Bill Mullins wrote:And if you look at the actual report, he excluded Hilliar and Sanders for reasons other than dissimilarity of language.
That is completely false. He only excluded them for the comparison of the Charlier Pass. . . . All other tests and comparisons included Sanders and included Hilliar, and thus his overall assessment does include both of them.
Chris, Olsson's own words contradict what you just said.
With respect to Sanders, he said: "It is not possible to make a direct comparison between Expert and Sanders." And in Table 21, with respect to Sanders' diary he says: "Not a viable comparison because diary written during adolescence". He could not and did not do a comparison with Sanders, and thus his "overall assessment" cannot include him. His conclusion says "I would specifically rule out Hilliar, Roterberg and Wilson." -- Note the omission of Sanders. The only mention of Sanders in the conclusion is with respect to some hand-written letters, about which there is no discussion or analysis given other than a statement that the language is different.
And about Hilliar: "any form of plagiarism is of itself despicable, and this discredits Hilliar completely
in my view." [bold emphasis mine]. Hilliar is discredited completely, not just with respect to the Charlier Pass.
Bill Mullins wrote:He did not critique that only 5 candidates were studied -- he critiqued (like I have) your misrepresentation of the results of the study.
The study included all major candidates who could be included in a linguistic study. Exactly what is your problem with my characterization?
That when you describe the results you significantly exaggerate how good a match Olsson found between Gallaway and Erdnase.
Let's suppose that "match-ness" can be described on a scale of 0-100, and that we normalize it so two samples known to written by the same author on the same subject/genre around the same time under the same circumstances would get a score near 100. For example, two different sections of the same book written by a single author ought to score high, if there even is such a thing as an author's style that stays consistent and can be measured. Richard Kaufman's "Genii Speaks" for Jun 2017 ought to match his "Genii Speaks" for April 2017 pretty closely.
The first parts of Gallaway's book ought to match later parts to a high level. The first parts of Expert should match the latter parts to a high level. If they don't, that says that the whole concept of studying the works for similarities of style to show common authorship is bogus.
Olsson did not say that the score comparing Gallaway and Erdnase was in the high 90s. He just said that the score for G-E was higher than the score for Wilson-E, or Hilliar-E, or Roterberg-E, or Sanders-E. It could be that G-E was 37, and that W-E, H-E, R-E and S-E were all in the 20s, and that you can't count any of these as a "match" unless the score is greater than 75.
Ever since you announced on the Forum that you'd hired an analyst and that he found similarities, people have been asking for details -- how good is the similarity score, and how high should it be to say that the two works come from a common author? You never answered that, and when you released Olsson's report, it didn't answer it as well.
Olsson says that X > Y1, Y2, Y3, Y4. He never says what M is, that X should be greater than to be a match, and he never says that X > M. You, on the other hand, say that X is very high, near 100. Olsson's report does not support that conclusion.
And Clay brought up this same issue on 11/29, and you have not addressed it.
Who else should have been included to make the study that much better?
1. He should have used samples of the same size. Look at the note for Table 3: "Values not adjusted for length." Why not? "Gallaway . . . has more very long words in common with Expert than any of the other candidates. However, it should be noted that Gallaway’s book is longer than some of the other books and so, proportionately, we should expect this result to at least some extent." So Olsson admits that length skews results. Why did he not correct for this?
2. In the cases where an author had more than one work available (Roterberg, Sanders, Hilliar, Gallaway) he should have compared them against each other to get a sense of how consistent in style each author is. For cases where only one work was available (Wilson, Erdnase) he should have broken the work into parts, and compared part vs part, for the same reasons.
3. He used a number of different tests without giving us any validation of them. Why did he use the particular tests? Have they been show to be more effective than others? Have they been shown to be effective at all?
4. He should have used more and better samples of writing from Sanders. Olsson analyzed two Mining Articles, and a few years of adolescent diaries -- both of which he admits so different in genre to be problematic to do a good analysis on. He says he had late access to some handwritten letters, but did no substantial analysis of them, other than reading them. As Leonard and I
have pointed out, Sanders had the ability to write in a substantially different register than he used in his technical mining work. When you were putting samples together for Olsson to analyze, you should have asked me. I've got tons of stuff on him.
Gallaway had magic and gambling books in his library.
Again with the exaggerations. You continue to assert without any evidence that Gallaway had magic "books". He had a copy of Expert, which is a book with some magic content. If you know that he had any other magic books, you should say why or back off this claim. When you continue to repeat something so demonstrably, obviously wrong, it really dilutes the rest of your arguments.
That means the fact that we have no knowledge about E.S. Andrews having written anything of substance is a disqualifying deficiency
Implicit in this statement is that since we don't know of other writings by Andrews, they must not have ever existed. Which is not a valid argument. Most things written 115 years ago are lost. If you look at the of the copyright registry issued by the Library of Congress in 1902 that includes Erdnase, some of those books are currently findable online or in worldcat. But many of them are not available for review anywhere as far as we know. Andrews may have written letters or other private documents that allowed him to build his skill. He may have written regular reports to the railroad while he was on the road. Your own research into Gallaway should show you that works that you cannot find still exist -- where are Gallaway's many newspaper articles? Where is his Ft Payne newspaper? Most 19th century newspapers might as well not have ever existed, for all that we can research them now. I myself have published many tens of thousands of words that neither you nor the general public will ever see - they are classified technical reports that are not for release.
Also implicit is the idea that a writer can't have only one publication, which is wrong. Anna Sewell wrote only one thing -- "Black Beauty." Closer to home, Gerritt M Evans wrote only "How Gamblers Win". Koschitz wrote only his "Manual of Useful Information." By your logic, each of these must have actually been written by someone else, who has other publications.