This note covers one final case of multiple pdfs posted at NARA under one record number. Like the other cases discussed in the previous two notes (available here and here), this note involves CIA info in FBI records. Unlike the previous notes, this time I go into problems in the 2022 CIA document index which I have tried to use to explain things.

After acknowledging and discussing these problems, I close with some thoughts on the purpose and practice of the “duplicate records”, and with their implications for determining how many records still have redactions.

Like the last two posts, this note is hardcore inside baseball. I will, however, try to explain how some of this matters for people who just want to track ARC releases. If that describes you, see the end comments and skip the rest of this note, Otherwise, read on — if you dare! Mwa ha ha ha!!

Yet another introduction to the problem

An enduring puzzle in ARC releases has been the oddball “duplicate records” in a number of NARA releases. These were pdf files posted at NARA’s JFKARC website which purported to be “ARC record number so and so,” but actually turned out to be fragments of “record so and so.”

This happened primarily for FBI records, and as we saw, you could get 3, or 4, or 12, or even more fragments from a single record in one particular release, e.g. the big April 2018 release. To make matters even more mysterious, full copies of the records (incorporating the redactions used in the fragments) were often released at the same time.

This continued for the big December 2022 releases as well, and as we saw in the previous two notes, we were able to track some fragments from 2018 to 2022.

2022 also saw the release of a lengthy CIA document index (here) which shed some light on all this. Many of these fragments showed up in the doc index, and after comparing many fragments to whole files, I suggested that the fragments were pages from the complete FBI file which had CIA equities, i.e. these pages contained info supplied by the CIA.

I then theorized that the fragments represented separate releases of CIA reviews of this material, whose publication was meant to enable us to track the progress in opening the record, and distinguish which redactions were done by the CIA and which were done by the FBI.

The discussion below will focus on another such set of duplicate records. This set supports my understanding of the duplicates, but also presents some problems for my theorizing.

More “duplicate” releases: ARC record number 124-10185-10099

This note deals with another FBI administrative/liaison file, numbered 62-116395. I believe it was originally an FBI/Church Committee liaison file, but portions of it were also provided to the HSCA, which is how it wound up in the ARC.

There are some interesting notes on this file and its relevance to the JFK assassination in a long memo addressed to ARRB staffer Laura Denk, available from the MFF website here.

The ARC record we are looking at today, 124-10185-10099, is of course not the entire liaison file (Denk says the entire file is 62 volumes). Instead, it is identified in the JFK database as “62-116395-441 THRU 5TH NR 474”, or about thirty serials (sections).

As far as I can see, this chunk of the liaison file has nothing related to the assassination. The JFK database comments on this record call it NAR, “not assassination related.” This was the FBI’s judgment and I believe it is correct. However, since the ARRB decided to release every piece of paper in the HSCA files, this too had to be published.¹This is an overstatement perhaps, but not a big one.

This excerpt of the liaison file is stuffed full of actual CIA records, not just CIA supplied information, and must have presented quite a few headaches to the CIA redactors. It also must have presented a puzzle to the CIA guys who were compiling the CIA document index released in December 2022. Let’s take a look at what the CIA doc index shows us about this file.

The doc index lists three file fragments from 124-10185-10099. Our working assumption is that these fragments contain CIA info and/or CIA docs. Oddly, however, only one file fragment was released at NARA in December 2022, plus, of course, the redacted full file version. In the 2018 and 2023 releases, only the redacted whole file version was released. No file fragments were released at all.

In other words, of the three file fragments listed in the 2022 CIA doc index, two of them have never been published at all: not in 2018, not in 2022, not in 2023. These are not just fragments, they are phantom fragments!

Could it be that these are not just phantoms, but fictions? It seems not. The 2022 CIA doc index gives us the number of pages in each fragment, and the date of each fragment as well. This enables us to search the 288 page redacted full FBI file to see if there are CIA docs incorporated in the file with X pages and Y date. And the answer is yes, there are.

The first fragment, labeled c06716627 in the CIA doc index, is two pages long, dated 1975-07-29, and contains “operational details.” This matches up with an FBI memo on pages 254-255 in the 288 page pdf (here are links for the 2018 version, the 2022 version, and the 2023 version. As this comparison shows, the redacted material is gradually released in 2018 and 2022, and the 2023 version is released in full.

The second fragment, labeled c06716626 in the CIA doc index, is eight pages long, dated 1975-07-29, and contains “people” info. This matches up with a CIA list on pages 221-228 in the 288 page pdf (here are links for the 2018 version, the 2022 version, and the 2023 version.

This second fragment is a CIA memo, listing all staff members of the SSCI who meet the qualifications for top secret clearance. What is redacted in the list? Their social security numbers. These are also gradually released from 2018 to 2023, hopefully after the staff members have passed away. Apparently a number are still alive, and their SSNs are therefore still protected. This is the only reasonable way to handle this.

So these file fragments listed in the doc index have a basis in the text of record number 124-10185-10099. Moreover, this text is being periodically released. But the CIA reviews of the these components are not being published at NARA. Maybe they forgot. Or maybe my theory about publishing this stuff to show agency review work is not correct.

Well, what about the file fragment that was published in 2022? There were two pdfs published at NARA in 2022, one pdf is 170 pages long, and one is 288 pages long. The 288 page version is the full FBI file, that we also found in 2018 and 2023, with slowly shrinking redactions in each version.

The 170 page pdf released in 2022 is listed in the 2022 CIA doc index, and uses the file naming convention we saw for the other file fragments in our previous two notes. However, 170 pages is a pretty big fragment. Could it really be just one doc, stuck in the middle of the FBI file? Nope. On closer examination, this turns out to be not one CIA doc in the full FBI file, but a collection of different CIA documents in the full FBI file. Moreover this collection is not just pulled out one by one from the full FBI file. Instead, dozens of CIA docs, from the beginning, middle, and end of the FBI file are randomly thrown together.

Comparing the dozens of docs in the 170 page fragment with the corresponding pages in the 2022 release of the full FBI file, it seems that the releasing is the same. That is, text redacted in the 170 page fragment is redacted in the 288 page file. Text released in the fragment is released in the file. However, the whole thing is such a gigantic mishmash that I cannot guarantee this 100 percent.

To be one hundred percent sure, you would have to print out every page in the 288 pager, evey page in the 170 pager, and fit them together like a documentary jigsaw puzzle and eyeball each page repeatedly. I did one eyeballing on my computer screen and that was all I could stand. My time on this world is not infinite.

Anyway, it is quite hard to understand why these file fragments are listed this way in the 2022 CIA doc index, and why the one fragment was released this way. Why was the eight page list described in the CIA doc index, NOT released in 2022 as a separate pdf? Why was it not combined with the 170 pages of the big fragment, since that is already a collection of different documents? The whole thing is a puzzle.

More puzzlements

It seems to me that this third case study confirms my ideas about the nature of the “duplicate” FBI records. That is, they are indeed fragments of a full FBI file which are being separately released. They do indeed consist of pages with CIA docs or pages with CIA info. This mysterious process is therefore related to CIA review of its equities in FBI docs.

But why are these review files being published at NARA? Why are they being published inconsistently and sometimes in a horrible mess? Why are the review files divided in such arbitrary ways? It doesn’t seem like NARA or CIA or the FBI are doing this to clarify their methods and progress, does it? I really don’t get it.

One final puzzlement here is the question of who these redactions belong to. Could it be that, arbitrariness and inconsistency aside, this system at least tells us that all the redactions in this particular FBI file are being taken by the CIA? In other words, the 2022 CIA doc index is a list of docs where the remaining redactions are the sole responsibility of the CIA. That was sort of my assumption when I first look at this system.

However, on closer examination of this particular file, it seems this is not the case. Look for example at page four of the 2023 release of the full 124-10185-10099 file (MFF link here). There are two paragraph-length redactions here, where the wording strongly suggests this is an FBI composed doc consisting of FBI owned info. Moreover, this page is not part of any of the CIA’s phantom fragments.

This means that BOTH the CIA and the FBI are postponing text in the 2023 version of the FBI file. So the 2022 CIA doc index does not mean that redactions in this doc belong solely to the CIA. Yet this record appears only in the CIA doc index. It does not appear in the 2022 list of FBI docs which still have redactions. This is also disconcerting. There are not yet, it seems, complete lists of remaining redacted docs in the ARC. Yikes!

We are back, it seems, to the situation in 2018, when it was bafflingly unclear, after tens of thousands of releases, which records were still redacted and which were not. Of course the situation now is greatly simplified because the 2021 update of the JFK database clarified the status of a huge number of records. The 2022 CIA doc index also clarified the status of records released in 2022. It is only the four thousand plus records that were not released in full in 2022 that we are still uncertain of. Combining these two sources (2021 database plus CIA index) thus helps a lot. But there is still no exhaustive list of redacted docs.

As I said after the ARC record review mandated by President Biden’s memo ended in 2023, we need another updated version of the JFK database to know what’s what. Truer words than I knew.

My two cents

So! We sort of know what is the nature of the fragments of FBI files released in the past. We have a (flawed) theory of why they were released. But really, why should we care? I mean, can’t I just skip the fragment and look directly at the full file to see if there are still redactions in the record, and where they are?

The answer, I now believe, is “yes, you can.” But is there a list somewhere that tells you the record still has redactions, without going through the whole file? It seems the answer is “No, there isn’t.” Remember, the record we are looking at, 124-10185-10099, is not in the most recent FBI TP list.

However, there is also no running count of records released in full for each of the half dozen NARA releases in 2023. I know which of the CIA records is released in full, because I have compared all the 2023 releases with the 2022 releases. So for these FBI records, one must simply do the same.

In fact, I’ve done the same, The FBI “duplicated” releases are the last bit of these records I’ve checked. Based on these recent checks, I can now put up some new figures later this week.

My suggestion that we ignore file fragments in the CIA doc index may bother some. Of course this reduces the number of docs on the CIA doc index that we are looking for. Since these fragments do not actually represent separate records, however, this is a logical result.

Are these fragments the only sort of “duplicates” in the CIA doc index? Are these the only sort of “duplicates” in the various NARA releases from 2017 on? Well, no. Non-FBI “duplicates” are, however, much fewer, and for now I will put these aside, and go have a bottle of Taiwan beer.