Applied The Archive Playbook

Loss and Burial Chain: What the Records Report

Updated

Intermediate

What you'll learn

~45 min
  • Ask an AI tool, in plan mode, for a page that lays out the chain of records about a loss and what followed it
  • Check the plan, before anything is built, for what a careful researcher needs from such a page
  • Tell a place of memorial, where a name is recorded, from a place of loss, and read the honor roll's 'place of loss' against the loss report
  • Catch a summary, your own or an index's, that uses a word the record behind that step does not use, and replace your own with the record's words
  • Keep 'recovered', 'identified' and 'accounted for' apart, and point to the one page in a file that makes an identification

The story, and what you will end up with

The family story says Walt’s name “is on a memorial in India, and after the war he came home and is buried at Norrlid.” The casualty file you requested in February arrived on 28 March 2026. A casualty file, in this case, is the file of reports and letters about one service member’s death kept by the responsible authority: the official body in this case whose file it is and which identifies the dead. In lesson 08’s page sorter you sorted its twelve pages and labelled each one: a report of burial, a report of disinterment (the taking up of a burial from its grave so it can be moved), a report of interment (interment is another word for burial), an identification memorandum, a shipment record, a report of final interment.

From those pages you have written your own notes: seven steps, from the loss of the aircraft to the final burial at Norrlid, each with the record and page it rests on and a short summary in your own words. You also have two records that are not steps: a memorial plaque the squadron put up at Nohorkhat airfield in June 1945, and the Tamarack County honor roll, an online index typed by volunteers in 2009. The honor roll is the easiest of these to find, and it reads like a summary of the whole story: “Died 22 Mar 1945”, “Place of loss: Nohorkhat Airfield Memorial”, “Remains recovered 1946”. Each phrase sounds official. None of them is how the casualty file puts it, and one of them has already crept into your own notes.

By the end of this lesson you will have a page on your own computer that lays out the chain step by step, each step with its record’s own words, and points out where your summaries or the honor roll say something the records do not. You build it by asking your AI tool in plain English, and then you read it.

Many families’ files do not end as this one does, with a grave at home; a real file may stop at any step.

Before you start

First, if your AI tool is still open, leave it: type /exit and press Enter (if you use Antigravity, press Ctrl+D with nothing typed instead), and wait until you see the $ or % prompt. The next line goes there, in the terminal, not in the AI tool. It makes a folder for this lesson and goes into it:

Terminal window
mkdir -p ~/archive-playbook/09-loss-chain && cd ~/archive-playbook/09-loss-chain

Start your AI tool with the start line you kept from the setup lesson, the one with approval prompts off. Then ask it to set the folder up:

Set this folder up so we can undo any change you make, download https://learn.intellimetrics.net/starters/history/09/loss-chain.json into a data folder, and save a checkpoint.

Look for it saying three things: the folder is set up with Git (the program that saves snapshots, from setup), the file is saved as data/loss-chain.json, and a first snapshot, your checkpoint, is saved. If it does not mention one of them, ask: “Did you save a checkpoint with the file in it?” If the download does not work, download loss-chain.json in your browser, check where it went (the browser’s list of downloads shows the folder), and tell your AI tool, for example: “loss-chain.json is in my Downloads folder. Copy that one file into a data folder here, and save a checkpoint.” In Windows PowerShell or a desktop app, “my Downloads folder” works as is; only say “my Windows Downloads folder” if you use Ubuntu (WSL) on Windows, which keeps its own Downloads folder apart from Windows.

Everything in the file is invented for this course: every person, record, place and repository in it, and nothing in it describes a real loss. The casualty file’s page titles were written for the course too, and are not offered as real form titles.

Open this box once, because it says what each code used further down stands for.

ℹWhat is in loss-chain.jsonclick to expand

The person as the records name him, ENGSTROM, Walter O., TRAINING-SN-204583, then three parts.

Ten records, each with its title, its reference, its record date (the date the record was made, which is not always the date of what it reports) and its full text, line by line:

CodeWhat it isRecord date
R-19Squadron operations log, 21 March 19451945-03-21
R-20Loss report for the aircraft, with its crew list1945-04-05
R-24 p05Casualty file, page 5: Report of Burial1945-04-02
R-24 p07Casualty file, page 7: Report of Disinterment1946-02-14
R-24 p08Casualty file, page 8: Report of Interment1946-02-20
R-24 p09Casualty file, page 9: Identification Memorandum1946-06-11
R-24 p11Casualty file, page 11: Shipment Recordnot given
R-24 p12Casualty file, page 12: Report of Final Interment1948-05-22
R-25Nohorkhat Airfield Memorial plaque, its inscription1945-06
R-26Tamarack County honor roll, online index, no source cited2009

Your six places, each with the kind you gave it (place of loss, of burial, of memorial, or other):

CodePlaceYour kind
PLACE-1Lonsaing drop zone, BurmaPlace of loss
PLACE-2Isolated grave west of Wetkyauk village, BurmaPlace of burial
PLACE-3Nohorkhat Temporary Cemetery, Assam, IndiaPlace of burial
PLACE-4Norrlid Cemetery, Kettle Ford, WisconsinPlace of burial
PLACE-5Nohorkhat Airfield Memorial, Assam, IndiaPlace of memorial
PLACE-6Nohorkhat to Kettle Ford (shipment)Other

Your seven steps, each with its event date (when the step is said to have happened), its place, the place as the record names it, the record behind it with the words you quoted from it, and your summary:

StepWhatEvent dateRecord
C-1Loss1945-03-21R-20 and R-19
C-2First burial1945-03-28R-24 p05
C-3Disinterment1946-02-14R-24 p07
C-4Transfer1946-02-20R-24 p08
C-5Identification1946-06-11R-24 p09
C-6Transfer1948-04-28 to 1948-05-19R-24 p11
C-7Final disposition1948-05-22R-24 p12

Last, a note that R-25 and R-26 are other records, not steps. Here is a reference example of one step in the JSON file your agent downloads and maintains (you do not type or edit this JSON yourself):

{
"step": "C-3",
"name": "Disinterment",
"event_date": "1946-02-14",
"place": "PLACE-2",
"place_as_the_record_names_it": "isolated grave, about one and one-half miles west of Wetkyauk village, Burma",
"sources": [{"record": "R-24 p07", "quotes": [
"Disinterred from: isolated grave, about one and one-half miles west of Wetkyauk village, Burma",
"Transferred to: Nohorkhat Temporary Cemetery"]}],
"my_summary": "Remains recovered from the isolated grave."
}

Switch to plan mode

Your AI tool is already open. Switch to plan mode now: Shift+Tab until it says plan, or /plan (the keys for each tool). In plan mode it usually reads your file, asks its questions and shows you a plan before it builds anything; the order can vary.

The conversation

Each message below is an example. Say it your own way.

Say what you want

Say what is in the file, and what you want from it:

data/loss-chain.json has my notes on Great-uncle Walt's loss and burial, with the records behind each step. Help me build a page that lays out the chain and points out anything that disagrees. Ask me what you need first.

Your AI tool reads the file and comes back with questions. It may also list disagreements it has already noticed, such as the two distances given for the first grave. You do not have to act on those now.

Answer its questions

Answer each in a sentence. These are questions AI tools often ask, with one way to answer each. Yours may ask them in a different order or show a list of choices:

If it asksYou might say
Who is it for: you, or relatives too?Just me for now.
A page in your browser, or something else?A page I can open in my browser; it only needs to show, not edit.
Only show disagreements, or also say which record is more reliable?Show disagreements side by side and never say which record is right.
Should it also flag problems in your own notes?Yes, flag my own notes too.
The plaque and the honor roll: their own section, or beside the steps?Keep the plaque and honor roll in their own section.
Anything else you want on it?That’s enough.
Should the word check flag small words like “the”?Skip small words like “the”; only flag words that change what happened.

If it asks several questions at once, answer the ones that matter to your research in a sentence, and for the rest say that its suggestions are fine. If it asks something not listed here, answer it the same way, in a sentence.

Read the plan

Next it shows you a plan: what it will build and how it will check it. Read it for six things:

  1. Every step is tied to its record and page, in event-date order, with both the date of the step and the date the record was made.
  2. Each place is given as the record names it, in the record’s own words, beside the kind of place you gave it.
  3. The plaque and the honor roll are kept apart from the chain. They mention these events; they are not steps in it.
  4. Where records disagree, they are shown side by side, each in its own words, and nothing says which one is right.
  5. Any word in your summaries or the honor roll that the record behind that step does not use is pointed out, with the record’s own words beside it.
  6. Your file is read and left as it is.

A first plan often leaves out the fifth. Say so in a sentence or two, in your own words, for example:

Also point out any word in my summaries or the honor roll that the record behind that step doesn't use.

Your AI tool may then show you what such a check finds on your file and ask how strict to be, for example whether small words like “the” count. Answer it in a sentence; the table above has one way.

When the plan has all six, approve it: if your AI tool offers a choice to go ahead, pick it; otherwise say “Yes, go ahead and build it.” In Codex, press Shift+Tab first to leave plan mode. In Copilot, saying it brings up a choice to go ahead; pick that. Let it build. When it has finished, ask it: “Did you install anything, and where?” Nothing, or something inside this lesson’s folder, is fine for these lessons. If it names anywhere else on your computer, ask it what that is for, and to remove it if you do not want it there.

Look at what it built

Ask it to open the page:

Open the chain page in my browser.

Your AI tool may open the page straight from its file, or start a small program on this computer that serves the page at a local address, one ending in a number, such as http://127.0.0.1:8000/ (127.0.0.1 is an address that means this computer); either is fine. If it gives you an address without opening it, copy the address into your browser’s address bar and press Enter. If it says the address is already in use, say: “Something else is using that address. Don’t stop it. Use a different address for my page.” Leave the program running while you work; when you are done, ask your AI tool to stop it. If your AI tool does not take a message after opening the page, wait a minute. If it still does not, close the terminal window, open a new one, type this lesson’s folder line again, start your tool, and say: “Start the page again so it keeps running while we talk.” The page and your saved work are files in the folder, so nothing is lost. Your page will be laid out in its own way, and some parts may be folded shut until you open them.

Check it for these. If one is not so, tell your AI tool what you see and what you expected, for example: “The honor roll is shown as a step in the chain. It should be in its own section.”

  • Seven steps, in this order: the loss on 21 March 1945, the first burial on 28 March 1945, the disinterment on 14 February 1946, the transfer to Nohorkhat Temporary Cemetery on 20 February 1946, the identification on 11 June 1946, the shipment from 28 April to 19 May 1948, and the final burial at Norrlid on 22 May 1948. Each shows its record and page. Your page may rename a step (the shipment as “Shipment” instead of “Transfer”); that is fine if it also shows your own name for it.
  • The plaque, R-25, and the honor roll, R-26, appear apart from the chain, not as steps.
  • The honor roll’s “Place of loss: Nohorkhat Airfield Memorial, Assam, India” is shown beside the loss report’s own words, Mission: supply drop, DZ Lonsaing, Burma (DZ is short for drop zone), and is not taken as where the loss happened.
  • “recovered” is pointed out twice: in your summary of C-3, and in the honor roll’s “Remains recovered 1946”. Beside each, the page shows what R-24 p07 says: “Disinterred from” and “Transferred to”.
  • The first grave’s two distances are shown side by side: “about one mile west of Wetkyauk village” on p05 and “about one and one-half miles west” on p07. Neither is picked.
  • The dates are shown side by side: 21 March 1945 in R-20 and on p09, “Died 22 Mar 1945” in the honor roll. Neither is picked.
  • The identification step shows p09’s own words, “The responsible authority has identified the remains…”, and nowhere in its own words does the page say that Walt was recovered or identified.
  • The shipment record, p11, is shown with no record date of its own, not one borrowed from another page.
  • When you ask “Has my data file changed since the checkpoint?”, your AI tool compares it with the checkpoint and says no.

Your page may point out more than this. It may show that the honor roll’s service number, TRAINING-SN-204533, differs from the file’s TRAINING-SN-204583 (lesson 06 is where you weighed that number), or that the report of burial is dated 2 April 1945, three days before the loss report that still lists Walt as missing. Both are true of the records. Neither changes the list above.

Read the chain

Work through what the page shows, in this order.

  1. Two dates on every step. The loss is dated 21 March 1945, but the loss report that gives it, R-20, was made on 5 April 1945, fifteen days later. That is lesson 07’s rule that record date is not event date, on one step.
  2. The place of loss that is a memorial. The honor roll gives “Nohorkhat Airfield Memorial” as the place of loss. You marked that place, PLACE-5, as a place of memorial, and R-25 is its plaque: an inscription the squadron put up in June 1945, a place where a name is recorded. The loss report gives a mission, a supply drop at DZ Lonsaing in Burma, and a last known position, “second pass over DZ, about 1340”. So the careful reading: the honor roll’s “place of loss” is a memorial, and the loss report does not put the loss at Nohorkhat. Notice too that your own PLACE-1 is your reading of R-20’s mission line, not a place R-20 gives for the loss; some pages point that out as well.
  3. Two distances. p05 and p07 describe the same isolated grave, one mile and one and one-half miles west of Wetkyauk village. A hasty reader picks one, or calls the other a typing error. A description in a record is that record’s words, not a position, and the page keeps both.
  4. Two dates of death. R-20 and p09 give 21 March 1945; the honor roll gives 22 March. The 22 March matches the day the squadron’s morning report, the daily record in which a unit clerk noted each change in a man’s status, carried the crew missing, that is, entered them as missing in action; it is not the date of the loss the loss report gives (lesson 07 shows that report). The page shows all three and decides nothing; the weighing is yours, as in lesson 05.
  5. The identification. Read p08 and p09 together. p08, the report of interment at Nohorkhat on 20 February 1946, names Walt “(pending identification)”. p09, dated 11 June 1946, nearly four months later, says “The responsible authority has identified the remains”. In this file, identification is its own step, made by the responsible authority, on its own page, and no other page makes that statement.

Notice what one word does

Find your summary of C-3 on the page: “Remains recovered from the isolated grave.” Before you read on, compare it with what p07 says. Which of your words is not on the page?

p07 says “Disinterred from” and “Transferred to”. “Recovered” is the honor roll’s word. It reached your notes the easy way, copied from the index because it read like a record’s word. Now see what changes when your summary uses the record’s words. Ask:

Make a copy of my data file in this folder. In the copy only, change my C-3 note to 'Disinterred from the isolated grave; transferred to Nohorkhat Temporary Cemetery.' Then tell me what the chain page flags for the copy.

Look for it saying that on the copy “remains” and “recovered” are no longer flagged in C-3 (a small word such as “the” may still be), and that the honor roll’s “recovered” is still flagged. If it changed your own file instead, say: “Put my data file back the way it was at the checkpoint.”

You can correct your own words; you cannot correct a record. The honor roll says what it says, and the page keeps showing p07’s words beside it. That does not make the honor roll wrong. It may be using the word loosely. It means the word is the index’s and not the file’s, so your chain keeps the file’s words. And the three words are not loose synonyms: in the US, the agency that accounts for missing service members counts a person as “accounted for” only if they are returned alive or their remains are “recovered and identified” (its own wording is quoted under How real archives differ). A word that means “found” is not a word that means “identified”, and neither is a word for the whole outcome.

Recovered, identified and accounted for are three different words. Only the responsible authority identifies.

Now remove the copy:

Delete the copy now, and save a checkpoint.

One real run

Here is part of one chain page, built with Claude Code on 10 November 2026. Yours will be laid out differently; check it against the list above, not against this.

Each record in your file carries a kind, the terms from lesson 02: “Original” is a record in its first form, and “Derivative” is a later version made from others, such as a typed index. The page shows them beside each record.

Place of loss
Value Record Exact text
DZ Lonsaing, Burma R-20 Original Mission: supply drop, DZ Lonsaing, Burma
Nohorkhat Airfield Memorial, Assam, India R-26 Derivative Place of loss: Nohorkhat Airfield Memorial, Assam, India
Words in my summaries that the cited record doesn't use
Step My note Flagged words Checked against
C-2 First burial in an isolated grave near Wetkyauk. first, near R-24 p05
C-3 Remains recovered from the isolated grave. remains, recovered R-24 p07

Fictional training example.

Look at the first row. The page lists DZ Lonsaing under “Place of loss”, with R-20 beside it, as if R-20 gave that as the place of loss; R-20’s own words in the same row are a mission line. That is exactly the kind of wording to catch, on a page you built as much as in an index. If your page does it, tell your AI tool: “R-20 names Lonsaing as the mission, not as the place of loss. Label that row with R-20’s own words.”

The same page flagged “first” and “near” in your C-2 summary. Those are your own fair words for a first burial near a village; a flag asks you to look again, and you decide whether the record bears your words out.

How real archives differ

The plaque and the honor roll in this case are invented. The record types and the three words are real:

  • Casualty files. The US National Archives describes Individual Deceased Personnel Files as “Also known as Mortuary or Casualty Files”, whose documents “will vary based on branch of service and time period but may include correspondence, memorandums, teletype messages, and other documentation relating to the death of individual personnel during service” (National Archives at St. Louis, Frequently Requested Records). To locate a real file, that page asks for the “Full name of service member”, “Service number”, “Dates of service” and “Date and place of death”, and tells you to “Check which records are already on the Catalog”.
  • Accounted for. The US Defense POW/MIA Accounting Agency (DPAA) states its mission as “Provide the fullest possible accounting for our missing personnel to their families and the nation” (DPAA, Vision, Mission, Values; Internet Archive copy, 8 October 2026), and says: “According to Section 1513, Title 10, United States Code, and Department of Defense (DoD) policy, unaccounted-for DoD personnel from past conflicts may only be “accounted for” if they are returned alive or their remains are recovered and identified” (DPAA, FAQs; Internet Archive copy, 12 July 2026). The definition joins its two words with “and”: remains that are recovered but not identified do not make a person accounted for.
  • Memorials. The American Battle Monuments Commission (ABMC) says: “When a missing service member is recovered, identified and finally laid to rest, ABMC places a rosette beside their name on the Walls of the Missing” (ABMC news item). The wall keeps the name; it does not say where a loss happened. Its listing has limits: ABMC’s database “contains only the names of those interred or memorialized at ABMC cemeteries and memorials. The database does not contain the names of war dead returned to the United States for permanent interment at national or private cemeteries” (ABMC FAQ). A man the file reports as returned home for burial, as this case’s file does for Walt, would not be expected in it. And a date on such a listing may be an administrative one: the same FAQ says “Without confirmed information to the contrary, a War Department Review Board established the official date of death of those missing as one year and a day from the date on which the individual was placed in missing status.” In this case that rule does not explain the honor roll’s 22 March (see the two dates of death under Read the chain).
⚠What your page and this course do not do

The page you build and this course do not identify anyone, do not rank or screen candidates, and are not a route to locating or visiting a site. For families, the US agency names a route. Its contact page (Internet Archive copy, 10 June 2026) says: “Service Casualty Offices serve family members. Each military department maintains a service casualty office.” It adds: “The officials in these offices serve as primary liaisons for families concerning personnel recovery and accounting.” This course gives no other contact route, and nothing you build or find with it is a reason to approach a family, or an agency, with a claim about who someone is.

Keep going

Small changes do not need plan mode. Say each in a sentence, then look at the page again:

  1. “Change my C-3 summary to say what p07 says, and rebuild the page.” This time it is your own file you are changing, on purpose. The flag on your summary goes; the one on the honor roll stays.
  2. “Add a step to my notes: Family told of the first burial, with no date and no record.” The letter from Walt’s mother, Ruth Engstrom, asking in May 1945 for news of him, says only that a Report of Burial “would follow if anything was learned”; it is page 3 of the casualty file (p03), which you sorted in lesson 08 and which is not in loss-chain.json because it is not a step. No page you hold shows the family being told. Look for the new step marked as having no record, with nothing filled in for it. If your AI tool fills in a date, a place or a record, say: “That step has no record. Leave it empty and say so.”
  3. “Give me a version I can print, without my summaries.” A chain about a death can hold details a relative has not seen. Read a printed chain yourself first, and decide what to share and with whom.
  4. “Save a checkpoint.” Your AI tool saves a snapshot you can return to (people who use Git call this a commit).

The finished page holds only this practice chain. When you have your own relative’s records, build the page on them yourself, and use it only for your own relative, from records you or your family hold. It runs on your real records by itself, with your AI tool closed. Do these steps in order; the first two happen now, while this folder still holds only practice data:

  1. Go through the data check, the questions in setup about what the material is, who decides how it is used, and which AI tool would see it. Go on only if it allows this use.
  2. While this folder still holds only the practice file, ask your AI tool: “How do I build the chain page myself from a data file in another folder, and save the page beside that file?”
  3. Write down what it tells you.
  4. Ask your AI tool to stop the page’s program, then leave the tool: type /exit and press Enter (if you use Antigravity, press Ctrl+D with nothing typed instead).
  5. Keep your real file in a folder of its own, outside this course folder, and do what you wrote down. The page is saved beside your real file, not in the course folder.

Try another case

Before the quiz, write two or three sentences for yourself. An online family tree says of a cousin’s great-uncle: “Remains recovered and identified 1950.” The only 1950 page in the file the family holds reads: “Transferred to: town cemetery, plot 4.” What would your chain say for 1950, and what would it leave open?

A good answer gives the 1950 step in the page’s own words (“Transferred to”), points out that the tree’s “recovered” and “identified” are not on that page, and leaves the identification as a step with no record held, since only the responsible authority’s own statement could fill it.

Key takeaways

  • Ask briefly, and let your AI tool ask back. Say what is in the file and what you want from it; its questions fill in the rest.
  • Read the plan as a researcher: check what it will do with records, places and disagreements, not how it is coded.
  • A memorial records a name: the honor roll’s “place of loss” at a memorial does not say where the loss happened.
  • Use the record’s word, not a summary’s: “Disinterred from” and “Transferred to” are what p07 says; “recovered” was the index’s word, and you can correct only your own.
  • Identification belongs to the responsible authority: a researcher, a family story, an index and a page you built do not make one.
KNOWLEDGE CHECK

You are building a chain for a different serviceman. A regimental association website says: 'Killed near the river crossing; remains recovered and identified 1947; commemorated at the county war memorial.' In the casualty file you hold, a 1947 page reads 'Exhumed from: field grave near the ferry. Moved to: temporary cemetery, plot 6.' and a 1948 page reads 'The responsible authority has identified the remains in plot 6.' What does your chain say for 1947?

What’s next

The chain ends at Norrlid in 1948. Some of what you know comes from images: the snapshot of Walt by the barn, the photograph of the memorial plaque, and the aerial photographs of the family’s farm. Next, in the Photograph and Scan Provenance Catalog, you record where each image came from, keep an archive master apart from its enhanced copies, and date a photograph by a range instead of a single day.

intellimetrics · Arlington, VA
© 2026 intellimetrics.net
Search lessons