Statistical Methodology – understanding the “replica comparisions”

Statistical Methodology – understanding the “replica comparisions”

What do the numbers show? The numbers count differences between the replica and the edition being studied. Making no distinction between correcting spelling mistakes, replacing inadvertenly omitted text, or introducing inadvertent errors, those numbers in no way represent “scores” or “relative quality.” They are qualitative, not quantitative measures. They tell us there were changes, they do not tell us the nature of the changes. And the nature of the changes varies widely in these editions.

The comparative data, if properly understood and analyzed, is quite good and quite useful however, and so is being included here for reference purposes. Carefully studied, it can show the student a great deal about the relative quality of the editions. This “qualitative” information which can be extracted does not in any way resemble the “raw scores” which simply count, but do not evaluate, changes

Is a high score good or bad?

There has been an assumption on the part of some that “more changes” is less desirable than “fewer changes” which would only be true if ALL changes were undesirable. The only way that would be true is if you view fixing spelling errors as ‘undesirable’ since all editions “change” (correct) some of Schucman’s original spelling and grammar mistakes. Some editions insert material from both earlier and later versions where there is evidence that it is closer to the original dictation, and one edition, the CORRECTED HLC footnotes each such modification.

Since the first oral dictation of the Course, a great many changes were introduced, many are just plain typing mistakes and certainly need to be fixed. Others are corrections of earlier scribal errors, and others are the addition of modifications which have no apparent reason. This is true of our current editions also, there are changes which are clearly repairs to earlier mistakes and we see some changes which have no explanation. Not all changes are created equal!

The Course originated in verbal form, and Schucman took notes. In these notes there were some errors. Some corrections were dictated and are recorded. The notes were edited and recopied by hand at least six times, and with some sections it is certain there were many more than six iterations. Several of these “versions” are available to us, and in each we can see that as Schucman and Thetford, and later Schucman and Wapnick, edited the material some previous mistakes were corrected and some new mistakes were introduced, and there were some changes which aren’t obviously one or the other. So, when an apparent ‘error’, even a simple spelling mistake, shows up in the HLC, the CORRECTED HLC methodology was to check both earlier and later material to see if Schucman fixed this herself later, or to see if the error stemmed from a typing mistake between earlier material and the HLC. If the material appeared “correct” in previous versions, it is generally reasonable to suppose the error in the HLC was a copying mistake. If corrected in later material, that was evidence that Schucman had noted the problem and had introduced a repair.

In any of these cases, where there is evidence of a mistake in the HLC, and a correct form is available in earlier or later material, there is both evidence of a mistake and evidence of what Schucman intended to put there! In the CORRECTED HLC, with that sort of evidence, a change would be made and footnoted. The later CIMS(2) material adopts some of the CORRECTED HLC modifications of this sort, but does not footnote or attribute them.

In dealing with illegible passages in the HLC manuscript itself, to generate the replica, a similar methodology was adopted. Both earlier and later material was consulted to see what was there before and after this version. That was our only evidence for what was originally on that page.

One may view such changes as desirable or undesirable, and one may view documenting them as desirable or undesirable, but neither sort of change is of the same quality as the apparently inadvertent and unintentional mistakes which are surprisingly numerous in all but one of the editions. It is agonizingly apparent that little proofreading to catch errors was done on any of them except the CORRECTED HLC as there are a great many changes which are almost certainly not intended by the respective editors. Indeed, one might question whether any of these editions after CIMS(1) did any systematic proofreading at all!

One edition, CIMS(2) “OE” sets out to modify punctuation and original emphasis on a massive scale without any apparent evidence that there was an original “error.” This is a completely different standard of editing than that generally employed by the others. Rather than replicating the HLC typescript exactly except where there is evidence of error, and there is a lot of that, this editing methodology re-interprets the typescript, replacing the Scribes’ emphasis and punctuation patterns generally with its own. Rather than “proofing” or “copy-editing” for “error” this becomes a kind of ‘re-scribing” or re-writing of part of the material for reasons other than “apparent error.” I suspect the number of readers who would knowingly choose a ‘re-scribing’ over a reasonably accurate ‘replication with corrections’ might be few.

One edition, ACIM 1972 sets out not to fix, but to alter Schucman’s capitalization conventions. Helen capitalized pronouns relating to a person of the Trinity or a Name of God. Most of the time. She’ll sometimes capitalize one “He” in a sentence that refers to God and not another. That would be an “error”, most editions “fix” it and that would also register as a “change.” Such capitalization is actually exceedingly helpful where ‘he’ is being used to refer both to God and man, as is often the case. It makes it much easier to parse sometimes complicated sentences. In ACIM 1972 all the pronoun capitalization for divinity is removed. This very much simplifies things but removes a little bit of useful, though certainly not crucial information. I don’t know of any sentence where the context makes it impossible to tell, if you study it carefully, which pronoun belongs to whom. The capitalization just makes it easier, and most newcomers to this work need all the help they can get!

Different editions, different objectives, different methodologies

Because the Corrected HLC is a very different kind of document, its comparison is handled very differently. While you can see each change just as with the others, you can also see the editors’ explanation for each change in the accompanying comparison file against the replica. The Corrected HLC comparison was not made by Raphael.

It must be borne in mind that there are two very different kinds of editions, the undocumented which has no footnotes and the documented, which footnotes many things. Not all footnotes represent a change in the text, some simply offer information. Yet they all get counted in the “score” as do annotation numbers for paragraphs, sections, and chapters, save in those analyses in which that information was stripped, for some reason. The reason was probably to try to provide numerical results which could be compared meaningfully. This is a bit like comparing apples and oranges and evaluating only the number of seeds each has. Unless the only important difference to you is the number of seeds, such a measure, although it may be accurate, has no meaning. We don’t, by the same to Wapnick, evaluate students in schools on the basis of their shoe size, even though it would be exceptionally easy to do so!

In the case of CIMS(2) and ACIM 1972, while lip-service might be paid to “scholarship” there is clearly no trace of “scholarly” methodology being applied in any systematic form. The objective was not a high degree of accuracy, the objective was not to change only when there was solid evidence of a mistake, the objective was to get something “quick and dirty” out the door rapidly. With the pre-2006 material, when it was very hard to get any copy of the HLC, that is completely understandable. With the CIMS(2) material, which came out four months after the Corrected HLC, and actually copied a few of its corrections, there appears no reason to do a very low quality edition when one of considerably higher quality is already available. It’s one of the many mysteries of the Course history! Why, when a good edition is available, would anyone go to the trouble of publishing a poor quality one?

In the case of the Corrected HLC, we have a thoroughly scholarly approach by serious ACIM scholars, one of whom has an academic background in Biblical Scholarship and brings those tools to the task. The methodology was clear and up-front and thoroughly conventional in the field of textual scholarship. Each change was documented and explained. The reason for that is the basic scholarly one, it is an invitation to others to review each decision, research it more, and if evidence of error is found, offer up corrections.

Some people are certain to disagree with some decisions about what to change, but at least in this edition you can see what was changed, and what was originally there, and make up your mind which reading you prefer. No other edition gives you that option. None reveals at all what they changed. You’ve got the editors’ choice, and often you’ve got an inadvertent typo. If you think there’s a mistake, we very much want to hear from you. We know there’s room for improvement and improvement is the direction we plan on taking!

For the most part the choices of what to change and what not to change were not based on opinion or preference or guidance or inspiration, though some of the footnotes which don’t point to changes are of that sort. Nothing was changed without evidence that there was a mistake, and the change made was the best which the then-available evidence allowed, and that did not include the original notes which were withheld from the Corrected HLC editors by those who possessed them.

How many mistakes are there which should be corrected?

By our estimate there are at least a thousand instances of inadvertent spelling, grammar, punctuation, capitalization, or copying errors in the HLC typescript. It is certain that more will be found when a systematic comparison of the HLC to the earlier Sub-Urtext is undertaken. Many of them involve more than one word, so could run up some pretty high scores. Just changing “no-one” to “no one” as copy-editors are wont to do runs up 219 points. We’re almost a quarter of the way to a thousand with just one single change! While preferences for spelling variants will differ, not many people are going to consider choosing one spelling variant over another a very important matter! Schucman herself switched back and forth between the two, finally settling in her last editing work on the more popular “no one.”

If the quantity of changes is understood to be a qualitative assessment, the reader will be very badly misinformed by these numbers. The numbers, by themselves, tell us nothing meaningful about these various editions and should probably not be included because they seem to those who do not understand statistics, to offer a relative evaluative “grade”. In fact, they have no “qualitative” value whatsoever. What is meaningful is looking at what was changed and why … and in all but one edition no explanation is offered for any change and a great many are rather obviously inadvertent errors.

Of the contemporary HLC editions, only one sets out to document why its editorial board altered the material in the way it did and only one appears to be largely free of inadvertent typos or seriously inadequate proofreading. That is the Corrected HLC. All editions are mostly identical to the “replica” and all deviate from the replica hundreds or thousands of times. Only in the Corrected HLC are the deviations noted with the original reading reproduced so the reader can see what was changed, read the explanation, and make up his or her own mind as to what is likely correct.

The others all make changes, and except for CIMS(1), make more changes than the Corrected HLC according to Raphael’s score. The reason CIMS(1) has so few changes is that its editors only attempted to correct the most obvious of typos and left all the more subtle ones for the future to unscramble. None footnote, document or otherwise explain why their changes are made, however, except for the Corrected HLC. Indeed, unless you essentially “proofread” those other editions yourself, line by line, against the original typescript or replica, you’d never know there are changes. Raphael’s comparisons are extremely useful then, in that they show us what was changed, but in no case do they show us why it was changed.

Thanks Raphael!