The Appendices of The Lord of the Rings

This feels like a problem best addressed using computers. Not sure how well OCR would handle some of the accented characters, but it would allow for matching of large swathes of the text to find differences.

Hi Stu and Trotter,

With regard to finding errors in the Appendices, I do not feel that using computers will help with this other than to log the errors. If we had exact reproductions for each page it would cut down reading time but any software that finds errors and differences in text is filled with pitfalls in its self. After many years of building digital archives I have found that the only true way to compare and correct differences in text is by eye. Reading each passage word for word is the only sure fire way to minimize error and even then it is open to so many variables.

Also there is the question of which edition to be used as a true version and which are described as with error. The latest corrections that Wayne Hammond has found are from working with the manuscripts which I am pretty certain most of us do not have full access to if any access at all. Plus even the most recent printings are now textually incorrect to the First edition.

I have been updating my eBooks for many years now as errors and revisions come to light and also once I have an early edition of Rings I will adapt my digital version to show the changes over time but all one can do with regard to changes in text is to note them and keep both until an official version is released with those changes included.

I may be thinking too deeply about the Appendices but would we by finding errors across many different editions add to the complication as there is no true correct version in a retail state that we can describe as error free.

It does sound like a fun exercise to take part in if I can help with digitizing any of this.

I wish all of you all the very best for 2016.

Phil

onthetrail wrote:
Hi Stu and Trotter,

With regard to finding errors in the Appendices, I do not feel that using computers will help with this other than to log the errors. If we had exact reproductions for each page it would cut down reading time but any software that finds errors and differences in text is filled with pitfalls in its self.

I'd still think it is useful as a base-point for identifying differences that can then be explored further. In Software Development we regularly use these kinds of tools understand what has changed between source-code versions, often across multiple files and documents. They aren't a substitute for the human eye/brain, and an understanding of the content/context, but they definitely do help identify differences that the brain will simply skip over otherwise (single character changes are really hard to identify). Humans are notoriously bad at making these kinds of comparisons.

Sorry but I don't see how any software can help unless a digital reproduction is used of all the versions being cross referenced.

Software will not find errors from a paper document that has been scanned unless it is OCR'd, at which point the resulting document would then need to be referenced against a digital version and the errors that come from OCR corrected or the scanned and OCR'd document corrected by eye. I do not own one digital copy that has no errors and some are quite major so this would not represent a sound base copy with which to correct the OCR'd document. My own ongoing digitals are still full of errors and I have been working on them as long as they have been available.

I do see your point and you are quite correct about software finding differences in source code but I feel this is quite different when dealing with paper books. Humans are notorious for missing finer detail but some are well practiced in seeing errors on a page and can quickly identify changes in text that should ultimately be the same.

onthetrail wrote:
Sorry but I don't see how any software can help unless a digital reproduction is used of all the versions being cross referenced.

Software will not find errors from a paper document that has been scanned unless it is OCR'd, at which point the resulting document would then need to be referenced against a digital version and the errors that come from OCR corrected or the scanned and OCR'd document corrected by eye. I do not own one digital copy that has no errors and some are quite major so this would not represent a sound base copy with which to correct the OCR'd document. My own ongoing digitals are still full of errors and I have been working on them as long as they have been available.

I do see your point and you are quite correct about software finding differences in source code but I feel this is quite different when dealing with paper books. Humans are notorious for missing finer detail but some are well practiced in seeing errors on a page and can quickly identify changes in text that should ultimately be the same.

Yes, I was suggesting OCR. OCR can be very accurate, depending on the quality of the original (fonts, quality of printing, etc). It can also be very inaccurate, as you say, especially if the quality of the original type is poor. However, comparisons of OCRd copies are going to give you false positives, not false negatives, so if the OCR process itself is not arduous, there is essentially zero down side (unless the number of OCR errors is so high as for it to be useless -- which seems unlikely at this point in time). Plus, presumably current versions of the Appendices are already digital, so if the goal is to work backwards from what is there now and understand what has changed to get there, digital is the only sane choice (IMHO).

That said, I won't be doing any of the work (digital or otherwise), so my opinion is somewhat unimportant compared to those who will actually get involved.

Hello... I'd like to offer my help if I may. Having never done something like this before I'm not really familiar with the best way of going about it. Especially since this sort of thing, I'm guessing, has been done before, at least to some degree (Hammond and Scull?). I'd be surprised if I could find anything that hasn't already been noted. But still, I'd like to try! Maybe I can start out with an example and see if I'm headed in the right direction:

I noticed that in the 1st, 2nd and continuing editions/printings (at least until the 80s, 90s?), on page 319 (1st edition), Appendix A.ii.Footnote no. 1 the note begins: "The wild white kine..." but at some point it was changed to "The wild kine..." omitting the word "white" as seen in the Hammond 50th Anniversary edition (2004, 1st printing) p. 1039.

I didn't see anything in the Reader's Companion or on Hammond/Scull's site concerning this... but I'm sure it's most likely that I'm missing some information that's already out there. Anyway, let me know if this is useful and if so (or if not) how I can help further.

Thanks!

Yes, this is exactly what I was looking for, I will try and put up a spreadsheet that everyone can access and update over the weekend, so that interested parties have a place to update with items that they have found.

It does not matter what editions you are comparing, if you find a change then please record it, as it can be checked in editions that members have, and you may not have.

Having checked on this, the latest Return of the King that I have does state "The wild white kine", so it looks like this has already been looked at.

Re: OCR, one of the issues we describe in our blog post, as mentioned by Trotter at the start of this thread, was an aspect of layout rather than of text: an indent which seemed out of place, indeed is unique in the Tale of Years. It's not necessarily wrong as a matter of style, but because it was unusual it was questioned, and the questioning revealed two errors previously undetected. The second issue we dealt with concerns a date in the Bolger family tree (first published in 2004) which a reader noticed was different from that given in Peoples of Middle-earth, so was not a point of textual change over the course of editions and printings of LR.

Christina noticed the "wild kine" / "wild white kine" difference in Appendix A after the 2004 printings of LR. We sent in a correction for the 2005 printing, and noted it in The Lord of the Rings: A Reader's Companion, p. 809. Like so many errors, it had crept in with resetting for the 1994 HarperCollins edition.

Wayne

I have just a look at Appendix F in the 1955 ROTK and the latest single volume ROTK that I have.

It would not be possible to use OCR to do a comparison like this as the page formatting is to dissimilar between the editions.

I noticed a couple of very minor issues, which look to be to a matter of editorial changes, rather than actual incorrect text in the latest edition.

As an example on page 1129 (line 17), the first edition has the word 'dealings' and the latest version says 'dealing'. I prefer 'dealings' as in the first edition, but both are correct, and it would not seem appropriate for any changes to be made in this case.

It is nice to see that it is only very minor areas like this that I have found and I have to say thanks to the efforts of Christopher Tolkien and the Estate, HarperCollins and Wayne and Christina for this.

I also think that the latest version is a much better reading experience than the first impression :)