Deaths by month this week: verified reports and corrections

A monthly index is rarely typed straight into a monthly page. It is assembled from smaller pages that were published first, and the assembly is where the errors live.

Nobody is named here and no death is reported. What follows is about the mechanics of folding a set of weekly pages into a month.

What to take away

• Weeks and months do not nest. Four or five weekly pages never map cleanly onto one calendar month, and the overlap is where entries go missing or appear twice.
• The reconciliation is a comparison of two records, not a copy of one into the other. If it is a copy, nothing is being checked.
• The most valuable output of the pass is the list of disagreements, because those are the entries that were never as settled as they looked.

Why the two units do not line up

A calendar month has twenty-eight to thirty-one days. A publication week has seven. No month is a whole number of weeks, and no week that spans a month boundary belongs to one month.

That leaves a straddling week at each end of every month, and every straddling week contains entries that belong to the earlier month and entries that belong to the later one. Splitting them requires a rule, and the rule has to be written down, because a person doing it from instinct will split it differently on a Tuesday than on a Friday.

There is a second complication underneath the first. A weekly page is normally organized by the week an entry was published or confirmed. A monthly index is often organized by the month of death. Those are different fields, so the split is not simply a matter of cutting the week in the right place. An entry confirmed in the first week of April may belong to March, or to February, or to no month this year at all.

The four ways an entry goes wrong in the handoff

It is dropped. An entry sits in a straddling week, both months assume the other one took it, and it appears in neither. This is the most common failure and the hardest to notice, because nothing on either page looks wrong.

It is doubled. The mirror image. Both months claim the straddling entry, and the same death appears twice in the year. Doubling is easier to spot than dropping and is usually caught, which is why archives tend to undercount rather than overcount.

It is moved without a note. An entry that appeared on a published weekly page turns up in a different month, with no trace of the change. Anyone who cited the weekly page now cannot find what they cited.

It is upgraded in transit. A weekly entry marked as provisional loses its qualifier during the fold, and reaches the monthly page looking confirmed. Nothing new was learned. The uncertainty was simply lost in the copying.

The pass, in order

The reconciliation is a comparison, and comparisons need both sides present.

1. Take the weekly pages that touch the month, including both straddling weeks, and list every entry on them.
2. Take the monthly page as it currently stands, and list every entry on it.
3. Match the two lists on identity rather than on name. Two people can share a name, and one person can appear under two.
4. Work the differences. Every entry on one list and not the other is either an error or a decision, and each one needs a note saying which.
5. Check the qualifiers, not just the presence. An entry that is on both lists but confirmed on one and provisional on the other has not been reconciled.
6. Write the result down, including the entries you decided to leave out and why.

Step six is the one that gets skipped, and it is the one that pays. A month with a reconciliation note attached can be re-checked by somebody else next year. A month without one can only be rebuilt from scratch.

What the pass cannot fix

It cannot make a weekly page complete, and it cannot make a month final. Registration and publication run on their own timetables, and in most systems the two are separated by days at least: the statutory steps a family goes through, of the kind set out in the official guidance on what to do after someone dies, take time before anything is available to be reported at all. Entries will keep arriving after the fold is done.

Nor can it settle a disagreement between sources. Where a weekly page and a monthly page carry different dates for the same death, the reconciliation records the disagreement and sends it back for sourcing. Picking the more convenient version and moving on is how a wrong date becomes permanent.

Historical practice has the same shape and the same weakness. A researcher reconciling a run of old papers is doing this pass by hand, and the frequency at which a title published is what determines how fine a grain is available at all, a point the Library of Congress guide to newspaper research makes for anyone working back through print runs.

Where this sits

The weekly source pages are described in deaths this week, and the monthly target in deaths by month. The same handoff happens one level down, from daily pages into weeks, and one level up, from months into a year: see deaths today and deaths by year for what each of those units can hold.

Common questions

Why not skip the weekly pages and build the month directly?

Because readers want something sooner than a month, and because a weekly page is a checkpoint. Building the month directly means nobody compares two independent versions of the same period.

Which month does a death at the end of a month belong to?

Whichever month the page's stated rule assigns it to. The answer matters much less than the rule being written down and applied the same way every time.

How often should reconciliation happen?

Once per month at minimum, shortly after the month closes, and again if a batch of late entries arrives. Doing it while the month is open produces a result that is out of date before it is finished.

What if the weekly page and the monthly page disagree and both cite a source?

Then two sources disagree, and that is the finding. Record both, mark the field as disputed, and go back to whoever is closer to the event rather than choosing between two secondhand reports.

Should a corrected entry stay visible on the old weekly page?

Yes. Someone read that page and may have relied on it. A note on the old page pointing at the correction is worth more than a tidy archive.