Last time I blogged about the <choice> Tag Project, I established a semi-comprehensive list of situations that required <choice> tags, and ended with a series of unsettled questions. Since then I have investigated the issue more thoroughly and now, having answered those questions, I can put together a more definitive guide.
Aside from Blake’s letters, there are four currently published works in the Blake Archive that contain <choice> tags: An Island in the Moon, The Pickering Manuscript, Tiriel, and Genesis. These four works contain a total of 494 <choice> tags. So between those works and what I could find in the currently published letters, I had a good sample size to work with. It was large enough to contain a great variety of <choice> tag situations, but small enough that I could perform a survey of all of them.
Hence from these works I was able to distill what seem to be the basic principles involved in using <choice> tags. From these principles it is now possible to outline prescriptively how Blake Archive transcribers should use them going forward.
The question with which I ended last time was how to structure <choice> tags when a word begins on one line and ends on the next; do the <choice> tags go on both lines or only the first one? Genesis and Tiriel both lack such situations, which is unfortunate because both are very recently published works and would hence make for a good precedent. Fortunately, An Island in the Moon contains plenty of these situations. The first such situation only has <orig></orig> on the second line, but each one after that has the whole thing (aka <choice><orig></orig><reg></reg></choice>) on both lines. This anomaly provided a good opportunity to take a look at what difference it makes when actually searching for the word.
Part of what the <choice> tags actually do is highlight the word inside of <orig></orig> when it comes up in a search. In the above situation the part of the word that lies on the second line “ation” has <orig></orig> around it, but no <choice> tags. As you can see, <orig></orig> doesn’t actually do anything without <choice> tags.
The word “Osteology” on the other hand, has <choice> tags around both parts. Hence in a search for “osteology” both parts of the word come up highlighted (capitalization doesn’t make a difference with the search function).
Could we do nothing with the second line? Technically we could and the search would still work, but it is both aesthetically more pleasant and technically more accurate to have the whole word highlighted when someone searches for it, which requires all three tags on both lines.
It can be a little more complicated when words are split up for other reasons.
The biggest lesson from the Genesis manuscript is that it isn’t just unconventional spellings and words split by line-breaks that require <choice> tags, but any situation where the word is split up into multiple parts. To put it more concretely, if part of the word is underlined, italicized or superscript, we need a <choice> tag so the search function will recognize the word in spite of the gap between letters.
Whether or not the entire word should be inside the <choice> tags, or only part of it as in the image above, is something on which there does not seem to be perfect consistency, but the best way to make the distinction seems to be whether or not it would cause errors with the XML to put the whole word inside. The testing app is a good way to check. While it is desirable for the whole word to be highlighted when someone searches, it still works (i.e. it still takes the user to that word) as long as the regularized spelling is hooked to part of the word, so we have to stay in the realm of what is possible. Genesis is rife with situations where an underline ends partway through a word, and these situations are handled by <choice> tagging the latter (non-underlined) part.
For famous people, it’s a good idea to add the first name, not just the corrected version of the last name. Napoleon is an example, as we see in the case of Blake’s May 28th 1804 letter to William Haley:
It makes a lot of sense because someone searching for mentions of Napoleon in the Blake canon would likely think to search for Napoleon before they would think to search for Bonaparte. Either way, we’ve got them covered!
Likewise with names that are less famous.
I’ve found that we typically use line notes to cite our source when we learn who one of these people is from Erdman or Bentley.
The rule for abbreviations also applies to abbreviated forms of titles and name suffixes, including those containing superscript letters.
As to the question of whether or not we provide the American version of Anglicized spellings, there is ample evidence in published works that we do.
One other thing I was still wondering about last time I looked into <choice> tags was the distinction between <reg></reg> and <corr></corr> for the correct or standard version of the unconventionally spelled word. An Island in the Moon for example contains a dozen instances of <choice> tags that contain <corr></corr> instead of <reg></reg>.
However all 120 <choice> tag situations in Tiriel and all 52 in Genesis use <orig></orig> and <reg></reg>. Since these were published more recently than An Island in the Moon, the absence of alternative versions suggests we’ve gotten away from using them, which seems reasonable. For my part, I can’t figure out what the difference would be between “Chymistry” and “askd” in the examples above that would necessitate using a different tag for their regularized spellings. Does <corr></corr> work?
Yes it does. But there’s no real point in having two tags that do the exact same thing. Maybe at one point the idea was to make some kind of distinction between a misspelling that resulted from hasty writing and a now unconventional spelling that would have been normal for Blake, or something to that effect. This kind of distinction might be somewhat more interpretive in nature than would be consistent with the Blake Archive’s editorial principles.
This guide will not make every decision of when and how to use <choice> tags easy, but it should prove a good toolbox for dealing with such decisions!