The Source of the Source, of Course, of Course



6 March 2015

Citing ... What?

The source of the source, of course, of course.*

EE user Simon (aka “Nomis”) poses a question that puzzles many researchers. When we cite a source that cites its own source, we always cite first exactly what we’ve used. We don’t “borrow” our source’s citation and leave the impression that we consulted the original. But, of course, we do want to record whatever our source gives as its own source.

Typically, we introduce that second layer of our citation with the word citing .... But, Simon asks, what exactly should come after that? What details need to be included?

Here’s the answer you don’t want: It depends. Let’s start with basics and then complicate matters a bit.

Ideally, we provide a full citation that covers all the essential elements for that type of source. That ideal may or may not be achievable.

Some sources cite their own sources very well. We might then copy exactly what they cite—in which case it's wise to put the quoted words in quotation marks. (Some editors to which we submit our work may choose to drop those quote marks; but in our own data collection stage, we want to be precise enough that, in our future analyses, we will know exactly what words we've copied from elsewhere.)

Some sources poorly cite their own sources. In those cases, we might also decide that the best course to follow is to quote exactly what our source gives us. Again, we put those copied words in quotation marks. (In these cases, editors will likely do the same, or they’ll ask us to go find that source-of-our-source.)

We could, of course, track down our source’s source and use it without waiting for an editor to ask us. That’s the best of ideas. In that case, eventually, we’d be citing the original. We would not need to cite the derivative unless, say, the derivative misrepresented what the original said and we want to correct that. But until we get to the point of using and citing the original, we need a workable "source-of-the-source" citation for what we’re using right here and now.

If the source of our source is a book, incompletely cited, our first thought might be to check a bibliographic catalog like Library of Congress or WorldCat, either of which is likely to offer a full citation. But that’s risky. Many books have similar titles. Many authors write books with similar titles. Many books have gone through multiple editions in which both content and pagination can differ. Creating a citation for a source we haven’t actually used is never a safe idea.

Sometimes, our source’s “original” source data is so convoluted that copying it exactly would be almost pointless. Take, for example, the “California Passenger and Crew Lists, 1882–1959” collection at that wonderful fount of imaged original documents,  The provider’s source-of-the-source data cites 22 different microfilmed collections, for a total of 848 words. No sane person is going to create a reference note that quotes all that source-of-the-source data.

When we used this example for our QuickSheet: Citing Databases & Images in 2010, the collection covered “only” 10 microfilm collections, not 22. (It also used slightly different dates in its title.) However, its database entries did not identify which of the film publications the data was extracted from—much less the roll number and frame number.

The only logical choice in a situation of this type, whether we are dealing with 10 different fiilm publications or 22, is a generic citation with a caveat. This is what we decided upon in 2010.

“California Passenger and Crew Lists, 1893–1957,” database entry, ( : accessed 6 March 2015), manifest R.M.S. Makura, Sydney, Australia, to San Francisco, arriving 16 January 1931, p. 1, Albert Dion Barr Smith; citing National Archives microfilm publications A3408, A3419, A3361, M1388, M1410, M1411, M1412, M1416, M1438, M1439, M1494, and M1852; exact publication and roll not cited for individual images.

If we are citing the actual image rather than the database entry—as we should, of course—then the image itself or the provider’s digital bread crumb might identify the exact film publication we are dealing with. In this case, it doesn’t. 

With some digitized microfilm, we can simply page back to the first image or two and find there the “title page” to the microfilm publication, from which we can create a streamlined citation to the exact film. In this case, we can’t. Neither Ancestry nor the microfilm itself provides the exact title and publication number. The best we can do, then, is the same thing we did for the database—substituting the words “digital image” for “database entry,” of course.

But what would we do today, considering that the collection has grown from 10 film publications to 22? No, we would not likely cite all 22 of them. The only practical option when we have only generic data for a collection this size would be an explanation, rather than a citation:

“California Passenger and Crew Lists, 188s–1959,” digital images, ( : accessed 2 June 2010), manifest R.M.S. Makura, Sydney, Australia, to San Francisco, arriving 16 January 1931, p. 1, Albert Dion Barr Smith; citing 22 National Archives microfilm publications, covering 700+ rolls of film; exact publication number, roll number, and frame number not provided for individual images.

We’ve gone a bit deep into the weeds here, for those who love the weeds. For more practical examples, you might try EE itself, pages 27, 52, 180, 301, 344, 348, 404, 446, 462, and 577.


*With apologies to “Mr. Ed” for adapting his signature song.