came out required hours with Hatch-Redpath and the Dos Santos index; now it's almost instantaneous, and the results can give unparalleled insights.

Thanks for the excellent and very interesting post, Van.

[DavidR]

I add my thanks for this excellent tutorial, Van. It is so important to take context and linguistic realities into account when considering the use of particular biblical words. BW and other contemporary tools can make a "quick" word study very quick indeed, but it's easy to get misled by the simplicity and speed, and neglect to do the kind of detailed research that you demonstrate. Kudos!

[Dale A. Brueggemann]

If you're using the parallel to help explain the NT (an excellent thing to do!), here's another suggestion....

I do this frequently; however, I've never thought of just making a matrix-table to compile results when I do (fur dumm! ).

[Van Parunak]

Since this approach seems to be of interest, some folk may be interested in a way to render it more quantitative. Here is how to estimate the statistical significance of a difference between two sets of correspondences.

For the record, in the question of translations of קהל, this analysis suggests that the difference between 70 instances of εκκλησια and 34 instances of συναγωγη is is in fact statistically significant. But I'd be cautious about rendering exegetical decisions solely, or even mainly, on the basis of numbers. 34 instances of συναγωγη, while less than 20% of the total usage of that lexeme in the MT (compared with 91% for εκκλησια), is hardly trivial, and I'd focus scarce time resources first on trying to understand contextual features that might explain why we have two different Greek translations for קהל. So I don't usually do the computations outlined in the thesis--I lay out the matrix, and use it to guide deeper engagement with the text. But if you do want a statistical assessment, it's available.

(The link is to the first half of the appendix to my thesis. I didn't post the second half, since it has been supplanted by a later publication [1].)

[1] H. V. D. Parunak. "Prolegomena to Pictorial Concordances." Computers and the Humanities, 15:15-36, 1981. http://www.cyber-chapel.org/LinguisticDensityPlots.pdf