This is the first of a series of posts in which I am going to put dictionaries and databases of ancient Hebrew literature in electronic format to the test. I review ten electronic dictionaries which include discussion of biblical Hebrew vocabulary items, and describe what it’s like to use them alongside other tools, electronic and otherwise, in the course of an exploration of the statistics, morphology, and semantics of the verb אכל ‘to eat.’
In the study of ancient literature, few things matter more than how useful and accurate tools like lexica, grammars, and concordances are. We might especially expect accuracy in reference works that relate to the Bible, a book, after all, pored over with great care by many. To the extent that reference works fail to describe textual details helpfully and accurately, I will pound the point home. If the sight of blood disturbs you, turn back now.
My comments relate in almost equal measure to print and electronic versions of resources where available in both. The discussion will be of interest to anyone who loves the fine detail of the language and literature of the Hebrew Bible, regardless of whether one plans to acquire electronic tools for research purposes or not. If nothing else, my discussion of specific passages may prove interesting. Referenced resources are referred to by author, acronym, or short title. Full bibliographical data are supplied in a downloadable file. In this post, I concentrate on the statistics of אכל. Statistics, properly understood, reveal more than is sometimes thought.
My first-hand experience with electronic dictionaries of biblical Hebrew is limited to those available through Logos. Except for eHALOT, they form part of the Scholar’s Library.
If you’re like me, you often have a dozen books open to various places in the midst of a hot and heavy study of a text or an item of Hebrew, Aramaic, or Greek vocabulary. In the Libronix system sold by Logos, you may open up multiple resources at the same time and make a side-by-side comparison in tiled windows. You may copy and paste discussion from an electronic commentary or dictionary into a document of your own. Hebrew/Aramaic, Greek, Arabic, Syriac, and Ethiopic transfer over without difficulty in the easy-to-read SBL Hebrew, Gentium, Scheherazade, Serto Jerusalem, and Abyssinica SIL scripts, respectively. Hover over a scripture reference, and a verse of pointed Hebrew appears in a popup window. Click on a reference to a resource that is part of your electronic collection, e.g., GKC, Waltke-O’Connor, or Joüon-Muraoka, and the referenced paragraph appears in a new window. The quantity of data and resources available is immense and growing exponentially. Accessing it is like drinking water from a fire hose.
These features make it possible for the electronic study of a Hebrew vocabulary item or a particular passage to quickly reach a depth of inquiry that is impossible to match if one is limited to dead-tree resources.
On the other hand, dependence on electronic resources alone is self-defeating. Some of the best resources do not yet exist in electronic format. The important thing is to learn to use electronic, print, and online resources conjointly and effectively.
Let’s say I’m interested in exploring the semantics of אכל ‘eat.’ How often, in what binyanim, and in what contexts is it used? In what idiomatic expressions does it occur? How should one go about translating it in context?
[You knucklehead Hobbins! Everyone knows what אכל means. The dictionary is the place you go when you don’t know the meaning of a word - ed. That’s where you’re wrong. The most important discoveries in the fields of philology and linguistics are made over familiar terrain falsely so-called.]
Resources useful in exploring the semantics of ancient Hebrew vocabulary are numerous. In a piece entitled “A Brief Guide to Dictionaries of Biblical Hebrew,” I review dictionaries anyone with a serious interest in the vocabulary of classical Hebrew will want to be at home with. Other reference works, like those which canvass epigraphic Hebrew, Dead Sea Scrolls Hebrew, later Hebrew, and/or cognate languages, not to mention dictionaries of names, places, and concepts are also essential. Qimron, DJPA, DNWSI, and Olmo and Sanmartín come to mind, as do eABD, DBSup, RAC, TRE, THAT, and ThWAT (=TDOT). Unfortunately, I don’t have all of the resources I need in my personal library. In my case, and probably yours, a trip to a research library is essential to the completion of an honest-to-goodness research project. That brings up a dimension of serious study of the Bible I will post on in the future: the necessity of a working relationship with a friendly neighborhood research institution.Print resources I have on hand for doing linguistic and philological study are alluded to below. The dictionaries I have in electronic format which include discussions of biblical Hebrew vocabulary items are the following.
eDictionaries relevant to the Study of Biblical Hebrew
(1) Francis Brown, with Samuel Rolles Driver and Charles Augustus Briggs, The Enhanced Brown-Driver-Briggs Hebrew and English Lexicon. With an appendix containing the Biblical Aramaic. Based on the lexicon of William Gesenius, as translated by Edward Robinson, and edited with constant reference to the thesaurus of Gesenius as completed by E. Rödiger, and with authorized use of the German editions of Gesenius’ Handwörterbuch über das Alte Testament. Electronic ed. based on the Oxford: Clarendon 1906 edition as corrected in 1951. Oak Harbor: Logos Research Systems, 2000. Abbreviation: eBDB. The classic dictionary of biblical Hebrew, it remains an essential reference tool. A significant plus of eBDB: transliterations of Arabic, Syriac, and Ethiopic words are provided.
(2) Samuel Prideaux Tregelles and Wilhelm Gesenius, Gesenius' Hebrew and Chaldee Lexicon to the Old Testament Scriptures. Bellingham: Logos Research Systems, 2003 [1846]. Abbreviation: eTregelles. A proto-BDB as it were, with greater attention to some details. This dictionary deserves to be better known.
(3) M. E. J. Richardson, ed., The Hebrew and Aramaic Lexicon of the Old Testament. CD ROM Edition. (Translation and revision of Walter Baumgartner, Ludwig Koehler, and Johann Jakob Stamm, eds., Hebräisches und aramäisches Lexicon zum Alten Testament [5 vols; Leiden: Brill, 1967-1997). Abbreviation: HALAT]). Leiden: Brill, 1994-2000. Abbreviation: eHALOT. An essential reference tool. On the negative side: HALOT and eHALOT contain an inordinate number of typos and other errors. Its glosses and definitions, unsurprisingly, sometimes read as if they were translations of a German, not a Hebrew base. To be used in conjunction with (e)BDB.
(4) William Lee Holladay, A Concise Hebrew and Aramaic Lexicon of the Old Testament based upon the Lexical Work of Ludwig Koehler and Walter Baumgartner. Twelfth corrected impression originally published 1991. Leiden: Brill, 2000 [1971]. Abbreviation: eHolladay. A briefer version of (3). Concision, of course, is valuable in its own way.
(5) James Swanson, Dictionary of Biblical Languages with Semantic Domains: Hebrew (Old Testament). Abbreviation: DBL. Electronic ed. Oak Harbor: Logos Research Systems, 22001 [1997]. The chief advantage of this dictionary is the classification of vocabulary into semantic domains according to the typology devised for New Testament Greek by Louw and Nida.
(6) Enhanced Strong’s Lexicon. Woodside Bible Fellowship, 1995. Abbreviation: ESL. For every Hebrew word, an overview of KJV translation equivalents is given, with number of occurrences. The resource is of interest to students of translation.
(7) Robert L. Thomas, Hebrew-Aramaic Dictionary of the New American Standard Exhaustive Concordance. Updated Edition. Anaheim: [Lockman] Foundation Publications, 1998 [1981]. Abbreviated title: DNASV. Derivations and “general meanings” are taken from BDB. “Specific meanings” = NASV translation equivalents are also given, with number of occurrences. The resource is of interest to students of translation.
(8) R. Laird Harris, Gleason Leonard Archer, and Bruce K. Waltke, eds., Theological Wordbook of the Old Testament. Electronic ed., Chicago: Moody Press, 1999 [1980]. Abbreviation: eTWOT. Not without its uses, this resource is nonetheless inferior to Botterweck-Ringgren (ThWAT/TDOT; not yet complete in English), Jenni-Westermann (ThAT/TLOT), and, more often not, to the OT coverage in the “Outside the NT” sections in Kittel (ThWNT/TDNT). The sooner TDOT is available in electronic format, the better. TLOT is available through Accordance. For TDNT, see the next item.
(9) Gerhard Kittel, Gerhard Friedrich, and Geoffrey W. Bromiley, eds., Theological Dictionary of the New Testament (10 vols.; tr. Geoffrey W. Bromiley and [vol. 10] Ronald E. Pitkin; Grand Rapids: Eerdmans, 1964-1976 [= Gerhard Kittel and Gerhard Friedrich, eds., Theologisches Wörterbuch zum Neuen Testament (10 vols.; Stuttgart: Kohlhammer, 1933-1979)]. Abbreviations: ThWNT/eTDNT. TDNT vol. 10 (1976) is missing the Literaturnachträge of ThWNT vol. 10 (1979) 946-1294. The omission is unfortunate. TDNT’s discussions of NT vocabulary items, where they touch upon the equivalent vocabulary items in the Hebrew Bible, are an important resource.
(10) Johan Lust, Erik Eynikel, and Katrin Hauspie, A Greek-English Lexicon of the Septuagint. Revised Edition (Stuttgart: Deutsche Bibelgesellschaft, 2003 [1992, 1996]). The electronic edition is available through a number of Bible software firms. Abbreviation: eGELS. The look up of a Hebrew vocabulary item via the search function turns up instances of lexicographical significance in which a Hebrew base text at odds with MT is suggested.
Each of these resources has value. The verb אכל ‘eat’ may illustrate. Strictly speaking, TDNT is not a Hebrew dictionary at all, but one may find the appropriate entries in eTDNT via a search for ‘eat’ in eTDNT from the Libronix Basic Search function. The results make it clear that the entries ἐσθίω and τρώγω deserve the closest attention.
Defining the Corpus of a Word-Study: The Example of אכל
So then, how often is אכל used in the Hebrew Bible? Logos has a module entitled Bible Word Study. The module was probably not designed for the purposes I will put it to, but let’s see where it takes us. I put אכל into its search engine. A wealth of data is generated over the course of a few minutes. At the top, a sum total of אכל occurrences is given: 816. A bar graph presents a book-by-book breakdown. The size of each bar takes into account the size of the book it represents relative to the others. The feature is helpful. Lev, Deut, Hos, Joel, Amos, Nahum, and Qoh tower higher than the others. Anyone familiar with these books will not be surprised. Among print resources on hand, I notice that VOT claims 820 occurrences of the verb אכל in Hebrew (755+4+45+1+1+21–7 occurrences of אכל in Aramaic). DCH and TLOT, claim 809 occurrences. Even-Shoshan lists 807 occurrences.
The causes behind the variation in count are not obvious, and there is no easy way to figure the matter out. A manual comparison of Mandelkern, Lisowsky, and Even-Shoshan, while possible, would be time-consuming, and would not explain the higher counts of VOT and the Bible Word Study. The latter two resources give book-by-book statistics. Adjusting for VOT’s inclusion of occurrences of אכל in Aramaic, their counts differ by a margin of 1 or 2 in four cases.
I return to the Bible Word Study. Below the total and bar graph of אכל, a series of Keylinks is in view. Five are listed: eBDB, eHALOT, DBL, eTregelles, and eHolladay. I click on “More” and eight additional links are uploaded: eStrong, eNAS, references to אכל in introductory textbooks and helps, in a reference grammar, and eTWOT. Of the 10 resources I originally thought of for the purposes of a word study of אכל, all but eTDNT and eGELS appear. Their omission is understandable. Reference tools of lesser importance also appear. A few I thought would appear do not.[1]
I notice that eBDB claims 806 occurrences; ESL, 802.
Below Keylinks, a section entitled Grammatical Relationships appears. Grammatical relationships in which אכל occurs are groups into categories and exhaustive lists provided. This is little bit like having The Dictionary of Classical Hebrew (DCH) in electronic version format at one’s fingertips, without, unfortunately, extra-biblical occurrences of אכל included. One may upload the actual occurrences in context, and after that, the Andersen-Forbes Phrase Marker Analysis of contexts of choice. All of this is magnificent.
Below Grammatical Relationships, a unit entitled Translation appears. I will return to it in a future post. Last but not least, a concordance of occurrences of אכל appears, listed according to morphology. Entitled Biblica Hebraica Stuttgartensia: with Westminster 4.2 Morphology, it also gives a total of occurrences: 820. This contradicts the 816 showing up at the top of the Word Study beside the bar graph, also derived from BHS/WHM 4.2.
If I put the verb אכל into the search engine of the Andersen-Forbes Phrase Marker Analysis (AFPMA), another total is given: 792. As noted earlier, VOT, another Andersen-Forbes product, gives 820. Divergent totals of this magnitude probably point to issues of broader import. It will be interesting to find out what they are. [I discuss them in a future post.]
The coda of the concordance provided strikes the eye. The data presented is corrupt. The phrase אֲכָלַנִו חֲמָמַנִו occurs in context twice under the heading (2) v #2 (WTS) המם . אֲכָלַנִו חֲמָמַנִו appeared in context earlier, once under the heading “verb, qal, perfect [ketiv],” and once under the heading “verb, qal, perfect [qere]. הֲמָמַנִו is misspelled as חֲמָמַנִו in each instance.
To be sure, the masoretes record two occurrences of אכל by means of a single, per se impossible form: אֲכָלַנִו = אֲכָלָנוּ (ketiv) and אֲכָלַנִי (qere). But a presentation of the data four times without ever disambiguating it is not helpful. There are four sets of qere/ketiv readings in the MT אכל corpus. They all show up in the concordance with a total of 820 provided. 820 reduces to 816 following standard counting procedures, in agreement with the figure at the top of the Word Study.
But is the standard counting procedure correct? After all, both אֲכָלָנוּ and אֲכָלַנִי are attested in MT Jer 51:34. A comprehensive presentation of אכל data in MT would provide one total counting qere/ketiv readings as two occurrences, and another that does not. The Word Study seems to do this. If so, the totals should be explained for what they are in pop-up windows.
BHS/WHM 4.2 has some rough edges, but I prefer its Isa 37:30 וְאִכְ֯ולֻ֥ פִרְיָֽם to The Hebrew Bible: Andersen-Forbes Analyzed Text (AFAT) Isa 37:30 וְאָכוֹלק̇ פִרְיָם. The latter appears in a window if one hovers over the reference in BHS/WHM. AFAT makes the aspirated פ of פִרְיָם look like an error. It would have been wiser to read the qere in the body of AFAT, and superscript the ketiv with the vocalization that best seems to suit it.
How does Even-Shoshan fare? אכל in Jer 51:34, Ezek 16:13; 44:3, and Isa 37:30 is listed as אֲכָלַנִי, אָכָלְתְּ, לֶאֱכָל־, and וְאִכְלוּ, respectively, the qere in every case. In two cases, the ketiv is given in parentheses. Why not the other two? Mandelkern registers all eight readings, but you have to know where to look. Qere/ketiv readings are not presented in a consistent fashion.
Let us summarize results so far. None of the resources currently available, electronic or print, gives its users a comprehensive, error-free, and consistent presentation of the raw data relative to אכל in MT.
The advantage of electronic resources is that they are readily modifiable. Future editions of BHS/WHM, and AFAT, I expect, will be free of the problems just noted. The wait between one edition and another, furthermore, will be miniscule compared to the span of time that separated the appearance of the corrected BDB (1951) from its predecessor (1906).
Alice in אכל-land
I'm still wondering how many occurrences of אכל in MT there are. A Hebrew Morphological Bible Search specifying אכל as a verb, Hebrew words only, with BHS (WTS) and Hebrew Morphology (Westminster) as additional parameters, yields 862 occurrences according to BHS/WHM 4.2 tags. I scroll down and take a look at the book of Ruth, a text I feel I know well, to see what’s up. The morphological analysis is wrong two out of six times. הָאֹכֶלin לְעֵת הָאֹכֶל in Ruth 2:14 is highlighted as if it were a verbal form,[2] not the noun that it is. אוּכַל in לֹא אוּכַל לִגְאוֹל in 4:6 is highlighted as if it derived from אכל. The correct derivation, of course, is from יכל. The bar graph at the top of the Word Study, described as dependent on BHS/WHM 4.2 data, gives a total of 5 occurrences of אכל in Ruth, not 6. Better, but still not correct. The correct total is 4.
A Morphological Search specifying אכל as a verb, “all resources of specified morphology” as the database, and Hebrew Morphology (Andersen-Forbes) yields a total of 2586 occurrences. The number is mysterious. The following forms are parsed in The ESV English-Hebrew Reverse Interlinear Old Testament [ESVIB] as if from אכל:
וַאֲכַלֵּם (Exod 32:10);
וָאֲכַלֵּם (II Samuel 22:39); and
וָאֲכַל (Ezek 43:8).
In each case, the correct derivation is from כלה. As far as the book of Ruth is concerned, across all three resources indexed (AFAT, ESVIB, and The Lexham English-Hebrew Interlinear Bible [LIB]), the analysis appears to be wrong one out of five times. הָאֹכֶל in לְעֵת הָאֹכֶל in Ruth 2:14 is highlighted as a verb, not the noun that it is.
A Search specifying אכל as a verb, BHS/WIVU as the database, and Hebrew/Aramaic Morphology (WIVU) yields 861 occurrences. Not to be outdone by its competitors, הָאֹכֶל in לְעֵת הָאֹכֶל in Ruth 2:14 is once again misconstrued as if it were a verbal form, not the noun that it is.
Is it really that hard to parse הָאֹכֶל in לְעֵת הָאֹכֶל ‘at mealtime’ correctly? Not to my knowledge. If I examine AFAT Ruth 2:14 in situ, it turns out הָאֹכֶל is understood as a noun. Ditto for BHS/WHM 4.2 and BHS/WIVU. The databases parse the form correctly, but one is led to believe otherwise based on the searches I just described.
As a last resort, I visit the Logos website, and immediately discover the problems noted are well-known. It also turns out I was performing the search incorrectly (see footnote 2). Workarounds exist. I follow instructions, and soon have in hand the lemmatized analyses of אכל of the three databases of the Tanakh in the Scholar’s Library: BHS/WHM 4.2, BHS/WIVU, and AFAT. It is fabulous to have all three. Comparative study of their contents is an effective way of flagging problems and identifying promising avenues of research. Aside from two obvious errors in WHM and three in AFAT, the databases agree in all but four instances, with WIVU deriving אָכְלוֹ in Ex 12:4; 16:16, 18, and 21 from the noun אֹכֶל, and WHM and AFAT parsing the same forms as suffixed infinitive constructs of the verb. In short, according to WIVU, the verb אכל occurs 810 times in MT; according to WHM and AFAT (after the elimination of error), 814 times.
In the next post, a foray into the field of text-criticism will serve to define the contours of the אכל corpus with greater sharpness.
[1] In a word study, Keylinks to GKC, Joüon-Muraoka, UT, CAL, GELS, and similar resources are not yet possible. But a “Basic Search” on a string of Hebrew consonants with GKC or a similar resource open yields a decent set of results.
ק̇ Ԡאִכְלו ּ.
[3] The errors: WHM 4.2 lists אוּכַל in Ruth 4:6 as from אכל: it’s a form of יכל. AFAT lists אֲכֶלְךָ in Exod 33:3 as from אכל : it’s a form of כלה. Both list וּלְאָכְלְכֶם in Gen 47:24 as a suffixed inf. construct: it’s a suffixed form of the noun אֹכֶל. If inf., it would have been אֲכָלְכֶם (see Josh 23:13; Gen 3:5). AFAT misparses Hos 13:8 וְאֹכְלֵם as Hiphil: it is Qal.
A printable version of Parts 1-3 of this demonstration is available here.
You might also like to look at SDBH, http://www.sdbh.org/, which is the United Bible Societies' work in progress on a Hebrew dictionary organised by semantic domains. The format is intended for a specialised audience.
Another project in progress, and at an earlier stage, is KTBH, http://www.ktbh-team.org/, which I worked on for a time. So far only a few sample entries are available (I wrote the ones on the כהן word group), and the e-mail link to the team is broken.
Sadly, neither of these projects yet has an entry for אכל for you to compare with.
Posted by: Peter Kirk | May 01, 2007 at 12:46 AM
Thanks for the links, Peter. I temporarily lost your comment while fixing some glitches. It's back now.
Posted by: JohnFH | May 01, 2007 at 12:48 AM