Analytical Greek New Testament (AGNT) Project History
summarized by Timothy Friberg (current as of August 2021)
The AGNT project began when I attended a course organized by Wycliffe Bible Translators in Dallas TX during the summer of 1976 called “New Testament Greek Discourse Analysis.” It was a fascinating course, but one in which I quickly learned how rusty my Wheaton College major in Greek (1968) had become. My paper for the summer was on the discourse structure of the Greek text of Galatians, one that I expanded to become my masters thesis topic at the University of Minnesota (1978). I had the idea—become vision—to give the entire Greek New Testament a morphological analysis that would enable me to do a better job at my Galatians task and that would also aid numerous other Bible translators as they tried to understand the Greek text underlying their New Testament translation efforts.
Barbara, enabler as she is, quickly came alongside me as I started to analyze each succeeding Greek word, first in Galatians and then generally throughout the New Testament. John Werner, who had been one of several Greek experts both consulting the summer course in Dallas and at the same time himself learning the theory behind propositional analysis of discourse, started to check my work long distance.
The University of Minnesota Computing Center (UCC) under Dr. Peter C. Patton offered grants of computer time by which Barbara was able to run concordances of my analyses across individual New Testament Greek books, howbeit in Latin letters, the better to analyze individual words set together in groups of like occurrences. Through Wycliffe and other sources I was able to identify a number of Greek professors across the USA that were willing to check the morphological tags assigned to Greek words in concordance format. I mailed off UPS packages to distant addresses, interacted with professors by regular mail and telephone and gradually was able to put together an analysis that came to be called the Analytical Greek New Testament.
Dr. Patton of UCC and I discussed possible published forms of our project. UCC and we Fribergs finally signed a contract with Baker Book House of Grand Rapids MI to publish our product in three forms. The first was to be the simple text of the Greek New Testament with an interlinear placement of our morphological parsings. This came to be called AGNT, the Analytical Greek New Testament. The second was a concordance of the basic AGNT text, but for this format with tags following Greek words in the line of text. The presentation form was in two volumes, the first a concordance of Greek words and the second a concordance of Greek tags. This output was generated by the concordance programs developed by Richard Hotchkiss of UCC, which he gladly modified to fit our exacting requirements. This came to be the Analytical Concordance of the Greek New Testament, sometimes called ANCONC and sometimes merely CONC. Finally, the third form of our project was an Analytical Lexicon of the Greek New Testament, ANLEX.
Baker published AGNT in 1981 in a volume edited by Barbara and me, but happily acknowledging the crucial input of many cooperating volunteer scholars. The two-volume ANCONC set was published by Baker in 1991, after extensive editing by John and Claire Hughes of Whitefish MT. Its named editors were Philip S. Clapp, a retired Greek scholar, and the two of us. Finally. ANLEX was published by Baker in 2000 as a volume edited by Neva Miller, a lifelong teacher of Greek, and the two of us.
At the time that AGNT was published, Barbara and I headed back to Asia from our lengthy study programs at the University of Minnesota. I continued to correspond with Philip Clapp, finalizing the format of ANCONC. Actually, for a number of years the computer tapes of ANCONC prepared by UCC had sat unused and lost to us. When I sought to learn if they were still extant, I was pleased to learn that they had been found. From there copies were sent to Baker, who contracted with the freelancing Hugheses, and finally were printed by Baker as ANCONC. Philip Clapp died after his work was finished but before the concordance set was printed. It was our delight to present the printed volumes to his widow at her home in Portland OR.
Similarly, I corresponded with Neva Miller for years, also from our location in Asia. And we visited her when we were in the States for further discussions relative to ANLEX. Neva herself finished up her work on the manuscript, but died (November, 1997) before ANLEX was published by Baker. Baker’s designated editor on the book, David Aiken, and I carried the project to completion, mainly extensive proofreading and formatting decisions.
Around 1990 we employed John Baima of Silver Mountain Software to handle electronic aspects of our AGNT project, both because that wasn’t our particular expertise and because we were located in Asia. This need arose because several Bible software developers started to approach us about licensing AGNT. So John made the text compatible to the needs of those vendors. SIL also expressed interest in our database for several of its programs, eventually the Bible Analysis Research Tool (BART) and Translator’s Workplace (TW).
About the same time Robert Smith, long-term volunteer attached to SIL’s translation department, urged us to simplify our approach to tagging in certain systematic ways. I told Bob that I was very interested in his suggestions and would implement them if he were willing to work with me on the not-insignificant simplification process. During the next years Bob and I exchanged volumes of correspondence and I interacted with John, who implemented the changes in the electronic database. The original AGNT tagging system is that still found in the printed form of AGNT (Trafford; and the forthcoming corrected and updated electronic version of the same, Classic AGNT); the electronic forms now used widely by vendors are entirely the simplified tags.
While he was volunteering with SIL at its Dallas center in the late 1990s, Ulrik (now Sandborg-) Petersen of Denmark converted our AGNT Appendix from being an explanation of the original tagging system to a presentation that fully explained our revised and simplified tagging system. That revised Appendix is now available to vendors using AGNT to display for users of their software.
While Bob Smith was working with me on our simplification schema, he repeatedly brought to my attention the desideratum of producing an English interlinear to accompany our morphological tagging. I was at first highly resistant to such an approach, mainly because I didn’t want to facilitate would-be experts in the Greek text that in fact knew next to nothing about New Testament Greek. Finally, Bob and I agreed on a way by which we might provide assistance to the user for this or that unrecognized Greek word without making it a presentation of the text in English that could be read from left to right as a translation.
Thus I began in the late 1990s providing draft English gloss assignments to the project database taken from our ANLEX lemma write-ups. This was a slow and plodding undertaking, mainly because it was only hobby time activity to me in our Asian location. But I did seek out volunteer Greek scholars to come up with their own draft assignments for what came to be known as ERGs, English reference glosses, later to be compared against mine. As a footnote to the undertaking, I trained some forty volunteers over a five-year period, but most of them fell away, almost all of them because their volunteer task wasn’t a priority in their busy lives. But I did find a few disciplined souls out there, in particular Robert Merz and Merlin Zook. Both of these men worked hard for the AGNT project on ERG assignments until their deaths.
I was all the while thinking about possible enhancements to the AGNT project, a wish list for days when volunteers might abound. (See elsewhere for a discussion and prioritizing of just such a list!) We contracted with SIL to use our modules (both AGNT and ANLEX) with the proviso that they would find volunteers for us in lieu of paying royalties, which I was loath to assess against the organization for the benefit of whose members the AGNT project was initially conceived. And there has been a small but serious stream of volunteers in recent years.
Barbara and I are credited as creators of the AGNT database, even though we gladly acknowledge that there wouldn’t have been such a database without the selfless labors of dozens of others. When royalties began to be paid us, first by Baker for our three volumes in their “Greek New Testament Library” and later by various electronic vendors using AGNT, we decided that it was best all around to give the royalties fully and entirely to the Lord; we would fill the role of trustees of the income and the ongoing project.
Over the years we have done just that, seeking to find appropriate places to share the royalties. Though our volunteer helpers have never received pay, we have at times offered long-term volunteers the privilege of designating recipients to this or that gift. Our first electronic agent, John Baima, was given a fifty-percent cut of royalties for his not-insignificant contributions to helping us maintain, develop and mark our texts. John was very active in actual hands-on computing for our project. About thirteen years ago, due to John’s busy involvement in his own Silver Mountain Software company, regular employment and other involvements, we replaced him with John Hughes of Bits and Bytes, Inc.; John with his wife Claire were the team that originally edited our ANCONC for publication. We currently pay John 49% of royalties for his role as our electronic agent. John has solved the matter of his own busyness in many other endeavors by contracting out AGNT computing projects to various programmers. Ulrik Sandborg-Petersen has been significant among them. Another major programmer is Mark Nodine.
The target audience for AGNT modules has turned out to be two separate groups. First, there is SIL with their programs for translators, of which AGNT has become a core component. Much more recently, the UBS Paratext (now jointly being developed with SIL) has contracted with us for a use of our database. Both SIL and UBS use our AGNT without payment of royalties. They are for their parts both supposed to help us find volunteer Greek experts. That actually only works out for SIL—and that irregularly. For UBS their identifying volunteer helpers is still fully theoretical; they have chosen instead to fund paid computer programmers. (This first grouping also includes certain scholarly undertakings not producing income for the researcher.)
Second, royalty-paying vendors of various kinds. As of this date they are all electronic presentations of one or more aspects of our AGNT project. In every case the AGNT presentation is only one part of a larger inclusive database presentation. Among those now in a contractual arrangement with us are Logos, Olive Tree, and WORDsearch. As such they report to and pay royalties to our electronic agent, now John Hughes.
Several years ago Ray Gordon came on board as a volunteer expert, recommended by Mark Testardi, but quickly claimed by SIL as their volunteer to our project! Ray worked hard on taking the draft ERGs that I had finally finished, checking them against the work of other volunteers and coming up with an ERG module that was finally marketable as a beta version. In the process of his work, we extended the concept of ERGs to include PERGs (phrasal ERGs), PLERGs (phrase literal ERGs) and Greek phrases.
Ray has also worked hard on revising ANLEX (an ongoing project, but one that Ray has had major input toward). In particular, he has taken a number of our smaller tasks, commissioned and completed in past years, and integrated them into the ANLEX write-ups. Ray has also begun the nontrivial task to mark the P/ERGs in AGNT with a special pointer so that they are uniquely pointed toward their source location in ANLEX. This will be a major undertaking, but one that needs to wait as follow-up to a more exact (and/or final) fixing of the ANLEX write-ups with respect to the various revisions they are now experiencing. Indeed, we have recently engaged Tony Pope to check our P/ERGs and to assess the correctness of our ANLEX write-ups. Tony’s relevant expertise to the AGNT project is Greek lexicography. As this is just beginning and Tony is a volunteer, this vigorous checking process may go on for several more years.
Early 2013 Ray, Mark (Nodine) and I worked hard on readying our P/ERGs for a new release of UBS’s Paratext, then current as Paratext 8.0. The first beta year of that module’s availability resulted in moderate input for its correcting and upgrading. That because ERGs are fairly well set, but PERGs have only begun to be refined. For example, Ray took all phrasal meaning units discussed in ANLEX and entered them into the AGNT database as PERGs. But there are many more Greek phrases, not found in ANLEX definitions, whose unitary and figurative meaning does not equal the more literal meanings of the individual parts thereof, that is, of the individual Greek words. This is a place for further refinement and enhancement. (Indeed Tony’s current task will give everything a new scrutiny.)
One task I undertook several years ago was (again) to read through the entire Greek New Testament (both GNT4 and Byzantine Textform, also BT) to give each next distinct unit a thoughtful evaluation, that is, to include Greek word, AGNT tag, ANLEX lemma, ERG and PERG assignment. The interplay between several of these was checked; for example, if a tag notes that two different analyses are possible, a check was made whether that choice should be reflected in lemma or P/ERG as well. Mark Nodine wrote the program, that is, making a horizontal spreadsheet with this information for both databases in parallel; he now provides with an updated version every six to twelve months.
We have also had an ongoing check of our Greek databases (GNT4 and Byzantine Textform) for a number of things, to include exact representation down to punctuation and accenting against the underlying texts, checking for unique (and thus probably wrong) tags, checking for tag and Greek reflex combinations not now accounted for in ANLEX, checking for lemmas not found in ANLEX, etc. I recently completed such a check getting ready a new version, that is, AGNT for GNT5. In relation to that we also added AGNT-NA28, since punctuation differences obtaining between GNT5 and NA28 do in fact affect a few AGNT tags. We now maintain six databases: Classic AGNT (GNT3), AGNT4 (GNT4), AGNT5 (GNT5), AGNT-NA28, BYZAGNT, and AGNT-WH.
Sam Pflederer worked for us full time in 2013 with impressive results. His main contribution was to add a field to both database kinds (that is, eclectic and Byzantine Textform) that will point the user from the tag associated with the Greek word he is reading to the explanation of that tag as given in our AGNT Appendix. This was a massive project, one that added roughly 40,000 annotations to tags in both databases. At the same time as he was doing the annotation project, Sam revised the Appendix itself. Among other things, he made sure that the examples cited were relevant to both GNT and BT; he added headings and made other relevant changes that would make the user’s perusal of the Appendix that much more fruitful; he made the Appendix more comprehensible to a beginning learner of Greek.
The current annotation field now contains, largely due to Sam’s contribution, a strict tag that gives the bare morphological information expressed in each Greek reflex, that is, the traditional expanded AGNT tags, though retained, are now supplemented by this minimal parsing; and an identification of “seconds” for relevant Greek futures, aorists, perfects and pluperfects. We have also very recently moved an identification of participles with imperative sense from the tags (formerly VR… now combined with all other participles as VP).
Sam has also produced a new field for each AGNT record, in particular, he implemented Jan Hoogland’s TGNT (Transformation Greek New Testament) Intrinsic Meanings. In most cases this is a one-for-one assignment of intrinsic meaning to Greek lemmas, but in a few others, it involves the nontrivial assignment of intrinsic meaning from a set of multiple intrinsic meanings, due to the fact that some ANLEX lemmas are really conventional conflations of separate Greek words. This was no small undertaking.
This was not officially an AGNT project, but done more as a favor to Hoogland. The fact that he didn’t factor into his system changing meanings for given lexemes makes his approach very problematic. I arranged with BibleWorks to run it in their BW10 release, but in such a way as it is not seen as an integral part of AGNT. Oh the things we are driven to do for friends.
Shortly before Sam closed out his year with the AGNT project to head out as a translator with his new bride first to Brazil for Portuguese language learning and then more recently to Mozambique for translation work, he made a comparison of the Maurice Robinson parsings for both GNT4 (actually WH) and BT and our own for the same two texts. I started to work to find the difference in underlying analyses between Robinson and me to ascertain which is correct. Somehow it wasn’t a satisfying instrument (I don’t blame Sam; he’s brilliant), so rather it was reworked by input from Michael Bushell (BibleWorks) and Mark Nodine.
There are a number of AGNT tasks currently being worked on. I will mention them here, giving up-to-date summaries of them.
1. We have had Carl Conrad working on a parallel AGNT version in that his take on the analysis of Greek middles and passives is different from the traditional analysis—but very probably right. He has finished his review of every AGNT tag dealing with voice; and he has completed an alternate statement to our Appendix discussion of voice. He also has also written a longer treatment of voice now found as two new appendices to ANLEX (4, 4a, both found elsewhere in this website). He finished his commitment by reviewing the individual write-ups of ANLEX toward a parallel write-up different from the concept of deponents; a small aspect of that will also be to change citation forms from first person singular to infinite form.
2. Daniel Hoopert is giving a half a day weekly reviewing the reanalysis of conjunctions that John Werner originally undertook for us, but didn’t complete because of advancing Parkinson’s Disease (but he came very close). Actually only a part of καί remained of all the conjunctions. Dan started working on that but diverted to the conjunction οὖν at my request. His work is very thoughtful. As long as he has the will and joy to continue, I would like him to look over each and every reflex of each and every conjunction (all sixty-seven of them). Mark Nodine provided Dan with a parallel spreadsheet (GNT4 and BT) wherein all the conjunctions of the Greek New Testament are highlighted for review in context. The verses where the GNT4 and BT differ, if only a single word, are marked so that Dan needn’t read both texts in full, but only one of them and then the other only when it is marked as being different.
Ray Gordon is continuing a strong beginning in producing a highlighting (two colors) enhancement of our ANLEX write-ups by indicating the exact place in those write-ups where the ERG from AGNT is taken. Testing among ANLEX users has shown that the typical user would be very happy indeed to have the source of the ERG in ANLEX highlighted so that he can from that place move up and down the lexical write-up to personally evaluate the choices suggested, thus saving himself much time.
For this task Ray is using the revised sections of ANLEX being prepared by Tony Pope, thus giving the latest revision version of our ANLEX.
Though we still need to finish implementing the parallel (togglable) tags that Carl Conrad has made for us, most of our current work is progressing determinedly on all our fronts. It really cannot be otherwise, because the various current aspects of our work are so interrelated.
There you have it, a bit of history of the AGNT project melding into our current tasks.
Click below to download a PDF copy of this page text.