This note takes up some general issues in the ARC, but since it looks at them in the context of the latest information on redactions, call it my sixth note on NARA 21, the updated version of the JFK ARC database (available here). NARA 21 incorporates the 2017-2018 ARC releases, and is therefore essential when looking at redaction issues.

The point I make in this note is that in looking at ARC redactions, it is also essential to look at duplicate records. Unfortunately, one of the more intractable problems in the JFK ARC is the large number of duplicate records it contains.

Duplicate records in the ARC

If there were a competition for duplication, FBI records in the ARC would certainly take first prize. There are multiple copies of big chunks of the massive FBI files on Oswald and JFK assassination records incorporated everywhere.

CIA records in the ARC also have a high rate of duplication, though on a smaller scale, and this created a major headache for CIA when they went to redact Agency information. The result has often been inconsistent redaction in many CIA records.

The Assassination Records Review Board (ARRB), the federal agency charged with collecting and releasing the JFK assassination records, was aware of the problem of duplicate records from the beginning of its tenure. The topic came up at the Board’s first meeting, on April 12, 1994.¹A copy of the transcript is available at the MFF website (here). The most extended discussion at that meeting comes between Steve Tilley, NARA liaison with the ARRB, and Board member Anna Nelson:

MS. NELSON: I have a couple of questions. One is, Steve, a lot of this must be duplication. Because the CIA must have in its collection what it sent to the House Assassination Committee?

MR. TILLEY: That’s correct, there is.

MS. NELSON: So that there is a lot of duplication here which I assume gets straightened out in the database.

MR. TILLEY: Yes. I think that was one reason for the database, was the attempt to deal with the massive duplication of items within this world, the collection itself.

MS. NELSON: That’s right.

MR. TILLEY: Yes, you are correct, there is duplication. Each investigative agency or group, if you will, that has looked at the assassination went out to the various agencies which had documents and asked for copies of those documents. You have several sets of everybody’s files in the holdings of each assassination agency, depending on how widespread their calls were.

Of course, in many instances not only did those groups ask for records from the agencies, but then they would turn to the National Archives and ask for records from the records of the Warren Commission which would then produce extra copies of those same documents again. Yes, there is a very, very large problem of duplication throughout these files.

MS. NELSON: It comes out in the database?

MR. TILLEY: Sure. Yes, we are able to find, we can search on all fields of our database. The question of duplication is solvable through that but, you know, it is still a big problem.[END QUOTE]

There’s more that could be said about this exchange; the JFK database is by no means an efficient solution to this very big problem. This note has other fish to fry, however, so I’ll leave this topic for later.

CIA record duplication

The topic of duplicates came up again in an August 6 1996 presentation to the Board by John Pereira and Barry Harrelson of CIA’s Historical Research Group.²The transcript for the meeting is available at the MFF website (here). The context here is a discussion of the “hard-copy” and microfilm versions of the CIA’s JFK documents; Pereira is speaking with Board member William Joyce:

DR. JOYCE: Mr. Pereira, you mentioned that 50 percent of the microfilm appears to duplicate what already exists in hard copy. You mentioned the [Oswald] 201 file as an example of that. Maybe I missed it, but could you characterize perhaps the portion that appears not to be duplicated in hard copy? Do you have any information about that?

MR. PEREIRA: The types of material that isn’t duplicative?

DR. JOYCE: Yes.

MR. HARRELSON: This is where you get into the 201 files which make up the bulk of the microfilm, and these files cover individuals who were not — or areas of their careers that were not involved in the assassination periods. So, there would not be documents from these files in the hard copy.

Most of the hard copy, apparently, was created as the HSC staffers asked for particular information. They would look at a file, and ask for copies of it. The ultimate was one cable, where we found 43 copies. So, most of the documents — I would say, every document is duplicated at least once, and used multiple times throughout the collection.[END QUOTE]

There’s even more that could be said about this exchange, though it makes an important point in observing that there is massive overlap between the microfilm and hard copy segregated CIA collection. Again, this is a topic I’ll leave for later.

The 43 duplicate cables

To illustrate the complexities that the large number of duplicate docs can create, one might well start with the example Harrelson gave, the cable with “43 copies”.

Cables are most conveniently identified using the cable number, consisting of the cable prefix and sequence number. The cable prefix indicates the sending station and is usually four letters (all caps). It is followed by a number indicating the sequence in which it was sent.

Harrelson’s cable is probably MEXI 6453. This is a cable originated by the CIA station in Mexico City, hence the MEXI prefix. It is also identified as number 5 in the CIA’s JFK docs series (actually number 5-1a).³This series was put together in the 1970s and has a very complicated history, too complicated to get into here. Two frequently cited exemplars of MEXI 6453 are ARC 104-10015-10047 and 104-10015-10304.

In fact, I have found more than 43 copies of this cable: 46 to be exact, and this is just looking in the ARC’s CIA docs. It also appears in both the SSCIA (Church Committee) and HSCA records. And to complete the fun, the MFF website has multiple copies of many of these records.

An excel file listing all these record numbers, with links to the MFF copies on line, is available here.

Note that these are not just straight copies of MEXI 6453. There are multiple versions of this document. First, there is a version that says “CS copy” (standing, I think, for Cable Secretariat). Most of these are stamped with the “5-1a” doc series number. I have found 24 copies of this version. This means they occur as 24 different records in the ARC, each with its own unique ARC number.

Second, there is a version labeled “commo-chrono” in the upper left corner (call it the “CC copy”). Most CC copies have hand-written on them “DUP OF 5-1a”. The CC copy also comes in two versions, an unannotated version, and a second version with a note attached to it which states that the “originating officer” is the same as the “authenticating officer” listed on the bottom left side of the cable.⁴For a brief note on the format of CIA cables, see here. I have found nine copies of the unannotated version and seven copies of the annotated version.

Third, in addition to the copies which indicate the cable is JFK DOC 5-1a, either by stamping or handwriting, there are five more copies that don’t mark it at all, and one copy that says “SAME AS 5-1a”.

Remember, when I say copies I mean that each of these documents has a different record number. That means each comes from a different source in the vast pile of paper that is the CIA segregated record collection.

Each of these records also has its own release history. The CS version of MEXI 6453 was released in full in Aug. 1995. This was in the third batch of CIA docs that the ARRB voted to open in full. MEXI 6453 DUP was released in the same batch.⁵The copies released were 104-10015-10047 and 104-10015-10304. The other copies were released at different times; I will spare you the long and incredibly boring details.

Both the original cable and the unannotated DUP cable show that both the “originating officer” and the “authenticating officer” was LADILLINGER. The note attached to the annotated version was redacted in various degrees in the seven copies of this document. The current versions all show: “Originated by L. A. Dillinger (P) wife of [xxx]”. ⁶(P) means the name Dillinger is a pseudonym, not the real name of the officer. Readers familiar with John Le Carre’s books will recognize this as what the MI6 calls a “work name”.

There is, however, one exception to this: ARC 104-10086-10015. This annotated copy of MEXI 6453 DUP was released in full in July 2017.⁷Readers may check out the name for themselves. This is not a brand new revelation, poking around on-line will turn up discussions of these husband and wife officers by people such as Bill Simpich, who got the information from other records, such as HSCA and SSCIA interviews.

As a result, even though the other six annotated copies have not been released, we already know their redacted content. This is what I mean by inconsistent redaction. I am sure that the intent was to finally release the redacted text of the annotated MEXI 6453 DUP. At some point in the release process, however, the redactors lost track of the connection between 104-10086-10015 and the other six duplicates.

This is by no means the only such case. There are dozens of similar records, where a redaction was released in one record, but not released in another record that is actually a duplicate of the first. Sometimes, as in this case, it was the last redaction remaining, meaning that records listed as redacted are actually open in full. In other cases, the number of redactions remaining is reduced by however many were opened in the first doc.

In either case, these records are primary candidates for formal release in the next round of removing redactions. This also illustrates a point I made in earlier notes in this series: the large number of redactions remaining in CIA records doesn’t indicate a large number of different names remain. Not when releasing one name opens seven records in full.