Choice of Layer 1 - census vs. the site that delivers the image

Hello, Alex. Welcome.

The reality we deal with as historians is this:  Historical records were not created and organized according to a universal format from which no one could deviate. The archives that house those records today do not have one immutable pattern that’s followed worldwide—or even by every archive within one nation. The websites that are created to deliver those records online are not built on one sole architectural structure from which no IT engineer can deviate.

Within the records we use, there’s an infinite variety of situations. Therefore, there cannot be one rigid pattern that must be used for any category of records or any country. One rigid pattern is possible with published books and articles.  That is not possible with historical records, whether we use them as manuscripts, microfilm, or digital images online.

EE provides examples for handling the most common situations.  It cannot say: When Conditions 1, 2, or 3 exist at Site X, Y, or Z, then this is the exact formula to follow.  It cannot say “when using this census from this country, it must be done this way.” Any such dictate would be unworkable.

As researchers, we use manuals such as EE to inform ourselves about these original manuscript records and digital deliveries—along with their quirks and the way deviations might be handled.  Then we study each specific record we use and decide what approach works best for the record set we are using.

As a working example, how about this:

(1)

• Choose an entry from the 1901 Wales census, as delivered by FindMyPast.
• Construct two First Reference Note citations for this entry: one with the census in Layer 1 and one with the census in Layer 2.

(2)

• Then do the same for the 1901 Wales census, as delivered by Ancestry.

That's the best way to see what works, what doesn't and why.  If you want to extend the experiment, try this:

• Choose an entry from any two censuses, from two different countries, imaged at FamilySearch (where all our readers have free access).
• Pick one census that is in a named database and one that is not.
• Then construct two First Reference Note citations for each: one with the census in Layer 1 and one with the census in Layer 2.  

After this exercise, if you care to share the results with us, with your thoughts and additional questions, I suspect the further discussion would help all our readers.

Hello, Alex:

It’s good to know you don't shrink from challenges!  In response to your Comments 1–10:

Comment 1.

The exercise you have done underscores two things:

1. Often the sequence of layers does not matter. Sometimes, as we’ve discussed, the architecture of the website accommodates one approach, but not the other—in which case we have no choice. But in other situations, both approaches work.
2. As researchers, we also need flexibility because of the nature of the research we’re doing—as explained in EE4’s 3.16 (Sequence of Layers: Online Images).

EE4 has tried to provide examples of all of this, using a variety of records. Without growing to the size of the OED, it’s not possible to demonstrate every permutation for every type of record throughout the world. Our 632 pages accommodates only a sampling.

Comment 2.

Permanence is still a pipe dream for anything online, IMO.

Comments 3–5.

Logical, all.

Comment 6.

The colon works. Citations, like narrative writing, have general rules but also need flexibility—especially when working with complex materials. Citation guides that specialize in standard published books and articles can mandate every punctuation mark. But when our citations have many complex parts, we do need the ability to adapt to assure clarity.

Comment 7.

Your use of square editorial brackets is appropriate here.

Comment 8.

As for choosing the layer for your folio/page reference, see EE4: 3.1, “The Rule That Has No Exception” for constructing layered citations: Details that describe one entity must not be attributed to a different entity.  Folio and page numbers appear on the original documents. Image numbers and frame numbers are a creation of the website that delivers the images. We don't assign the image/frame numbers to the layer in which we describe the original document; and we don't assign the original page numbers to the layer in which we identify the website.

You also point out that, on these censuses, the final pagination “would not have been added until the records reached the authority that took custody of them.” The same situation exists with the U.S.’s historical censuses: while the individual enumerators may have numbered their sheets of paper (either as folios or pages) before sending then to the Bureau of the Census in Washington, the Bureau then assembled multiple sets of returns into volumes and added their stamped numbers for the whole volume (which are also sometimes folio numbers and sometime page numbers).

Comment 9.

Thanks for commenting on 7.48. I do not have a current subscription to FindMyPast, so I cannot check the point. However, two factcheckers/proofers came in behind me before EE4 went to press, in addition to my own several re-checks before submitting the manuscript to others to catch remaining gremlins. Either situation is possible. Before a new edition goes to press, all issues will be checked again (and, judging by past experience) many of EE's current examples will be replaced or redone due to website changes. 

Comment 10.

Yes, sometimes the RG and piece numbers do appear on the images themselves, as opposed to the provider’s frame for the image or the provider’s sidebar.  If so, we can include those numbers in the layer in which we cite the original if we wish.

The issue then, in most cases, becomes one of juggling punctuation to achieve clarity. As you know, traditionally, in citing manuscripts, the semi-colon has been used to separate the organizational layers that have internal commas—document details; file details; collection details; record group details; archive details. (Particularly so in the U.S., whose formal archives have not traditionally assigned document numbers and piece number citations to facilitate the more cryptic citations possible in the UK, Canada, etc.) Thus, in countries that follow the U.S. approach, each set of details in our citation to a document is essentially a layer created by the organizational scheme assigned to that document.

Evidence Style citations follows the traditional format for documents in formal archives. It then extends that layered concept to records imaged online—i.e., those semicolons separate the layers between our details for the original document, the online provider, and (if possible) identification of the original archive. 

All things considered, using a Layer 3 as a "location layer" for the archived original that we have not consulted —when at least partial details are provided by the image or the website—is the clearest way to handle those iffy details.