Thread 5--Epistemology, weak evidence, planted fragments, and NAA

      Thread 5 was the longest. It contained 28 messages and dealt with the four separate ideas noted above. Although its topics were solid, its dialogue was best described as unfulfilled.
      It began with F warning JM about the dangers of epistemology:

Also, beware epistemology. It has a fatal flaw.

      The second message was a strange one from D, his last communication here. D challenged F, the originator of the thread, to show how the magic bullet could really be magical:

Produce you proof that a material object such as a bullet can have the quality of magic. I dare you to post your proof on alt.sci.physics.

The only problem was that F didn’t say anything about a magic bullet.
      This caused G, in the third message, to challenge D back:

What magic? No one has proposed that that “quality” is part of the explanation of what happened. Critics have used the term, but that is another matter.

That was apparently enough for D, who then backed out of the discussion for good.
      F used the fourth message to gently query D, but with no result:

I’m afraid I don’t understand what you are trying to say. I assume you disagree with what I wrote about the gravitational lens [not quoted above] or epistemology, but I can’t tell from your words. Re-write and re-send.

      Note how disjointed and useless the first four messages of this thread were.
      It got no better with the next few messages. JM then replied to F’s mention of epistemology in the original message, but failed to ask him what the “fatal flaw” was. After defining the term, she returned to the enduring strength of nonconspiracy and the declining probability that strong evidence for conspiracy will ever be found:

Hi F, Gee I had to look it up

[Greek episteme knowledge, from epistanai to understand, know, from epi- + histanai to cause to stand -- more at STAND]

First appeared circa 1856

: the study or a theory of the nature and grounds of knowledge esp. with reference to its limits and validity

But as it is not mine, I am not too worried : )~ (Sticking tongue out at you) Of course KR has gone into more details on his site about critical thinking and theories, even has an article by a certain F. However, even with flaws, the jest that the lone nutter theory is the only one to this date that has held up with supporting evidence. I also accept that with passing of time, especially after so many investigations the chance of finding anything to change the official conclusions is reducing. Even KR does not eliminate the possibility even if he does not think it will happen. Now, you know I am not as certain as KR or [name deleted] or [name deleted] that no such evidence will be found. But I know one person that had better come up with a better explanation then they just took the risk that not all bullets would be found in the body.

In the sixth message, F replied to JM’s last sentence, about planting bullets:

What else is there?? When you run across one person who can explain the inexplicably totally absent micro-distortion necessitated by the scraping of CE-399 along either of the two bones, you let me know.

JM had told F that he needed a genuine reason to propose that bullets had been planted outside the body. F responded that CE 399 had to be a plant because it lacked small distortions from hitting the two bones. F’s mistake, of course, is that it is not logically sound to argue that X could not have happened because Y was not found. This logic is erroneous because reasoning from the absence of something is very risky—you can do it only when you are sure that the event can happen in only one way, in this case that X must give Y. In simple situations this logic may be valid, such as for example, “Shoot a bullet through a person, and you will find an entrance wound and an exit wound.” But when things become more complicated, such as the question of whether a bullet that slides along two bones will pick up microdistortions, everything changes. F was essentially saying that there is no way that a bullet can scrape along two bones and emerge without detectable microdistortions. As if this weren’t risky enough, F posited this relation knowing that CE 399 had been deformed enough distort its base, squeeze lead out of it, and bend it laterally. Seen in this light, the absence of “microdistortions” seems inconsequential.
      JM then responded in the seventh message by proposing two scenarios under which the bullet would not necessarily have picked up markings on its side. Here are the key sentences for each:

First, is it really known if a bullet that goes thru muscle and skin only, slaps soft ribs, has to leave marking on the sides of it?
If this was a base direct hit to the wrist bone, why does there have to be scratches on the sides of the bullet?

The point is that one cannot be sure that CE 399 must have been damaged on its side until all such scenarios of JM’s have been investigated and ruled out. Because that is effectively impossible for this case, F’s reasoning was fallacious.
      That ended the “planting” phase of Thread 5. Note that none of the six replies to F’s original message on the perils of epistemology sought to challenge him on it or even to ask him what he meant by it. This was particularly unusual, given that the word “epistemology” did not appear in the essay on predictions—F had plunked it down from out of nowhere.
      I decided to step into the debate. I suspected that F was out of his league, and did not know the difference between the general sense of “epistemology,” which he was using, and the specific sense, as in “Eastern versus Western epistemology.” He seemed unaware that by stating that epistemology contained a fatal flaw, he was claiming that this great branch of philosophy was rotten to its core. He was thus putting himself above the great philosophers from Plato through Russell. So in the first message of a new subthread, I asked F what that flaw was. To help him understand his mistake and get to the right answer, I added a second question that referred to philosophers:

Care to share with us the fatal flaw in epistemology? What are all those philosophers going to think?

      Before F could answer, A jumped in with a long, rambling answer of sorts that initially touched on epistemology but then veered away to matters of weak evidence, timetables, and weak evidence again. One of the few correct items in this whole thing was its first sentence, on epistemology. It is important to discuss this entire answer because of all the errors in reasoning it contains. To keep matters simple, we take it one paragraph at a time.

I don't agree that epistemology has any particular fatal flaw. I do believe that there are many disagreements within the philosophical community as to how best to get at truth, and the limits to knowledge, etc. A good example in the Kennedy assassination are those who would propel Occam's Razor from a simple rule of thumb into a law of nature.

As noted above, the best sentence here may be, “I don’t agree that epistemology has any particular flaw.” At least A recognized the general nature of the word “epistemology.” His next sentence, about disagreements among philosophers concerning how best to find truth, the limits of knowledge, etc., was fine but seemed to be intended to prepare the way for undercutting my reliance on a simplified scientific method. Given that criminology is not pure philosophy, many of the uncertainties of philosophy will not carry over into criminology. So A was on semi-shaky ground here. But his last sentence, on the alleged informal status of Occam’s razor, shows that he was indeed out of his league. Occam’s Razor, of the Principle of Parsimony, is far more than a rule of thumb; it is a high principle of reasoning that must never be violated when seeking an explanation for a problem or event. It is also very relevant to the Kennedy assassination, for many if not all conspiracy theorists build into their theories components that have no specific supporting evidence. In so doing, they are violating the Principle of Parsimony.

As per my earlier post to you KR, I want to emphasize that it digressed from dealing with the prediction article to dealing with your entire site, a site I visited often, and have read almost all of the articles. For instance, I have read your article about what “a lot of weak evidence” adds up to, and believe the answer to that question is a whole lot more nuanced than you would provide. Indeed, I pointed out a discussion JM and I had relating to a jury that reached what almost any independent observer would say is the correct verdict in a case, but on limited evidence. They were basically permitted to see only the weak evidence in the case. Stronger evidence was withheld for technical reasons and not presented in the case until sentencing. But they pieced together a lot of weak pieces of evidence and gained an accurate assessment.

Here A confused belief with proof. He may have believed that the result of multiple weak evidence is more nuanced than I indicated, but he didn’t show anything evidence for it. Until he produces evidence to prove his contention, his opinion will carry no weight. The second part of the paragraph mischaracterized the evidence shown the jury. In a message below on the same murder case, JM showed that they got strong indirect evidence, not weak evidence. In fact, the weak evidence they got was actually the direct visual identification from the victim. Many people make the same error of thinking that indirect evidence is weak, when in fact it is usually strong. Indirect and weak are NOT the same thing! So this paragraph accomplished little and muddied important issues.

As per predictions, it ultimately comes down to an arbitrary time table, something you are far more explicit on in other articles. There are many predictions that haven’t been evaluated-- for instance, there are thousands of documents that haven’t been read and analyzed, making the claim “that the documentary record will suggest a conspiracy” hard to confirm or deny.

Here A returned to the nonexistent “arbitrary timetable.” It just isn’t so (See earlier comments.) The second sentence perpetuated A’s erroneous claim that physical evidence (documents, in this case) are predictions. The same sentence implies erroneously that meaningful conclusions can’t be drawn until all the newly released documents have been read and evaluated. But that’s not how science operates—working hypotheses are always based on the available evidence, no matter how much more remains to be investigated. To do otherwise would be to predict that the unevaluated evidence would change the current conclusions, which obviously cannot be known until the new evidence is evaluated. So at best, this message was off-topic, and at worst, it represented a serious misunderstanding of the thrust of the essay, which speaks of predictions only from explanations of the assassination.

I can cite quotes where you basically argue for some sort of time table as to when we should eliminate the CT theory (let's not even get into the incredibly nuanced and varied versions of CT that seem to be ignored by your site). I pointed out three examples where someone holding that viewpoint 20, 30 or even 200 years after a controversial event would have been undeniably WRONG in their position. And in those cases their isn’t the type of obvious CYA job that prevented full disclosure, including, and up until, today.

This paragraph contains five major errors. The first is factual—I did not advance any timetable for eliminating conspiracy theories. Anyone who believes that I did is fundamentally misreading this essay and others on my web site. I proposed only that the probability of finding any conspiracy is continuing to decline with time, not that conspiracy would never be found. The second is the abbreviation “CT theory” which cannot mean what A uses it to. CT can mean general conspiracy theory, as is used currently in the growing field of conspiracy theory, or it can mean a specific conspiracy theory, the sense used by JFK researchers. It we take CT in this latter sense, we get the nonsensical “conspiracy theory theory.” The third error is logical—even if we accept the wrong premise that my course and web site ignore the “incredibly nuanced and varied versions of conspiracy theories,” neither notion would be related to timetables. The fourth error is logical—reversing a longstanding conclusion in light of new knowledge is completely consistent with declining probabilities for that conclusion, as noted above. It is also consistent with the provisional nature of knowledge and with the use of working hypotheses. The fifth error is logical and procedural—appealing to evidence putatively hidden is no excuse for not being able to find the truth when the truth is understood to be provisional and progressive. (See discussions above.) One last point about the paragraph above is that the cases where longstanding public ideas have been reversed made news precisely because of the rarity of such events. That this happens, but only rarely, actually supports my point rather than refuting it.

I mentioned the strong evidence argument, because you have your class list the items of strong evidence and then a scenario that covers all or most of them. The scenario you choose is “Oswald acted alone.” My point was, and you seemed to pretty much agree with me, that the proper answer is that “Oswald acted.” That is much different in its implications for a conspiracy.

Many errors again. A fundamentally misunderstood our class exercise, whose point was to find the simplest scenario consistent with the strong evidence. That scenario must be the working hypothesis (á la Occam) until new evidence forces it to be revised. People who miss this point know little about the scientific method or the broader procedures of critical thinking that it represents. By choosing “Oswald acted” instead of “Oswald acted alone,” A revealed his misunderstanding of the scientific process. A was also mistaken if he thought that I would ever have agreed with him on such a fundamental error of reasoning; he was expressing what he wanted rather than what would really happen. Note that he offered no support for his claims that “Oswald acted” was the proper interpretation of the evidence or that I essentially agreed with him.

You then take the approach of offering the following methodology for evaluating controversial events:

“Given Scenario A, arrived by evaluating STRONG evidence X, Y and Z, you should ignore weak evidence A, B and C if it requires an addition to Scenario A.”

Again wrong in several places. A is first wrong at the very top, where he uses the phrase for evaluating controversial events. The methodology, which he misstates in the next paragraph, is used equally for any event. His misstatement about the methodology is claiming that the weak evidence should be ignored if it “requires an addition to Scenario A.” No. The weak evidence should always be ignored (with one minor exception). We use the simple, straightforward procedure of discarding all the weak evidence right at the beginning and formulating the basic explanation from the strong evidence alone (what many writers call the “physical evidence”). A has misrepresented the procedures we use in class.

So you argue that “Oswald acted alone” and all the anectdotal evidence (and I believe there is far more than anectdotal evidence to suggest a conspiracy, but we will concede that point for now) pointing to the existence of another shooter, particularly on the Grassy Knoll area, should be dismissed."

Still not a correct restatement of my procedure. We do not “argue that Oswald acted alone.” We simply note that this is the simplest scenario supported by the strong evidence. Maybe he acted alone, and maybe he didn’t. The available evidence allows us only to take as our working hypothesis that he did. The hard truth about weak evidence is that because it is unreliable, all conclusions drawn from it will be equally unreliable. Therefore, it must not be used when strong evidence is available. Who wants to poison his conclusions by mixing in unreliable evidence?

There is a competing methodology: 

“Given Scenario A, arrived at by considering STRONG items of evidence X, Y and Z; if WEAK items of evidence A, B, and C do not require that you contradict Scenario A, augment Scenario A to include an explanation for items A, B and C.”

Wrong again. This “methodology” obviously does not compete with the previous one because it deals with the fundamentally different circumstance of weak evidence’s being available in an area not covered by strong evidence. If a JFK researcher does not understand the difference between competing and noncompeting methodologies, how can he converse meaningfully? But again A misstated the obvious. I said that weak evidence could be used informally after the strong evidence, if desired, to offer limited and preliminary guidance in areas not covered by the strong evidence. This is a much weaker approach than A’s instruction to simply include the weak evidence. Whether A realized it or not, he was promoting weak evidence to a status that it cannot logically have and that I have not advocated. Whatever happened to the idea of reading the other guy’s stuff carefully?

Under that scenario, you acknowledge that the strong evidence suggests Oswald ACTING, but the abundance of weak evidence pointing to another shooter requires that you advocate “Oswald acted but with assistance.”

His competing methodology may require that, but that methodology is fallacious, for the reason stated in the previous paragraph. No circumstances require us to include weak evidence—that is A’s creation, and illogical to the extreme. It is also an interesting example of JFK researchers carelessly stating things as fact without checking them.

Just my thoughts.

Hardly “just his thoughts.” A had just finished stating several strong ideas, sometimes repeatedly, and contradicting what he was supposed to be quoting. His conclusions were presented strongly and directly, without qualifiers. And then he sought to soften the whole passage by saying they were “just thoughts”? He cannot have it both ways.
      So what did A’s long, rambling answer accomplish? Other than giving epistemology a quick vote of confidence at the very beginning, it confused every other issue it touched.
      Then F tried to answer my direct question about the fatal flaw in epistemology, in this third message of the subthread. The operative word here, however, is “try,” for F answered the wrong question, as I had suspected he might. He answered for my epistemology rather than for epistemology in general. But his full response was interesting in other ways, too. After devoting the first sentence of the first paragraph to a quick formal answer, he took several sentences to add an example that was totally flawed. He followed that with a short paragraph that was nonsensical, and finally added a P.S. that first jumped to Vincent Guinn’s NAA results and the probability that his two groups of fragments were distinguishable, and then ended with a question so ill-defined that it shows that he (F) understood little about statistics. Again we go paragraph by paragraph.

Eyewitness testimony, no matter how numerous and consistent the accounts, has no place at the table. Case in point; There is no way all the “hole in the back of the head” people are wrong. Period. It matters not to epistemology that the BOTH [“back of the head”—KAR] witnesses comprise over 90% the head wound witnesses. Those witnesses carry weight that is not accorded them under the epistemological method. Ditto the GK witnesses, eye, and ear nose alike.

Even granting that this first sentence was meant to apply to the epistemology of my JFK class, it still misrepresented my position on witness testimony, which is, to borrow a phrase from A, more “nuanced” than F recognized. (See above.) But the real errors here came in the remaining five sentences, which read like a passage out of Mark Lane. The thrust of these sentences is that because the great majority of witnesses to the head wound and the shots allegedly agreed on what they saw and heard, they must have been right. This argument is so flawed that one hardly knows where to begin a rebuttal. First, it assumes that the reports of the witnesses are being represented properly. Second, it assumes that they observed things that were so clear and obvious as to admit no other reasonable explanation. And third, it assumes that they really were the majority of the witnesses. Each of these assumptions is flawed in practice.
      My reference to Mark Lane was meant to recall a passage from Rush to Judgment where he discussed six witnesses from the grassy knoll who heard shots coming from the knoll. Lane concluded that when six unbiased observers agreed, the common denominator must have been the truth. So the shots truly came from the knoll. The obvious flaw here is of course that one could have easily found six witnesses who heard the shots coming from the Depository. By the same reasoning, their common denominator would also have been the truth. We would then have two mutually exclusive “truths,” that the shots came from the knoll and that they didn’t. So much for extrapolating from six witnesses to truth! One way out of this impasse would be to recognize that systematic errors could have affected either or both groups of witnesses. The most obvious such error, echoes, could have made the sounds come from a different direction from the actual shots. Another way out would be to recognize random errors among the witnesses, which would invalidate the small samples of witnesses such as Lane used.
      Now consider F’s unstated assumptions about the back-of-the-head witnesses. The first is that their reports were portrayed accurately. In at least some cases they were not, however. If you were asked to put your hand at the “back of your head,” chances are you would place it squarely in the back, or at the occiput. But if you compare that position with the photos of the witnesses in Robert Groden’s The Killing Of A President, where the back-of-the-head argument is advanced, you will see hands pointing to the dead center of the back, the right rear, the low back, the high back, the top, and the right side, all under the guise of the “back of the head.” Groden’s book and other similar presentations are an exercise in telling us what we should see in the pictures, regardless of what they actually contain. The sad thing is how many people have been taken in by this bogus approach. Now, if Mr. Groden were the only one doing this, I might not make so much of it, for he sees all sorts of things in all sorts of pictures. But our JFK class got a taste of the same thing from one of our guest speakers, who showed a series of similar pictures of witnesses also purportedly locating a wound in the back of the head. Their positions varied at least as much as Groden’s. When a speaker says one thing and shows something else, what’s a listener to do? The upshot of all this is that when people tell about how all the witnesses saw a wound in the back of the head, the speakers are being much more general than they admit. They are taking wounds from quite different places and calling them “back of the head.” They are thus creating a false consensus.
      But there is a deeper, more fundamental problem with all this emphasis on the location of the large head wound. It involves another unstated assumption, one by no means restricted to those who promote this argument, that the big wound was the exit wound and therefore could reveal the direction of the shot. This is false. The “gaping” wound in JFK’s skull was created when the head exploded, not earlier when the bullet exited. The autopsy physicians found at least part of the outline the exit wound, which was considerably smaller and on the right side of the head. After the bullet had exited from the right side of the head, pressure continued to build up isotropically inside the skull (equally in all directions) until the weakest part of the skull burst and formed the larger hole through which much of the right side of JFK’s brain matter was explosively ejected. Because the pressure increased isotropically (equally in all directions), the location of the burst cannot be tied in any direct way to the location of the exit wound; rather, it is related to weak points in the skull. This simple fact makes the whole back-of the-head argument fallacious. In the simplest possible language, finding the explosive wound in the back of the head does not necessarily mean that the bullet exited from the back of the head.
      Were the reports of the earwitnesses in Dealey Plaza represented properly? Anyone who claims that all the witnesses heard shots from the grassy knoll, as opposed to the depository, is mistaken. It turns out that most witnesses thought the shots were coming from whichever of the two places they were nearer to. There were enough reports of shots from each place to eliminate any meaningful consensus. People who argue strongly from the earwitness reports usually omit critical facts such as (1) a sniper with a rifle was seen in the Depository but not at the knoll; (2) the young black men on the fifth floor heard the bolt being operated and the shells hitting the floor above them, but there were no reports of bolts or bullets for the knoll; (3) a sniper was clearly seen firing from the Depository but none was seen firing from the knoll; (4) a weapon and shells were found in the Depository, none at the knoll; and (5) three fragments were traced ballistically to the rifle from the Depository, none to any other weapon.
      F’s second unstated assumption was that the reports of the wound and the shots were so clear and obvious as to admit no other interpretation than his. The previous paragraphs showed that the opposite was true for both types of evidence.
      The third unstated assumption was that the reports cited by F really did represent the majority of the witnesses. This assumption appeared to be correct for the wound but not necessarily so for the shots.
      So the witness reports, when examined carefully, are much less persuasive than proposed by F, and certainly not conclusive.

Does the Autopsy report carry epistemological weight?? The very same Autopsy report that describes the same BOTH head wound as the BOTH witnesses themselves described.

F.

(Here BOTH means “Back Of The Head.”) Of course the autopsy report carried weight. I don’t know where F got the idea that it described the same wound as his witnesses are alleged to, because it doesn’t. It described an occipitoparietal wound, which means one in the rear side, not the rear. It appears that F was trying to have it both ways here. The autopsists clearly described an explosive wound (as distinct from an exit wound—see above) that was ahead of the entrance wound. That’s all one needs to know in order to dismiss the talk about “back of the head.” It’s amazing how people try to cover simple truths with an abundance of words.

P.S. In order for Guinn to put a 99.9% probability on the CE-399/CE 842 match, would not the rigors of science demand that a proper random sampling be made. How many of the 2,000,000 WCC MCC bullets produced does science demand be tested and how many did Guinn test??

Here F misunderstood the background of both his questions. The first question, about the 99.9% probability, failed to mention that Dr. Guinn characterized the probability of the two groups of fragments being chemically distinct in very different ways at different times, including saying that it was almost impossible to put a probability on the groups. Assigning a probability of a mismatch requires that the nature of the underlying distribution (of concentrations of antimony in WCC/MC fragments) be known or estimable. The distribution can be obtained by randomly sampling a large population or fully sampling a small population. This makes the strict answer to his first question “no.” The second question was so ill-defined as to be nearly meaningless. There is no unique answer because the number to be tested depends on how well you need to know the distribution of concentrations, which in turn depends on the nature of their distribution. So you can’t predict how many samples are needed. (A later exchange on this same point revealed that F understood even less about the purpose of his random sampling than was shown here.)
      So in summary, there was almost nothing right in F’s entire answer. He misunderstood the meaning of “epistemology,” my position on witness testimony, what the eyewitness and earwitness testimony really represented, and the statistics of sampling bullets. Many errors for such a short answer!
      In the fourth message of this subthread, I pointed out some of these errors to him and asked him to rephrase it. I did not deal with the question on probability because it was not related to the thread.

Interesting answer, but to a different question than I had asked. You gave fatal flaws in MY epistemology, whereas your original statement and my question on it went to "epistemology," which means the general theory of knowledge. If people in newsgroups are to have any hope of communicating effectively, questions and answers must be phrased precisely. The difference between epistemology in general and a specific epistemology is huge. Care to rephrase your answer? Then we can go forward.

      He then declined to rephrase, and followed that with a sentence whose first part was wrong and whose second part repeated me without acknowledging it:

No, but I will clarify. My comments are directed your epistemology, and you may respond to them in light of that revelation. As you yourself have acknowledged, your model is not going to give a truly accurate answer without a reasonable and well though-out method by which individual eye, ear, and nose witness accounts are accorded weight and a rightful place among the, wherever that may be, in the evidential hierarchy.

His comments about my acknowledging the necessity of including witness testimony were wrong. As for needing witness testimony in order to get the right answer, recall the section above that noted how some of it may be used after-the-fact to get informal introductory guidance on topics not covered by strong evidence. The resulting answers are certainly not binding, however, nor could they be. This is far from requiring witness evidence.
      So F was seriously wrong again.
      I then responded, in the sixth message, to F’s errant portrayal of my suggestion on using weak evidence:

That's not what I said. The evidence is such that critical thinking (which is not a "model") cannot now say with certainty who did it, and probably will never be able to. Regarding witness evidence, one must on the one hand recognize its shadings (see my essay at http://karws.gso.uri.edu/PSC482G/Spring2000/Critical_thinking/Shadings_of_eyewitness_testimony.html ), but on the other hand recognize that none of it can be as strong as physical evidence. We get the basic answer from the physical evidence, and then maybe drop down to witness evidence to get a feeling for areas not covered by physical evidence. But feelings are not close to solid answers.

I also noted that I disbelieved such high probabilities and that Guinn had offered various statements on them:

You should remember that I have never supported probabilities that high. I also believe that the totality of Guinn's statements about probability should be cited, not just one. Most of his remarks were more moderate than that one. He also noted that matches do not prove origins. That part seems to have been forgotten.

      F then used the seventh message to challenge me again to answer his questions about sampling, and to propound several more erroneous ideas about NAA. Here is his second request for me to answer the unanswerable:

Again, you did not answer my questions. Again, I resubmit them in the hope that you will give me two direct answers. “How many of the 2,000,000 WCC MCC bullets produced does science demand be tested and how many did Guinn test??”

In response to my remark about not supporting such high probabilities, he wrote:

You said 99.999999% at the Providence Conference.

See my response, two messages later, for the true version of this erroneous statement.
      F concluded this message with a quick series of serious errors that showed that he knew very little about what he was talking about:

You know as well as I that, Guinn would have testified that bullet 6002A and bullet 6003A2 were the SAME bullet because they match better than CE-399 and CE-842. No amount of scientific reasoning can change that simple fact. NAA is not the answer.

As for F’s first sentence, nonscientist F needed to refrain from putting words into the mouth of distinguished NAA practitioner Dr. Vincent Guinn. Guinn is far too capable to have said any such thing or anything near it. F had evidently been taken in by the erroneous arguments of mathematician Stewart Galanor and nonmathematician Wallace Milam, which amount to saying that two distributions with different means must never overlap. Galanor and Milam have confused the actual task of distinguishing means with distinguishing individual samples, a serious error indeed. That allowed them to claim fallaciously that the NAA fails if two different bullets contain the same concentration of antimony anywhere inside them. For some reason, F chose to believe these two non-NAA people over the expert Guinn. Let us hope that it was not because the first two gave F the answer he wanted.
      As for F’s “No amount of scientific reasoning can change that simple fact,” let us simply say that the previous paragraph did change that simple fact, and quite easily to boot. F’s punch line “NAA is not the answer” was a statement of belief rather than of fact. The validity of his belief can be judged by how poorly he fared in the previous few paragraphs.
      The eighth message was again from F. It was a public response to a private message from A. The two had been passing strong messages privately and were here caught at it:

A: Is it just me, or is KR not answering ANY of the tough questions raised by any of the posters???

F: No, it's not you. KR is not fairing particularly well. He won't answer my Guinn question. He knows there is 1 answer. He just can't bring himself to say the words. Why...because his “re-working” is based upon Guinn's data. He was done before he started. :-)

They both were confusing not answering with not having an answer. They saw something not happening and immediately supplied what they considered to be the only reason for it. This very risky approach is taken by many JFK critics, and they nearly always turn out to be wrong, as A and F were here. They thought that the only reason I wasn’t answering was because I had no answer. When you think about it, this is a very egotistical assumption that amounts to saying they know every possible reason why I was not responding. It’s close to equating themselves with God. The truth in this case was more prosaic—I had spent nearly every waking moment of the previous few days preparing a poster for a national meeting in Washington, for which I was to leave the morning after they wrote their message. That activity got higher priority that answering them.
      The “reworking” mentioned by F referred to a passage in an earlier private message to him where I communicated that I was revisiting an idea from a year ago that might answer the remaining questions about the NAA data. But F was remembering wrongly when he stated that this was based on Guinn’s data. I had introduced that work at the Providence Conference and explained that the key to the improved interpretation was the FBI’s data, not Guinn’s. So much for witness testimony! Thank goodness for strong evidence from tape recordings of conferences! As for my allegedly being “done before I started,” time will be the best judge of that. Nonscientists should not try to predict what scientists can and cannot accomplish.
      In the ninth message of this subthread, I answered F in some detail. This was the night before I left for Washington, and I had time enough:

   You are right about one point and wrong about some other more-important ones. I stand corrected on my remark about not having supported probabilities as high as 99.9% that the two groups of fragments were different. In Providence I provisionally did, provisionally meaning that I was proposing a hypothesis that if true would lead to probabilities something in excess of 99.99% (but nowhere near the 99.999999% that you cited, which is closer to 1 by ten thousand times than mine). Recall that I even wondered aloud whether my hypothesis would survive the night. (I understand that opinions differed on whether it did.) But that hypothesis involves a very different interpretation of the source of the fragments than Guinn and others, including me, have proposed, so it isn't quite fair to lump its probability with the others. With respect to Guinn's data and their conventional interpretation, including the large heterogeneities, the probabilities are considerably less than 99.9%. So we are both right on that point.
   Now as to your question abut the number of bullets that would have to be tested to raise the probabilities to Guinn's 99.9%, there is no answer because the question is ill-posed. (That's a polite way of saying it doesn't make any sense.) It is ill-posed because testing more bullets doesn't change the inherent answer (the real probability) in any systematic way, but rather defines the heterogeneity and hence the final answer better. (It better defines the underlying distribution of concentrations of antimony, the "sample set" in statistical terms, and thereby gives you a more reliable calculation.) So you analyze more bullets to know the answer better, not to move it in any particular direction.
   I have taken this exchange as an opportunity to very briefly review my short presentation on NAA from last year's Providence conference. That plus another line of thinking that I have been recently pursuing are slowly but surely reinforcing my earlier idea that there may be other legitimate ways to view the NAA data that avoid the problem of heterogeneity. When I return from the conference that I leave for tomorrow and that I have been busily preparing for these last two weeks, I will take up this question again. If all goes well, I would like to have something to show by roughly midsummer (my midsummer night's dream, perhaps). These data have been trying to tell us something that we haven't been smart enough to catch, but I think I'm getting closer. Rearranging one's mentality (thinking about a problem from a different angle) can take more time than we like to admit. At least, it is often so with me. And in the best of worlds, I may not be “done before I started.”

In the tenth and last message of this subthread, F responded with meaningless carping. The passages from my previous message that he was responding to precede his answers.

KR: You are right about one point and wrong about some other more-important ones. I stand corrected on my remark about not having supported probabilities as high as 99.9% that the two groups of fragments were different. In Providence I provisionally did, provisionally meaning that I was proposing a hypothesis that if true would lead to probabilities something in excess of 99.99%

F: I don't recall the word “provisional” passing your lips at the podium in Providence. I do recall you saying “point nine nine nine nine nine” [0.99999] in a contemporanious setting.

Thank goodness for tapes of conferences! F was right, but only on a technicality. I had communicated the same idea by saying “hypothesis,” which I used five times in regard to my new idea about the NAA data. In reference to the same idea, Martin Kelly said “preliminary hypothesis” twice. I also said “I’m putting out what I think the explanation is,” and “That’s my hypothesis. I invite you to have at me.” These comments clearly mean “provisional.” With regard to the number of nines in the probability, the tapes also show that I quoted (provisionally) four nines, not the eight and the five nines that F remembered. How fitting that F, who has repeatedly championed the need to include witness evidence, has himself provided such a graphic example of the errors it often contains. His lapses of memory make the point better than I ever could.

KR: (but nowhere near the 99.999999% that you cited, which is closer to 1 by ten thousand times than mine).

F: Two too many zeros?? Perhaps, but what is the difference.

Sorry, F, but there is a huge difference. The four extra nines in F’s probability made it ten thousand times less likely that the groups of fragments were actually the same but just appeared to be different. One should not be cavalier about factors of ten thousand.

KR: Recall that I even wondered aloud whether my hypothesis would survive the night. (I understand that opinions differed on whether it did.) But that hypothesis involves a very different interpretation of the source of the fragments than Guinn and others, including me, have proposed, so it isn’t quite fair

F: It isn’t “quite” fair?? So “what” fair is it?? 

F was just trying to save face here. Let’s remove the problem and just say “it isn’t fair.”

KR: to lump its probability with the others. With respect to Guinn’s data and their conventional interpretation, including the large heterogeneities,

F: Heterogeneities?? Guinn expressed himself in homogeneities. Your word choice is interesting in light of Guinn’s.

Guinn misspoke—“homogeneity” expresses an absolute condition akin to “pure.” As with all absolutes, “homogeneity” cannot have degrees. “Heterogeneity,” by contrast, is not an absolute, and can have degrees—the composition of one object may vary more than that of another object. Some people speak of homogeneities when they wish to emphasize that all the objects are nearly homogeneous, but that doesn’t make it right.

KR: the probabilities are considerably less than 99.9%. So we are both right on that point.
   Now as to your question abut the number of bullets that would have to be tested to raise the probabilities to Guinn’s 99.9%,

F: I never said Guinn raised them to 99.9%. I mentionn only his HSCA work.

The passage did not say that Guinn raised them, either. F was being hasty.

KR: there is no answer because the question is ill-posed. (That’s a polite way of saying it doesn’t make any sense.) It is ill-posed because testing more bullets doesn't change the inherent answer (the real probability) in any systematic way, but rather defines the heterogeneity and hence the final answer better. (It better defines the underlying distribution of concentrations of antimony, the “sample set” in statistical terms, and thereby gives you a more reliable calculation.) So you analyze more bullets to know the answer better, not to move it in any particular direction.

F: So a political pollster is wasting his/her time by random sampling 1500 people??

And Guinn does not need to sample a requisite number of WCC MC bullets??

F should have left my answer alone. By revisiting the subject, he revealed how little he understood about statistics. F’s original question asked how many bullets Guinn would have needed to sample in order to raise the probability to 99.9% that the means of the groups could not have differed so much by chance. As explained above, however, that’s not the reason for sampling more bullets. The reason is to better define the underlying distribution(s) of antimony in them and thereby get a better estimate of the real probability. For example, if the real probability is 85%, I want my range to be 84–86% rather than 70–100%. That’s what more samples give you. The pollster samples more people for the same reason. A poll of that gives 55—45 for two candidates means much more with uncertainties of 1% than of 5%. The former would mean a clear difference, the latter would be very close to the popular phrase “statistical dead heat.”

KR: I have taken this exchange as an opportunity to very briefly review my short presentation on NAA from last year’s Providence conference. That plus another line of thinking that I have been recently pursuing are slowly but surely reinforcing my earlier idea that there may be other legitimate ways to view the NAA data that avoid the problem of heterogeneity.

F: Again, you couch your view in terms of heterogeneity. You and Guinn seem to be at odds. When you get back, please give us your take on the differences.

F’s use of “take” assumes erroneously that the differences between “homogeneity” and “heterogeneity” are just a matter of opinion. In fact, they are well understood and can be found in any dictionary. I gave the right answer above.

KR: When I return from the conference that I leave for tomorrow and that I have been busily preparing for these last two weeks, I will take up this question again. If all goes well, I would like to have something to show by roughly midsummer (my midsummer night’s dream, perhaps). These data have been trying to tell us something that we haven’t been smart enough to catch, but I think I’m getting closer.

F: I await your data.

F misunderstood again. I gave the data, or at least an outline of them, at the Providence conference a year before this exchange. At that time, I stressed that the data were not new—they went back to 1964. Obviously, it’s the new interpretation, or understanding, of the data that counts.

KR: Rearranging one’s mentality (thinking about a problem from a different angle) can take more time than we like to admit. At least, it is often so with me. And in the best of worlds, I may not be “done before I started.”

F: And you may well be. :-)

Once again, the nonscientist was listening to the opinions of other nonscientists and thinking that they were more reliable than those of scientists who specialize in this field (NAA). F, of course, is not the only person within the JFK research community who does this. It’s common. It’s also egotistical and almost always wrong.

F: I missed your response to this; “You know as well as I that, Guinn would have testified that bullet 6002A and bullet 6003A2 were the SAME bullet because they match better than CE-399 and CE-842. No amount of scientific reasoning can change that simple fact. NAA is not the answer.”

F was thinking that I was avoiding his question because I had no answer. Again he was committing the error of trying to interpret the lack of an action, in this case interpreting the lack of an answer as meaning that none existed. Without extra information, one can never know the reason for the lack of an action. F could not have known that particular outside information because it was in my head. I didn’t answer F because his statement showed that he knew so little about the basics of probability and statistics that it would have required a major effort to bring him up to a level where he could understand the answer. That was fundamentally his responsibility, not mine. I gave a brief answer above, however.
      That ended the second and longest subthread of Thread 5. It was marked by nearly total confusion on the part of A and F about the issues of epistemology, weak evidence, planted fragments, and NAA.

      The rest of Thread 5 consisted of two responses that were not followed up by anyone, and then subthreads of six and three messages. The first unanswered message was from A to me. He proposed a big NAA study of WCC/MC bullets and others in order to see, once and for all, the extent to which we can distinguish from chemistry alone which fragments come from which bullets:

I have already explained what you claimed to H was a major oversight on my part. The post digressed really into a general critique of your entire approach on your site. Everything I said was in some way or form advocated in one article or another on your web site, in which I have read almost every article posted.

H mentioned that you should actually do something to replicate the acoustics evidence. I would rather propose something directly in your field, and something which you hold near and dear to your heart. Why don't we do a blind Neutron Activation study with a large sample of MC bullets, as well as other Winchester bullets, as well as bullets from completely different manufacturers? You seem to elevate the NAA into the perhaps the strongest evidence in the case. Why don't we see if you can distinguish, blindly, which fragments came from which bullets and to what degree of accuracy?

I didn’t respond to this message because (a) I was busy with last-minute preparations for the trip, and (b) the new interpretation of the NAA data will likely make such a study unnecessary, and maybe even inappropriate.
      The second unanswered message was I replying to my asking F to use “epistemology” properly. He submitted three illogical, illiterate, irrelevant, and incomprehensible paragraphs that were distinguished by their misunderstanding of “epistemology” and their unjustified assumption that Oswald was framed:

Amen to that, bro'....given that by itself on its own merits wihout regard to the validity of alternate theories in explanation=your theory fails the test of the evidence. Thats a 'general' epistemology(called 'common knowledge' on the street)..and the specific epistimilogical question that next arises would be to ask= why is it that it was still a rifle found as against Oswald in evidence? Thats the only rifle found. Any thorough review of your theory's use of it as against Oswald is a failure(back to common knowledge)..and to get more specific, the next question is why it still appears to implicate him?

Here I erroneously used “epistemology” to mean “knowledge” rather than “theory of knowledge.” He brought Oswald into a discussion where he didn’t belong (because the essay didn’t mention him). He erroneously equated general nonconspiracy and nonconspiracy with Oswald as the gunman. Beyond this, his sentences were gibberish.

Thats a frame= the leaving of evidence appearing to implicate someone we know did not do as alledged=so to get specific, you might start wondering why Oswald was framed in the way in which it was done. The found rifle IS the evidence you say you seek..of conspiracy..someone placed it under the boxes, and it wasn't Oswald;your story on that failing.

This paragraph used “frame” instead of “frame-up.” It stated without evidence that Oswald couldn’t have placed his rifle under the boxes. Beyond this, the paragraph was marginally comprehensible.

Wanna' move forward? Start asking yourself why Oswald was framed in the way in which it was done. Thats the specific epistimology, in this case.

Another assumed frame-up and wrong use of epistemology, both couched in seedy grammar.
      All in all, these paragraphs represent a pitiable and failed attempt to rebut my comments to F on epistemology. If people who post on these newsgroups want their opinions to be taken seriously, they need to learn the English language.

      Then began a subthread with six messages, all but the last being a dialog between JM and F about whether the lead fragments were MC and whether any had been planted. This topic also had nothing to do with the original essay, which still had not been discussed. JM begins by asking F whether he agrees that all the fragments were MCs and all the weapons were, too:

F, I will let you and KR discuss epistemological method. The closest I can come to on that is episiotomy, but no babies are being born here. However, I acknowledge there is some conflict with Guinn's probablilty studies, but do you agree at least all fragments and bullets found were of the MC variety. So does that mean in your shooting scenario, and I am not sure what that is -?Daltex?- do you agree all of the weapons were MCs?

      F answered in a second message that amounted to yes, no, and maybe. Here are his answers to JM’s questions:

JM: but do you agree at least all fragments and bullets found were of the MC variety.

F: All of the fragments found are not available for testing so the answer has to be know [He means “no”—KAR]. All the extant fragments would seem to be WCC MC. However...

This was F’s “no.”

F: ...the funny thing about that is that there are other bullets with the same composition. Rare, but available. I think it unlikely any conspirator would choose a second set of ammo based on the similarity of antimony and silver. Having said that, I think the same WCC MC ammo was used.

This was F’s “yes.”

JM: So does that mean in your shooting scenario, and I am not sure what that is -?Daltex?-

F: Could be.

JM: do you agree all of the weapons were MCs?

F: There is no other kind of extant ammo evidence. That is all that can be said with certainty. The FBI has a history of playing with evidence, both in this case and others. Anything is possible when the FBI prepares Oswald's address book for the WC as an Exhibit and leaves out only one page. The one with FBI agent Hosty’s name and license number on it. They fiddled with that evidence. What else did they fiddle with??

This was F’s “maybe.”
      In the third message, JM reminded F that claims of planted fragments require positive evidence, and then pressed him for a more concrete answer as to whether any additional purported weapons must have also been MCs:

Well, I am not one who believes every person in a department, FBI or otherwise is always perfect, but we have received fair explanations of the fiddling we do know about, but will agree this does not rule out or prove other fiddling went on. But to go down the road you suggest above needs substantiation. If all fragments tested are legitimate, which I accept until otherwise proven, then we know the larger pieces that could be tested were from Oswald’s MC. If the other fragments that were MC type bullets, but too small to match ballistically came from another weapon, it still indicated it was also an MC. Is that a fair statement? 

      F in the fourth message gave the right answer but for the wrong reason:

Yes, it is. I would think that any conspirators attempting to frame a rifle, and thus its owner, would use like ammo. Barring an absolute confidence in the ability to switch evidence after the fact, it would seem to be a must.

Rather than taking the logical approach of using the low concentrations of antimony in the smaller fragments to show that they came from Mannlicher-Carcano rifles, F invoked an invalid conspiracy argument. He essentially said that the small fragments were from Mannlicher-Carcano rifles because the conspirators knew that would be the best way to frame Oswald. The problem with this long way around, of course, is that it is based on unsubstantiated premises—you have to know that (a) there was a conspiracy, (b) the conspirators used MC rifles to deliberately frame Oswald, and (c) they had to be sure that no fragment large enough to be traced ballistically would be found because it would betray the existence of a second rifle and hence a conspiracy before you can use this argument. As of now, we know none of these three. Note how point (c) makes this scenario impossibly risky for the alleged conspirators. F gave the right answer for the wrong reason.
      In the fifth message, JM appeared to misunderstand F’s position with respect to planting fragments:

Just as planting a bullet without knowing it would really be needed seems illogical >grin<.

F was referring to conspirators shooting with MC rifles in order to frame Oswald, whereas JM thought he was imagining bullets and fragments being planted afterward. I thought with JM initially, too, but understood after reading these messages several times.
      In the sixth message, J followed up on JM’s remark by proposing two reasons why conspirators could have felt they could have planted CE 399 without getting caught (numbers inserted to clarify the two reasons):

JM: Just as planting a bullet without knowing it would really be needed seems illogical >grin<.

J: If you mean 399, the assumption would only have to be that the bullet would be accounted for by the wounds. What are the principle sources of bullet fragments in a shooting like this? The bodies and the car. JFK’s body and the car were both whisked back to Washington. 399 entered the evidence stream very early. When did the other fragments enter the stream? [1] Maybe the conspirators were confident that they could account for 399 and that they could control which other fragments became evidence. [2] Perhaps the reason for planting 399 was to ensure that the rifle was identified early on and to have one piece of evidence linking the rifle to the shooting that was not coming out of the White House garage.

If, according to reason 1, the conspirators thought they could “account for” CE 399, they were mistaken—its provenance has been disputed almost from the time it was found. If, according to reason 2, they wanted to be sure to link Oswald’s rifle to the shooting, they fared somewhat better.
      J’s discussion forgot the power of the NAA results to show that CE 399 was genuine.

      The final subthread of Thread 5 consisted of three messages linked by thoughts about witnesses and statistics. In the first message, K seconded F’s position on the utility of witnesses by presenting a contrived example that did not support the general idea:

Eyewitness testimony has its value. There is a shooting on the street leaving two people dead. There is no evidence found on the street indicating perpetrators. 5 eyewitnesses tell you that the shots came from high above. You check the angle that the bullets went through the victims and then reconstruct the angles backwards. They lead to the roof of a nearby building. You find no evidence on the roof. No cartridges left behind. You assume that the shooters retrieved all of the empty cartridges so as not to leave behind any incriminating evidence. Then, being a thorough investigator you search every possible location where a stray cartridge could have rebounded into. You even pour water into the gutter and voila a spent cartridge washes out at the bottom. The markings on that cartridge match cartridges from a previous case where the defendant was acquitted because the star witness was murdered the day before she was going to testify.

Your eyewitness do not prove who committed the murder, but without following where they lead you might never have found the cartridge.

K then used a non sequitur to address F’s earlier insistence that the “back of the head” witnesses are much more important than the mysterious “epistemological method” acknowledges:

Just because many witnesses describe damage to the back of the head does not tell us whether that must be an entrance wound or an exit wound.

Other than not really responding to F’s point, K’s remark ignored the fact that entrance wounds are not huge, oozy, and bloody. Telling whether the wound in the back of JFK’s head was one of entrance or exit was the least of the problems.
      K followed up this strange comment with one nearly as strange. In response to F’s question about the number of WCC/MC bullets that should have been tested, K noted that the underlying population (which he calls the “natural occurrence”) needs to be known. The problem is that he then likened the task to determining whether a deck of cards has four aces or 52:

In order to make a statistical statement about an event, one needs to know the underlying natural occurrence. What are the chances of drawing 4 aces from the top of a deck of 52 cards? That depends on what type of deck it is. A regular bridge deck? Or a deck of all aces?

Right idea; terrible example.
      F used the second message to briefly agree with K (ten words for two agreements) and then to try to get K to answer the same question on the number of bullets to analyze that he didn’t get from me:

K: In order to make a statistical statement about an event, one needs to know the underlying natural occurrence. What are the chances of drawing 4 aces from the top of a deck of 52 cards? That depends on what type of deck it is. A regular bridge deck? Or a deck of all aces?

F: What is the proper sampling number in a population of 2,000,000??

K didn’t answer him because K was not trained in statistics.
      To end this subthread and thread with a third message, I chipped in again with a mega-non sequitur in which he “gives” the number of bullets by touting the perfection of nosewitnesses:

F: What is the proper sampling number in a population of 2,000,000??

I: Asuming the witnesses were alive at the time, the sampling is almost always perfect=when it comes to the 'nosewitnes'. You guys were knocking about the validity of eye-witness or ear witness=what about the "nose-witness'?

The olfactory sense is the most reliable=people alive have to be breathing=thats why the sense of smell is recognized as so closely connected to the memory of things. Live pople at the time may not see if they were not looking; may not hear if they were not listening..but no one fails to smell unless they were not breathing at the time.

At the time in question here, we know two things by the most reliable of witness. No one at all anywhere in the TSBD building smelled anything at all anywhere in the building of the smell of gunpowder. Given the reliability of the involuntary nature of response, if there had been such a smell anywhere at all in the building, some one's nose would have reported it. No one's did.Thats the first point.

Equally=given the involuntary nature of response= all those who responded with a memory of the smell ALL uniformily reported that smell of gunpowder down on the street. They were not even asked what they smelled= they volunteered the information=consistent with the involuntary nature of the sense in question. One such=who prefaced that he was fammiliar with fire arms=said he "always thought that was strange".

The less said about this nonanswer, the better. Maybe we should pretend that it never happened.
      Mercifully, that ended Thread 5. Overall, it completely avoided the central points of the essay and replaced them by bullets and magic, wrongly used “epistemology,” wrong remembrances of the Providence Conference by vocal advocates of witness testimony, nonsensical statistical questions, contrived scenarios about planting bullets and fragments, and strange claims for nosewitnesses. All in all, this thread went nowhere.

Ahead to Thread 6
Back to Thread 4
Back to Anatomy Of A Newsgroup Discussion