Dan posted Tue May 24, 2011 10:00 pm
viewtopic.php?p=458490#p458490In historiography, ad hocs are “suppositions about the past which are not already implied to some extent by existing beliefs” (McCullagh
http://en.wikipedia.org/wiki/Historical_method). That is, suppositions that have no evidentiary basis. They are quite rational, although they are unfalsifiable. Historical interpretations with fewer ad hocs are given preference.
I mentioned I'd post this week, but I am going away this weekend leaving tomorrow, so I won't be able to respond until next week.
To start off, where Dan does McCullagh ever say that one should choose as best explanation , “Historical interpretations with fewer ad hocs are given preference.”? He doesn't.
Let’s see what he/she says: From Justifying Historical Descriptions by C Behan McCullagh
P27
So far it has been shown that hypothesis of very great explanatory scope and power are accepted as true. Are the other criteria, about plausibility, ad hocness and disconfirmation relevant, and if so why? The requirement that hypothesis should be as little ad hoc as possible is simply to keep to a minimum the number of statements in hypotheses which are not implied by known facts, and which are therefore more especially likely to be false. The fewer ad hoc statements a hypothesis includes, the less likely it is to be false. On the other hand, the more plausible the statements in a hypothesis are, that is the more they are implied by existing knowledge, then the more likely they are to be true.
Ad hocness and lack of plausibility are not reasons for thinking a hypothesis false, or at least they are not reasons for thinking it is more likely to be false than true.
P 28 The conditions which bear upon the acceptability of arguments to the best explanation can be divided into three groups: those about the explanatory scope and power of a hypothesis; those about the ad hocness of the hypothesis; and those about facts which render a hypothesis implausible which disconfirm it. There is a sense in which the third group is the most important. For even if a hypothesis is of greater explanatory scope and power than another, if evidence incompatible with it cannot be explained away satisfactorily, then it is abandoned. Historians simply assume, as do most people, that the world is logically and materially consistent, so that for beliefs about it to be true, they must refer to compatible events and states of affairs. The first group of conditions, about explanatory scope and power, are the next most important, for if two hypothesis are not implausible or disconfirmed, then even if one is more ad hoc than another if it has greater explanatory scope and power it will be preferred.
Couple of things to note. Your say so, and opinion that many of the explanations of the S/R theory are ad hoc is simply that ..your opinion..they are not ad hoc. For example, the evidence indicates another manuscript existed besides the Manuscript Story Conneaut Creek. Your accusation that the hypothesis of another existing is "ad hoc" fallacious ..is nonsense. You have been using your own distorted fallacious reasoning throughout this entire thread.
Ad hoc is when counter evidence against a hypothesis, not mere opinion (such as your say so) refutes a hypothesis and then to rescue that hypothesis from the counter evidence an explanation (ad hoc)is thought up but is unjustified other than its purpose is to rescue the hypothesis.
Your say so that another manuscript didn’t exist…is not counter evidence. Your Book of Mormon witnesses are not reliable witnesses despite your reliance on apologist Anderson's say so. So let’s get that straight …you don’t even know what “ad hoc” is...or you do but deliberately are misusing it. You also didn’t know what Occam’s Razor was in past discussions because you didn’t appreciate that it was between hypotheses with equal explanatory power. This is a key concept you fail to acknowledge. You seem to not appreciate how explanatory power fits in and is necessary both for Occam’s Razor and for arguments to best explanation by McCullagh.
Although I have criticism of McCullagh explanations, I’ll leave that because it’s not essential for our discussion.
(I'm going to assume McCullagh is a 'he' for this discussion) In order to follow his reasoning better, he illustrates with an example of 3 competing theories to show how to decide which one leads to best explanation. I'll quote from the book below for anyone wanting to follow what McCullagh argues. What is interesting is that the 3 hypothesis are similar in many respects to the 3 hypothesis of the Book of Mormon.
Hypothesis # 1 ..King William killed by accident ..or K/A. This theory parallels in many respects to the Smith alone theory or S/A- both are the simplest of the 3 theories..least suppositions
- both most widely believed and accepted as long as one doesn’t look at other details and further reasoning inferred.
- both do not involved a conspiracy.
- both do not have independent objective witnesses.
- both have witnesses with vested interests and motivation to suppress truth if conspiracy involved.
Hypothesis # 2 King William killed by witchcraft or K/W..this theory parallels in many respects to the Smith divine theory or S/D
- both involve highly implausible extraordinary claims with no evidence provided for verification
Hypothesis # 3 King William killed by murder or K/M conspiracy …as a result of a conspiracy between the King’s brother Henry and W.Tirel ..this theory parallels in many respects the Smith-Rigdon conspiracy. K/M theory- explains Henry’s extraordinary prompt action following death of king to seize royal treasure which was conveniently close to New Forest at Winchester and being anointed and crowned king 3 days later
- a quick coronation could hardly be secured without preparation and without a nucleus of supporters already formed
- explains why Tirel’s wife’s family, the Clares, were so well patronized by Henry after he became king, and why many bishoprics and abbeys, previously retained by the crown were bestowed upon the families of his supporters.
- This hypothesis is consonant with Henry’s known disposition to cruelty and is itself perhaps explained by his declared belief that he deserved the crown rather than his brothers, since he alone had been born when his father was king.
The S/R theory Explains
- Rigdon’s extraordinary immediate conversion and and quick rise to power
- just as the church explains how unlikely it was that Smith could or even would have written the Book of Mormon …a better explanation it that it involved input from different people, more knowledgable, more interested in writing such a book and more capable at the time than Smith
- the many and varied witnesses who recall a spalding manuscript which matches closely to historical parts in the Book of Mormon..credible witnesses with no motivation or benefit to lie
-why only small knit group essentially family were involved as witnesses during the writing process
-why a blanket was used to conceal smith from public view
-
(I could go on, but Roger or Dale would be much better in all describing all the things from evidence the S/R theory explains. It is the theory which has greatest explanatory power and therefore argument with best explanation.)
Note, the 3rd theory, the conspiracy that King William was murdered has the most suppositions and according to your reasoning Dan it should be rejected and the simplest theory that the King was accidentally killed should be accepted as the best explanation. Fortunately though both Brooke and McCullagh are better critical thinkers than yourself. Thank goodness for that.
Theory 3, the conspiracy theory that the King was murdered, the one with the most suppositions is the one preferred by Oxford historian Brooke and although it is acknowledged as not being proven it is the one which best explains the evidence. McCullagh does not call those explanations for # 3 ad hoc, nor does he say it should be rejected on that basis. The simplist theory # 1 that the King was accidentally killed and the most straightforward most widely believed and accepted ..si not the one chosen as the theory to best explanation. # 1, doesn’t explain the evidence as well as # 3. McCullagh while acknowledging # 3 is not proven, argues it is the best explanation.
On this basis Dan, the reasoning in McCullagh’s book (he’s a philosopher of history not an historian) favors the S/R theory not the Smith Alone.
I haven’t written the details for the 3 theories from McCullagh's book, instead I'll post directly from the book so as to not take his words out of context and anyone interested can read it for themselves.
P 21- An interesting example of inference to the best explanation is provided by Christopher Brooke’s discussion of why William II (William Rufus) died, in his book “The Saxon and Norman Kings” (1963, pp. 178 - 96). William II died when he was shot by an arrow during a hunt in the New Forest in Hampshire on August 2 1100. Brooke says that the two fullest accounts of the incident, by William of Malmesbury (c. 1125) and Orderic Vitalis (c.. 1135), both declare that the arrow was fired by Walter Tirel, and that it was aimed at a stag and hit the king by accident. The truth of this story would help to explain both why those reports narrate it and why the king died, so we can take it as the first explanatory hypothesis to be considered. There are two quite reliable sources, however, which report Walter Tirel’s denial that he killed the king, or indeed that he was in the same part of the wood as the king when he died. And since the king was generally judged to be a wicked man Brooke believes Tirel would not wish to deny that he had been God’s instrument in killing him, if that had been the case and he had not murdered him deliberately. So Walter Tirel’s denials cast doubt upon the truth of the hypothesis. As Brooke says: ‘If William of Malmesbury was right in saying that the king and Walter were alone, only Walter could have told the story. Yet Walter firmly and frequently denied it. The contradiction is baffling’ (p180).
A second, different, hypothesis has been suggested by Dr. Margaret Murray in her book, “The God of the Witches”, and it is this which Brooke considers next. Chronicles of the king’s death report many strange portents of its coming. The night before he died the king dreamed of his death, as did a monk; and it is said that St. Anselm knew of it in France immediately by angelic messenger. Dr. Murray thought the king was a devil worshiper, and attributed his death to witchcraft. Brooke comments: ‘The cunning of Dr. Murray’s explanation is that it does, indeed, explain the warnings and prognostications, even the king’s dreams. Everybody knew he was going to die! Unfortunately, her explanation involves a number of assumptions for which there is no evidence, and which are wildly improbable’ (p181).
P 22
There no evidence of William II’s being a devil worshiper or involved in witchcraft; it is possible that Luciferians existed in the 11th century, since we have evidence of their rites in the 12th century, though only in Germany which William did not visit. Brooke thinks the stories of portents surrounding the king’s death are better explained as embellishments typically added by the ecclesiastical chroniclers to express their sense of divine judgment upon the man, at whose hands many of their houses had suffered. Brooke concludes: ‘Her theory, in fact, can be dismissed as fantasy’. (page 185).
The third hypothesis is one which Brooke puts forward himself, namely that the king was murdered as a result of a conspiracy involving his younger brother and successor, Henry, as well as Walter Tirel. This would explain Henry’s extraordinarily prompt action following his brother’s death in seizing the royal treasure, which was conveniently close to the New Forest at Winchester, and then being anointed and crowned king in Westminster three days later, on August 5. This was just a few weeks before his elder brother Robert returned from the First Crusade and sought the throne for himself. In Brooke’s opinion such a quick coronation could hardly have been secured without preparation, and without a nucleus of supporters already formed. The conspiracy hypothesis also explains why Tirel’s wife’s family, the Clares, were so well patronized by Henry after he became king, and why many bishoprics and abbeys, previously retained by the crown were bestowed upon the families of his supporters. Fear or hope of favour, might even have moved the early chroniclers tactfully to overlook Henry’s part of William II’s death, if they knew of it. This hypothesis is consonant with Henry’s known disposition to cruelty and is itself perhaps explained by his declared belief that he deserved the crown rather than his brothers, since he alone had been born when his father was king. And finally, this hypothesis would explain why, although it was agreed shortly after William’s death that Walter Tirel had shot him, Tirel himself denied any role in what he presumably knew to be a traitorous plot.
Thus the third hypothesis is contradicted by no observable data. It certainly explains more than the first, and seems less ad hoc than the second so it is the one to be preferred according to the present theory. But Brooke is too good a scholar to think as some philosophers of history would have him think, that the third hypothesis is therefore acceptable. ‘The circumstantial evidence against Henry can be piled up; but of positive evidence there seems to be none. The verdict must be “not proven”… The most we can say is this: if Rufus’s death in August 1100 was an accident, Henry I was an exceptionally lucky man’ (pages 195, 196).
The seven conditions whose satisfaction warrant belief in the truth of a historical hypothesis help us to understand why Brooke preferred the third hypothesis, his own to the other two.
P 23
(1) All those hypotheses explain the death of William II as it was reported by early chroniclers. The three hypotheses are incompatible, however. The first is that Walter Tirel shot the king by accident; the second is that he died (presumably by an arrow) because of witchcraft; and the third is a that Tirel shot the king deliberately as part of a conspiracy. Only one of these hypothesis can be true, though possibly none is.
(2) Explanatory scope. Each hypothesis explains additional facts besides the report of the king’s death. The first explains why William of Malmesbury and Orderic Vitalis recorded the details of Walter Tirel’s action the way they did; the second explains chronicled reports of ‘the warnings and prognostications, even the king’s dreams’; the third explains the promptness with which Henry secured the royal treasure and the throne for himself after King William’s death, Henry’s munificence to Tirel’s wife’s family, the Clares, and to other supporters shortly after ascending the throne, and even Tirel’s denial of any involvement in King William’s death. The third hypothesis explains three kinds of evidence, whereas the first two explain only one kind each, in addition to the death of King William.
(3) Explanatory power Although the first and third hypotheses imply the facts they explain quite strongly, the second does not. If King William did died by accident, as the early chroniclers said, then it is quite probable that they would have said so in writing their chronicles. And if his death was the result of Henry’s conspiracy to seize the throne before the return of his brother Robert, then, too, the other facts about Henry’s behavior which Brooke describes were quite probable. But the king’s being a devil worshiper, which is the second hypothesis, does not make it likely that he would have been killed by an arrow fired by Walter Tirel, or, more to the point, that his death would have been recorded in this way. There is no reason to think that the existence of devil worshiping and witchcraft made the king’s accidental death at all likely.
P24
(4)Plausibility. The accounts of William of Malmesbury and of Orderic Vitalis make it to some degree probable that the first hypothesis is true, since chroniclers often tell the truth, but reports of Walter Tirel’s denial that he killed the king by quite reliable sources, make the hypothesis improbable. The first hypothesis, therefore, is neither plausible nor implausible.
As for the second hypothesis, a decision about whether the evidence which it explains also renders it probable to any extent, depends upon one’s views of the occult. Do dreams and portents of events which subsequently occur make it likely that evil powers were at work, or not? If the answer is that they do, then the reports of those dreams and portents do confer plausibility upon the second hypothesis; but if the answer is negative, then the reports of those dreams do not contribute to its plausibility. A belief that, quite generally, evil powers of the devil and of witches do not exist, would actually render the second hypothesis improbable, and so implausible, though this is not an argument Brooke uses. Instead he argues that as it is improbable that King William ever met any Luciferians, it is also improbable that the King was engaged in devil worshiping, as the second hypothesis states. Brooke clearly judges the second hypothesis implausible, “wildly improbable”, for this reason.
The events which the third hypothesis explains do not make it at all probable. The fact that Henry I secured the throne quickly after William II’s death does not imply that he conspired to murder him. The only facts which render this hypothesis plausible are Henry’s known disposition to cruelty combined with his belief that he deserved the crown, indeed the horror and danger of such a plot must render its existence somewhat improbable. Thus the third hypothesis is really no more plausible than the first.
(5) Ad hocness As none of the hypotheses is clearly rendered probable by the total available evidence, all must be regarded as ad hoc, specially created to account for reports of King William’s death. It is tempting to say that the second hypothesis is more ad hoc than the others because it is so implausible, and the third hypothesis is less ad hoc than the others because of its much greater explanatory scope. But that would be to confuse ad hocness with other characteristics of the hypothesis.
(6) Disconfirmation Brooke clearly believed that Tirel’s denial of having shot the king ‘contradicts’ the first hypothesis. His argument is that if the first hypothesis were true, Tirel would have admitted killing him, and of having thus rid the world of the king who generally was judged to be wicked. This implication of the hypothesis is disconfirmed by the evidence, however, for reliable sources report Tirel’s repeated denial of involvement in the king’s death.
Brooke can point to no facts which disconfirm the second hypothesis, except perhaps the absence of any evidence of King William’s involvement in devil worship or witchcraft. He is inclined to use this lack of evidence to emphasize the ad hocness of the hypothesis, as one which is implied by no known facts and ‘can be dismissed as fantasy’. It could also be used to disconfirm the hypothesis, if it were believed that the truth of the hypothesis made probable the existence of some evidence in the chronicles of the king’s reign of his involvement in devil worship and witchcraft. The absence of such evidence would then contradict this implication, thus disconfirming the hypothesis, or the additional assumptions about the knowledge and interests of the chroniclers.
P 25
Brooke mentions no facts which disconfirm the third hypothesis. Notice how Tirel’s denial of involvement in the king’s death, which might have been thought to be incompatible with his participation in the plot, is cleverly accommodated by extending the hypothesis in such a way as to demonstrate its compatibility. If there was a conspiracy, then Tirel would want to deny any part in it for fear of being charged with treason.
(7) Relative superiority The third hypothesis is clearly superior to the other two on the criteria which have been described. It has much greater explanatory scope than they; it is not disconfirmed as the first is; and it is not implausible as the second is.
Why, then, does Brooke not accept that it is true? It may be observed that the third hypothesis does not so far exceed the others in all the conditions listed that it could not quite easily be proved false. One or two well-respected pieces of counter- evidence, such as a reliable report of Henry’s grief and anger over his brother’s death, plus a reported confession by Tirel that he had denied shooting the king for fear of punishment under the law, the accident being such a remarkable one – these would be enough to make the hypothesis less acceptable than the first. a historical hypothesis which could be so easily overturned does not deserve to be believed.
Brooke’s own reasoning for refusing to affirm the truth of the third hypothesis, is different. It is that the evidence for that hypothesis is ‘circumstantial’ and not ‘positive’, so ‘the verdict must be “non-proven” ‘. His meaning is clear. Although the third hypothesis implies and thus explains quite a lot of evidence about Henry’s behavior, the evidence of his behavior does not itself imply the truth of the hypothesis. Evidence which is implied by a hypothesis but does not imply the hypothesis is circumstantial. Evidence which implies the hypothesis is direct. The interesting question is, why do historians not regard circumstantial evidence as implying the truth of their hypothesis, and why do they prefer direct evidence?
The basic reason for historians doubting the value of circumstantial evidence is probably their awareness that there might well be other explanations of the data they are trying to explain besides the explanation they have suggested, and that one of these other explanations might well be the true one. (Examples of this argument are given by Karlman, 1976 pages 44 - 9.) Just because an historian cannot think of any better explanation does not mean one does not exist. Kitson Clark has shown that this fact is also respected by courts of law. He said that a jury in Liverpool in 1931 condemned a man, George Henry Wallace, for murdering his wife because the prosecution’s hypothesis that he had done so was not matched by any cogent alternative. However a Court of Criminal Appeal quashed that jury’s verdict completely ‘on the ground that there was insufficient positive evidence to support a conviction’ (Clark, 1967, pp 59 to 60).
Positive evidence, on the other hand can give both the historian and the lawyer good reason for thinking that an explanatory hypothesis is true. Taken together with an appropriate generalization, positive evidence implies the truth of a hypothesis in a way that circumstantial evidence does not. If Brooke had had a letter describing Henry I’s conspiracy to murder William II and seize the throne, written by one of the conspirators to a friend, then assuming that people generally tell their friends the truth, especially if it is against themselves, Brooke could have inferred that his hypothesis was probably true. Without such evidence, however, he knows it to be a figment of his fertile imagination, and perhaps very far from the truth.
It looks as if arguments to the best explanation are not much more useful than simple hypothetico–deductive arguments after all. The same reason we have for doubting the truth of the conclusions of hypothetico–deductive arguments has turned out to be a reason for doubting the truth of the conclusions of arguments of the best explanation as well. Does superiority of an explanation in respect of conditions 2 to 6 not insure its truth?
The short answer is that it does not, but if condition 7 is satisfied as well, that is if one hypothesis is so far superior to the others that can be thought of in respect of conditions 2 to 6 that there is virtually no chance of it being proved false soon, then it is reasonable to believe it true. In particular, if the scope and strength of an explanation are very great, so that it explains a large number and variety of facts, many more than any competing explanation, then it is likely to be true. There are some historical events which help to explain so much in such a variety of evidence that there is good reason to believe they could not be superseded by a better explanatory hypothesis. These are what Kitson Clark has called
‘The public facts’ of history, that is those facts which are so woven into the texture of human history that, unless human affairs are an illusion and all history is false, they are not false.
… For instance there can be no doubt that the battle of Waterloo was fought on Sunday, June 18, 1815, and there can be no doubt who won the battle. If the common opinion on either point was wrong then most of the history of Europe in the 19th century would be inexplicable, and a great many contemporary records which allude to the battle, describe the battle, or are concerned with the results of the battle have to be rejected as spurious. (1967, pages 41 to 42)
p 27
There can be no doubt that hypothesis whose explanatory scope and power far exceeds that of any other incompatible hypothesis about the same subject are generally accepted as true. Is interesting to ask why this should be so. If such an explanatory hypothesis is not itself implied by the evidence it explains, then why should it be accepted as true of the world? One reason is that when inferences are made about things which can be checked by observation, then they are often proved true. Thus faith in the reliability of this form of inference in yielding true conclusions has been reinforced. But the reliability of inductive procedures cannot be proved without assuming them. So it would not be irrational to ignore the conclusions of such arguments to the best explanation, if one denied the rationality of the form of argument. To ignore them would, however leave one bereft of knowledge and understanding which could be useful, even if it cannot be proved true. It is better to accept the assumption commonly made that it is rational to believe that the conclusions of arguments to the best explanation which satisfy the stated criteria are correct descriptions of the world.
So far it has been shown that hypothesis of very great explanatory scope and power are accepted as true. Are the other criteria, about plausibility, ad hocness and disconfirmation relevant, and if so why? The requirement that hypothesis should be as little ad hoc as possible is simply to keep to a minimum the number of statements in hypotheses which are not implied by known facts, and which are therefore more especially likely to be false. The fewer ad hoc statements a hypothesis includes, the less likely it is to be false. On the other hand, the more plausible the statements in a hypothesis are, that is the more they are implied by existing knowledge, then the more likely they are to be true.
Ad hocness and lack of plausibility are not reasons for thinking a hypothesis false, or at least they are not reasons for thinking it is more likely to be false than true. Implausibility and disconfirmation, however can provide reasons for this conclusion. For a hypothesis to be implausible, our present knowledge of the world must imply that it is probably false. And for a hypothesis to be disconfirmed is for one of its implications to be false, which means the hypothesis itself is probably false. This latter point can be amplified thus. If hypothesis (H) together with a number of auxiliary statements (A (1)… A (n) imply ( C ) and C is true then H need not be true, because another hypothesis could be the true explanation of C. But if H together with A (1)… A (n) imply C and C is false, then either H is false or A (1)… A (n) is false. In these cases the auxiliary statements are usually beyond question, in which case H is probably false. Clearly historians wanting to establish the truth of a hypothesis have to take evidence which renders it implausible or which disconfirms it, very seriously, especially if the facts explained by the hypothesis are not very numerous.
There are two more questions which must be considered before this exposition of the nature of arguments to the best explanation history is concluded. The first is about the relative weight given to conditions 2 to 6. In deciding the merits of explanatory hypotheses, are any more important than any other? The second is a worry about incommensurability. Paul Feyerabend has argued that often two explanatory theories cannot be compared because they do not account for the same set of observation statements. Is this so in history?
P 28
The conditions which bear upon the acceptability of arguments to the best explanation can be divided into three groups: those about the explanatory scope and power of a hypothesis; those about the ad hocness of the hypothesis; and those about facts which render a hypothesis implausible which disconfirm it. There is a sense in which the third group is the most important. For even if a hypothesis is of greater explanatory scope and power than another, if evidence incompatible with it cannot be explained away satisfactorily, then it is abandoned. Historians simply assume, as do most people, that the world is logically and materially consistent, so that for beliefs about it to be true, they must refer to compatible events and states of affairs. The first group of conditions, about explanatory scope and power, are the next most important, for if two hypothesis are not implausible or disconfirmed, then even if one is more ad hoc than another if it has greater explanatory scope and power it will be preferred.