With the general question of if an archive must explain themselves and the choices they make, the largest objection I could think of would be time. Digitizing and/or cataloging items already takes quite a bit of time. Add in having to explain why you made the choice you did, and you're now spending even more time. And what explanation is long enough? Or too short?
I think a potential solution to this problem is also a potential side step. By linking together texts/objects with metadata tags, you are showing a general theme through specific items that can span the entire archive (as well as helping with search and user access). Perhaps you do not have to excuse your choice as much as you have to prove that said choice has a place in the archive. Maybe that's enough.
That alone is a big ask, as it can be rather subjective as to what terms to use as a metadata tag for which journal. I have decided to explore using the text itself, and using major terms found in the text itself as potential metadata tags. That can take a bit of time, so you to automate it.
To illustrate this, I have used Voyant Tools to automatically read the PDF copies of Blast issue 1, Camera Work number 5, and The Dome, vol. 1 no. 5. By doing so I was able to gather a word map which I have provided below:
The "TermsBerry" that Voyant provides is also useful in this regard, but Voyant does not like making an image of it that I can link here, apparently.
As you can see, there are plenty of words that we can look at to see about using as metadata tags. However, there are also spelling errors, and some more useless words like "good" or "new." We discussed methods in class that we can use to clean up XML files, and I feel as if that is a good way to avoid the spelling errors. This may work rather well if I were to expand the scope even further, and include say all of the MJP's offerings. Then we can get search tags that appear in (at least) a majority of the magazines, and show a clear link between the offerings.