This is the third note in my “Zombies in the House” trilogy (just pray that it doesn’t turn into a quadrology).1The other two are here and here. Although the title is frivolous, this error in reviewing, or indexing, or whatever it was, has a lesson behind it. I doubt very much whether my note will have more than a couple of readers, but I have done my best to draw that lesson.
Duplicates or ghosts?
Bibliographers from long ago began speaking of bibliographic ghosts. What is a bibliographic ghost? It is when a single book, or pamphlet, or broadsheet, or whatever, is entered into your records twice (or thrice!) under slightly different titles.
In my last post, I discussed 25 transcripts of HSCA interviews/depositions of CIA personnel. It turns out that 24 of these 25 transcripts have duplicate copies listed in the JFK database. The transcripts are listed in the JFK database in two sets: one set has record numbers beginning 180-10110-100XX (call these the A copies), the other set has record numbers beginning 180-10131-103XX (call these the B copies).
The JFK database is missing titles for a bunch of these records, and most of the records are not available online, but we can still see that these different record numbers in fact refer to the same transcripts. Why? because the number of pages and the dates for one set of transcripts match exactly the number of pages and dates for the other set of transcripts.
There are also a handful of online pdfs where we have both A and B copies.
If not for these actual pdfs of A-B pairs, one might wonder whether one of these sets where actually bibliographic ghosts: in other words, there is only a single set of transcripts, but they have been accidentally entered twice in the JFK database.
The A copies of the transcripts LOOK ghostly. Since NARA began releasing most of the withheld or redacted text in 2017, there have been releases LISTED as A copies, but they are not. They are B copies, labelled in the NARA release spreadsheets as A copies, but linked to the same file that represented the B copies.2I sort of noticed this back in 2018 (here), but was not able then to identify the records causing the problem.
However, it is not possible that the A copies are ghosts. First and foremost, we have pdfs for both an A copy and a B copy in several cases. Where are these A copies? They are in the Mary Ferrell Foundation’s online collection.
The A copies in the MFF collection
MFF has A copies of the HSCA transcripts that go way, way back. They were some of the first JFK docs the MFF website put online, beginning in 2007.3According to the Archive.org Wayback Machine.
Some of these old pdfs are themselves scans of releases dating back to the 1993-94. We can tell because they use a very old identification aid on top of the file. Instead of the now standard RIF form, they use another form that has a date of 1993 on it. In addition, there is an HSCA original cover sheet on a few of these records that does not appear in any of the later B copy releases.
But if you check these A copy transcripts, they are in fact word for word identical with the B copy transcripts. Except for one thing: they have different redactions. And some of these redactions reveal information that the CIA still wants to redact in the 2022 releases.
This is where “zombie redactions” come from. Here are some:
Nosenko deposition 2 (1978-06-20)
In the second HSCA interview of Yuri Nosenko, the 2022 release of the B copy has redactions on four pages. The A copy at MFF releases all this redacted material except for one word.
Phillips deposition 2 (1978-04-25)
In the second HSCA interview of David Phillips, the 2022 release of the B copy has one redaction left, on pg 7 of the pdf, pg 4 of the typescript. The A copy at MFF releases part of this redaction, concerning Phillips moving his office from Rio de Janeiro to Brasilia (h/t to Tracy Parnell for identifying the location code for Brasilia).
The Murphy deposition (1978-08-09)
The 2022 release of the B copy is finally released in full. Prior to this, the A copy at MFF released material that was redacted in 2018 and earlier NARA copies.
The Tovar deposition (1978-06-29)
The 2022 release of the B copy has redactions on 2 pages. The A copy at MFF releases all the redactions on one of those pages.
This doc was discussed on a post at Tracy Parnell’s blog and another post at Fred Litwin’s blog. They were checking a post at Jefferson Morley’s JFK Facts which observed a long redaction in one release of the Tovar transcript without noticing the more complete A copy at MFF.
In fact, Morley got this wrong. Although his post was supposed to be discussing redactions in 2022 releases, the redacted copy of the Tovar transcript that he cited was actually the 2018 release, a rather careless error on Morley’s part. The 2022 release of the B copy also releases the passage he cited.
The Bustos deposition (1978-05-19)
The 2022 release of the B copy has redactions on 2 pages. The A copy at MFF releases all the redactions on one page, and one of two redactions on the other. There is thus only one real redaction on the 2022 B copy, the others are zombies.
The Shaw deposition (1978-05-16)
The 2022 release of the B copy has redactions on 5 pages. The A copy at MFF releases all redactions on three of these pages, and part of one on a fourth. So one full redaction has not been revealed yet, and part of a phrase is also redacted.
The Shaw deposition was another record cited by Morley to show that important information was still being redacted. Fred Litwin’s post, cited above, noted that the redaction Morley referred to was released in the A copy at MFF. This time, at least, Morley was correct that the 2022 release of the B copy had a redaction here. Unfortunately, he did not realize it was a zombie redaction, not a real one.
The signficance of redactions
So the A copies on MFF reveal some of the material still redacted in the 2022 B copies. How do we evaluate the fact that this material was supposed to be held back at least until six months from now? Does it just go to show how the redactions are hiding less than earthshaking material? Some of the redactions in fact fit into the categories that CIA has defined to justify its redacting:
As an example, the A copy shows us that the Nosenko zombie redactions are about Nosenko’s part in U.S. liaison activities:
N: Yes, I have a passport, which I once used being sent by the Agency in 1975, in England, for week where I was talking with British Intelligence and Counterintelligence Services, concerning my knowledge of the Second Chief Directorate, KGB in general.
Q: Is that the-only country to which you have traveled since your defection?
N: No. I have traveled on behalf of the Central Intelligence Agency also for the last four years. I was in eleven countries. And I never received or demanded or asking any special fees with all friendly intelligence services, never.
Several of the redactions are employee names, another troubled and troubling subject. In general, there are quite a few new names (to me) in the 2022 releases, but this is a subject I will write about separately.
Regardless of whether they fit into categories or not, however, if the information has already been released online, and has been there for over a decade, the redactions should just automatically be dropped at the earliest opportunity. There is no excuse for anything else.
My two cents
While we are only talking maybe two dozen redactions, I think it is worth thinking about why this happened. Poor indexing was surely a primary reason. The fact that it was quite difficult to tell that there were two duplicate sets of transcripts must have made it easy to overlook how few A copies were actually released and to forget about the ones that the ARRB actually processed.
There are many A copies of the transcripts that have never been released. And what is particularly bad is that a couple of the A copies on MFF have ARRB decision sheets on top. What is an ARRB decision sheet? It is a record of how many redactions were allowed to continue and why. The CIA should not need NARA to remind them about documents which have these things attached to them. They should have them on a list. They DO have them on a list.
So they’re duplicate docs, so what? The ARC is teeming, crawling with a vast number of duplicates. Process and release is the only way to finish this otherwise interminable process.
Don’t worry about finishing processing the duplicates and junk you say? Surely you jest. The nation has spent far too much time and energy on this to blow it off at the very end.