PLAUSIBILITY CONCEPTS
by F Richard Singer III edition date 11/11/2006
website: www.conceptualstudy.org email: richardsinger3@sbcglobal.net
Abstract: This paper is a reorganization of concepts formulated in my book A Personal Approach to Conceptual Philosophy. In Section 0, a personal plausibility attitude is initially conceptualized as a relation between a person and a proposition. The key tool for this involves the use of numerical intervals as a way to indicate compactly both a level of personal plausibility and an extent of personal uncertainty. The concept of a personal plausibility is then casually extended to also include statements that are vague to be propositions. The next two sections can be read in either order, depending on whether the reader is more interested in why plausibility concepts might be used or on how they can be developed. Section 1 suggests when and how a focus on plausibility might have advantages over a focus on truth. Two main types of reasons are discussed, namely achieving consensus and achieving clarity. Section 2 develops a thoughtful way of thinking about personal plausibility intervals by using a betting interpretation. It then develops the concepts of warranted and evidential plausibility, both for personal plausibility intervals and for impersonal plausibility intervals. Section 3 is about specific applications of plausibility concepts. It will not be written until I receive some input from readers of the other sections. I would like input of two types, namely imagined applications and implemented applications. An imagined application would suggest for a very specific use that has not been carried out. However an imaginary application could tell what happens in a fictional account of it being carried out.
Note: The concept of an attitude and the concept of negotiation that use in this paper are taken from Descriptive Psychology, a public net of theory neutral concepts and conceptual tools created by Peter Ossorio. These concepts are merely refinements of those widely used by people to think about persons and their behavior, and an ordinary comprehension of these concepts should be adequate for a basic understanding of their use in this paper. For more a detailed presentation of these concepts and more information go to the Descriptive Psychology part of my website or to the website of the Society for Descriptive Psychology sdp.org. For an introductory book on Descriptive Psychology, I recommend Persons, Behavior, and the World by Mary Shideler.
SECTION 0 THE BASIC CONCEPT
Main Perspective: Arriving at our cabin, we found a strange dog on the porch. I decided to chase it away. Charmayne asked me to be careful and remained in the car. We were both implicitly considering the proposition that this dog is dangerous, which for reference, I label as DG. I acted as if DG was only slightly plausible, while she was treating DG as moderately plausible. This indicates several feature of the main plausibility concept to be formulated. First I want to stress that by a proposition I mean a statement that proposes information that is clear enough for the purposes at hand. Altho plausibility could be conceptualized as an attribute of a proposition, as somehow related to the extent to which it is reasonable to believe that it is true, there are reasons why this is not the main plausibility concept that would serve the purposes I have in mind. Foremost is the fact that I need a concept that relates to how a person would act in a situation in which the proposition is relevant. In the case of DG, I got a stick before trying to chase the dog from the porch. I would not have taken this precaution if I had considered DG totally implausible. Had I considered DG as somewhat plausible, I would have been even more careful. Another reason I want a relational concept is that plausibility involves personal considerations that can be influenced by a variety of factors other than evidence. A dog once attacked Charmayne. While this is evidence that a strange dog might attack, our reactions to this evidence are naturally different. My estimate of danger relates to my personal interactions with dogs that seem aggressive, but who back down when boldly confronted. The essence of the concept I call a personal plausibility attitude is given below.
A personal plausibility attitude is a relation between a person P and a proposition C.
I am using the Descriptive Psychology concept of an attitude. An attitude is a behavior tendency that is context-specific, in that it is directed towards some focus of attention. Different plausibility attitudes of P towards some proposition C indicate differences in the types of action P might take. A plausibility attitude of P towards C will influence behavior only in situations where C is somewhat relevant. P may not be aware of these attitudes and may not even be thinking in terms of plausibility. Unless we experience some hesitation, we usually regard propositions that we rely on as if they were true.
Example 0: This example illustrates how a plausibility attitude can easily shift, especially when almost nothing is at stake. Let C be the proposition that the Dow closed over 8500 while I was out working today. Shortly I will listen to the nightly business report and will find out whether C is true or false. However before then I modified my plausibility attitude towards C several times as indicated below. This example is also a prelude to the main conceptual tool for describing plausibility attitudes, namely by using a pair of numbers to represent such an attitude in a manner to be explained shortly.
A0: I find C slightly to be somewhat plausible, but having paid little attention recently, I feel uncertain about what might be happening in the stock market. I do recall that the last time I looked the Dow was about 8100, and that there had been some small gains in the last few days. [20,40]
A1: My son calls and asks if I knew that the Dow had gained over 300 points by noon today. Immediately, C seems more plausible, perhaps even fairly plausible. [45,75]
A2: When I ask him about the starting level he said 8274. Suspecting that the 300 point rally probably did not hold I now feel extremely uncertain about C. [10,99]
A3: He then said that by mid-afternoon it had given back half the early gain but then there was a rally. From the tone of his voice I was fairly certain that the Dow had finished with this rally. At this point C seemed highly plausible and my level of certainty increased considerably. [90,95]
A4: He tells me the closing average was 8534. C now seems totally plausible to me. [99,100]
Plausibility Intervals: The number pairs above are called plausibility intervals for my plausibility attitude. They indicate the extent to which I considered C plausible at various points in time.
A plausibility interval for P’s plausibility attitude towards a proposition C is an ordered pair of whole numbers [l,n], where 0 £ l < n £ 100. Thinking in terms of a scale from 0 to 100, this means that P would say that C is at least as plausible as l but not as plausible as n.
The Use of Numbers: My house is at 7739 Ahern. The next house is not at 7740 Ahern. Unlike probabilities, to which we can apply the usual operation on real numbers, house numbers indicate ordering rather than measuring. The use of numbers for plausibility is more like their use for addresses than their use for probabilities. Since thinking in terms of percentage is common, the range from 0 to 100 provides a useful heuristic tool for articulating a variety of considerations about plausibility concepts. I restricted the values to whole numbers to avoid any connotation that plausibility attitudes are measurable. Above I used numbers in an off hand manner associated with the phrases below. For instance, suppose P considers C plausible enough to act on but not highly plausible. P might give an interval whose center is some number from 71 to 89, using 80 as center unless P was inclined to do otherwise. If P was completely uncertain about C, P might give the interval [0,100].
|
Barely Plausible: [1,10] |
Slightly Plausible: [10,30] |
Somewhat Plausible: [30,50] |
|
Moderately Plausible: [50,70] |
Fairly Plausible: [70,90] |
Highly Plausible: [90,100] |
The number pairs represent my intuitive feeling about plausibility. The [20,40] after A0 is a short way of saying that I find C at worst slightly plausible and at best somewhat plausible. I could have used the pair [11,49] for this had I wanted to also indicate even more uncertainty. The pair [25,45] would also indicate that I find C at worst slightly plausible and at best somewhat plausible, but that I was leaning more towards somewhat plausible. The [45,75] after A1 show that I have changed my feeling, and that I then found C at worst only somewhat plausible and at best fairly plausible, and that I was leaning towards moderately plausible. Note that the extra information not only shifted me in that direction it also increased my uncertainty. After the next information, I had almost no feeling about whether the Dow had closed over 8500, except that I would not have said that it was barely plausible. The information about the loss and the rally increased my attitude at stage A3 to highly plausible, which I represented as [90,95]. I read the [99,100] at A4 as totally plausible, meaning plausible beyond any doubts that I might actually have, altho of course I could imagine that for some strange reason my son was giving me incorrect information.
Plausibility and Vague Statements: The above concept of a personal plausibility is a relation between a person and a proposition. With modifications the concept of a plausibility attitude towards statements that are too vague to be propositions could also be considered. Let C be the statement that burning leaves is bad for the environment. To the extent to which C does not propose clear information, even with unlimited knowledge I could not determine if C is true. My problem is that I find the idea of being bad for the environment too vague in this context. On the other hand, I find that statement that a nuclear war would be bad for the environment clear enough to be a proposition. Altho a vague statement is conceptually incapable of being true or false, it is still at least possible to act as if it was in some cases. I consider the concept of a plausibility attitude towards C in that manner. Assuming that P has given no thought to the possibility that C could be vague, then action possibilities in relation to C abound. Attitudes toward C could influence such actions. Such attitudes would also effect the discussions about C. With a wide-open attitude P might burn leaves but also look for others ways to get rid of them. P might implicitly interpret C as applying in the city but not in the country. However attitudes other than plausibility ones could be involved. If P is indifferent to the small impact of his own action then P might act as if C was highly plausible in an argument, but still burn leaves.
Implausibility: Sometimes we want to focus attention on why not to act on a proposition C, in which case we might call it highly implausible instead of barely plausible. Likewise, but with less of a feeling about the implausibility of C we might call it fairly implausible instead of slightly plausible.
Precision: As explained above the use of numerical intervals is merely a way to compactly indicate both a level of plausibility and an extent of uncertainty. These can be used in an offhand manner with little or no regard for precision. If so used, it should not be surprising if a person having forgotten the interval he used to indicate an attitude A towards a proposition might later uses a different interval to indicate A even if A has not changed. For many purposes there is little reason to take a plausibility intervals as anything more that a casual heuristic tool.
To be more thoughtful about using a plausibility interval we can interpret them as if they represented an attitude toward betting. Using this interpretation, which will be developed in Section 2, my thoughtful [90,95] attitude given above relates to several betting considerations. I would comfortably risk $90 to win $10 on a wager that C was true. However I would definitely not risk $95 to win $5 on such a wager. Furthermore I would be at least somewhat uncomfortable about risking $91to win $9, and would be unlikely to risk $94 to win $6. On the other hand I would be very happy to risk $80 to win $20, or $70 to win $30, $60 to win $40, etc.
SECTION 1 RATIONAL FOR UTILIZING PLAUSIBILITY CONCEPTS
Plausibility and Truth: While attempts have been made to formulate a logic with more than two truth-values, or even a logic that could not be described in terms of truth values, none of these have been widely accepted as useful. Altho the ordinary concept of truth does not seem to be limited to its use in formal logic, it still seems to be used mostly as if a proposition was either true or false. Altho I use a more complex synthetic subconcept of truth, for a simple proposition, truth can be thought of as an attribute of the proposition, altho in cases an attribute that may be somewhat vague. The main ordinary maxim for using this subconcept in relation to simple propositions is the following.
A simple proposition is true IFF it accurately describes some aspect of something.
On the other hand, a personal plausibility attitude is a relation between a proposition C and a person P. It relates to truth primarily because it indicates P’s degree of willingness to act for some purpose as if C were true. Instead of the stark two valued true or false expectations, personal plausibility attitudes involve a wide variety of ratings. This is particularly helpful when dealing with remote or unknown factors. Consider the claim that it will rain tomorrow. We do not know enough atmospheric science to be able to give a definite true or false to that statement, however we do have some relevant knowledge and information. Forecasters even state the chance of rain as a percentage. I suspect that these are not probabilities in the purely mathematical sense. They might be better interpreted as impersonal evidential plausibility attitudes rather than as probabilities. Whatever probabilities might be involved in weather forecasting, there are a multitude of factors involved that make it unlikely that any well defined mathematical sample space is a central feature of the models on which they are based.
Plausibility and Legal Justice: One realm in which the utility of thinking about plausibility is recognized is in legal matters. Even the locutions focus on this, altho in our adversary system law enforcement official and prosecutors act as if the defendant is guilty and the defense may try to convince the jury that the defendant is innocent. Still in a criminal case what is called for is guilty beyond a reasonable doubt, and no demand is made that a jury must determine the truth. In jury deliberations, altho discussion may sound as if truth was to be determined, the actual goal is a consensus on plausibility of rather than certainty. If a juror follows this guideline then all that the juror needs in order to vote not guilty is a warranted attitude of at least slightly plausible towards the defendant’s plea. On the other hand if all jurors have warranted attitudes of at most barely plausible then a verdict of guilty should be rendered.
Likewise plausibility attitudes are even more apparent in civil procedures. These are to be decided on the weight of evidence, and this clearly need not be a matter of certainty. In one of the accounts that I examined, weight of evidence was described as a 50% consideration. Interpreted as plausibility, this would mean deciding for the plaintiff if a juror found the plaintiff’s claim at least moderately plausible and the defendant’s claim at most somewhat plausible.
Main Advantages of Using Plausibility: There are at least two types of reasons that focusing on plausibility attitudes at times might be more useful than on focusing on the truth of a statement. One is to achieve a greater consensus, or to see why this may be difficult. The gap between true and false is psychologically greater than the gap between highly plausible and barely plausible. More important thinking in terms of plausibility at least suggests that a somewhat greater consensus might be feasible. Another reason for focusing on plausibility attitudes is to bring a statement into better focus. This can be especially useful in a discussion when the same words suggest different concepts to the persons involved. Accounting for a plausibility attitude can be useful in bringing out these differences. Bring a statement into better focus can have behavioral utility even when no disagreement is involved. Reflecting on plausibility could modify the way a person actions might have been influenced by some statement not previously recognized as vague.
Consensus and clarity reasons are often interrelated. Increasing clarity can show that differences were less significant than might have been imagined. On the other hand it bringing statements into sharper focus may make it apparent that the differences are even more significant than might have been imagined. Even if clarity initially intensifies differences it can be a prelude to useful negotiations. Without clarity, negotiation is likely to degenerate in ways that mask even the consensus that might have been achieved.
Consensus: In moving in the direction of consensus, thinking about personal plausibility attitudes towards some proposition C could have a number of advantages over focusing on whether C is true or false. Merely thinking about plausibility attitudes sets the stage for consensus. On the other hand, to move from false to true involve a major change, especially when the proposition being debated is about something that cannot be easily verified or in which beliefs are firmly entrenched. The very recognition that one has a plausibility attitude allows for movement. Even with an extreme attitude such as [99,100], a person merely needs to give some credence to something not previously considered in order to move to [98,100]. This can also lead to considering more factors and even willingness to look for more relevant information that might help in moving towards consensus. Of course this will not happen if a person is set on defending some cherished position. Example D below is an artificial illustration in which considerable consensus is achieved. Section 3 is intended for more useful illustrations.
Example 1A: Jan is a close friend of Ed. They talk on the phone daily and see each other at least weekly. They know a lot about each other’s habit and interests. Joe and Bob also know Ed, but not as well as Jan knows him. Joe has a high respect for Jan’s opinions, but Bob does not. Sue has only seen Ed at a few meetings, but she knows that Jan and Ed are close friends. Sunday evening the four of them decide that they will stop by Ed’s before a Tuesday morning meeting. Knowing that Ed will offer them breakfast, the following proposition C comes up for discussion.
C: There are at least 5 eggs in Ed’s refrigerator.
Each of these persons finds C at least slightly plausible, but due to differences in personal information and attitudes they have different plausibility attitudes. Sue having almost no opinion can imagine C as anywhere from barely plausible to highly plausible. Because Jan knows Ed so well, she has a more definite attitude. Each person agrees to writes an initial plausibility attitude towards C prior to any discussion or any attempt to acquire more information.
Sue: [10,90] Joe: [40,60] Bob: [20,40] Jan: [80,99]
After considering each other’s initial attitudes, but without discussing the matter, they each decide to modify their own attitudes.
Sue: [40,99] Joe: [60,80] Bob: [10,30] Jan: [70,99]
Further discussion reveals the following information, resulting in additional revisions.
¨ Jan and Joe and Bob all agree that Ed is not allergic to eggs.
¨ Jan and Joe agree that Ed likes to eat eggs.
¨ Bob claims that Ed baked cookies Sunday, using at least 6 eggs.
Sue: [0,100] Joe: [60,80] Bob: [10,50] Jan: [50,99]
Ed’s son comes in and tells them the information below, which they all consider reliable.
¨ Ed shops once a week and bought 2 dozen eggs Friday.
¨ Ed did not eat any eggs Saturday, but I did.
¨ Bob was right about Ed baking some cookies Sunday afternoon.
Sue: [80,99] Joe: [80,99] Bob: [70,99] Jan: [90,99]
Some of the reasons for the shifting attitudes in this example should be apparent. Merely seeing that others had different attitudes encouraged some shift towards consensus, at least when respect was involved. The initial shifts are largely due to information about each other’s attitudes. Sue’s and Joe shift because of Jan’s higher plausibility attitude, and this also narrows Sue’s uncertainty. Bob’s shift is due to his lack of respect for Jan. Jan simply become more uncertain because of the attitudes of the others. Maybe they know something she does not know. Discussion that allows for sharing reasons brings out information that did lead to a slightly greater consensus. Further shifts are due mostly to information from their discussion, which tends to lead to greater uncertainty. Finally they obtained some additional information, which resulted in a high level of consensus. This information from Ed’s son tends to shift them all towards fairly-to-highly plausible. It also narrowed the uncertainty. Of course none of these steps is guaranteed to lead towards consensus, but they at least have a potential to do so, and are likely to in for a highly manifest claim. In this particular case, just looking in Ed’s refrigerator would produce a total consensus, but there are highly manifest claims that cannot be checked this directly.
Consensual Factual Negotiation: Negotiation is a way of arriving at a resolution of some concern that is satisfactory to the parties involved. The concern can be about what is the case or about what to do. The adjective ‘factual’ can be used when the first of these is the primary concern. A negotiation will be classified as consensual if the parties involved want to resolve matters in manner that considers the perspective of all concerned as much as can be reasonably expected and if the performance of the parties focuses on authentically looking for a way to do so.
Descriptive Psychology represents a version of consensual factual negotiation process as a social practice that can be broken into 4 stages: taking positions, criticizing and defending these positions, adjusting these positions if necessary, drawing conclusions. There is no essential reason that a consensual factual negotiation should focus only on truth claims. In some cases more progress might be made if personal plausibility attitudes were examined. Instead of taking fixed positions, some extra effort could be given to specifying personal plausibility for these positions. Instead of merely criticizing and defending positions, an explanation of why these attitudes seem personally warranted could be given, including reasons that the parties have for treating there own positions as less than totally plausible. Making revisions would involve indicating even minor shifts in plausibility attitudes. Finally, drawing conclusions would focus both on how much progress was made towards on consensus and on the degree of remaining difference. This stage should also focus on what to do about remaining difference, at least if a greater consensus seems desirable.
Achieving Clarity: People can act as if a claim is true without even considering that it might be vague. One advantage of focusing on plausibility is that by focusing on uncertainty it sets a stage for thinking about reasons that the claim at issue may be vague. On the other hand, focusing on truth tends to presuppose that the claim is clear. If the claim is about manifest matters, feedback can bring clarity. If the vagueness is due to problems with a shared version of a claim then discussing plausibility attitudes can help bring clarity. Focusing on your own personal plausibility attitude can be a way of achieving clarity about your own use of a statement, even if you do so without any discussion with anyone else. Of course focusing on plausibility attitudes is not the only tool that can be used to help achieve clarity.
Example 1B: Ray says that picking up a water snake is crazy and I disagree. I pick one up. It clamps down on the skin between my thumb and index finger. It draws a trivial amount of blood, but causes no pain. This initiates a clarification of our claims that helps us see that both of our plausibility attitudes are personally warranted. He says that altho the snake is not poison, its bite could cause an infection. If this is all that Ray meant then I find his moderately plausible. His statement may have been somewhat vague to each of us when we started, but feedback shows that the main lack of clarity was not in our individual versions of but with our shared version. I tried to demonstrate that the statement was implausible by picking up the snake. A discussion of plausibility attitudes could have provided a less dramatic way to show that C proposed different types of information us.
Influencing Behavior: Altho the purpose of making a statement may be to propose information, statements can also be used to influence behavior. I am told that it is almost six o’clock and so I come in from the garden. In this case the statement is clear and compact. The influence on my behavior is useful to me, and it would be redundant to add an imperative sentence telling me to come in. This is the case with a wide variety of statements about ordinary manners in which information is proposed in order to influence behavior. With such statements considering plausibility is largely irrelevant. However some statements are motivated much more by a desire to influence behavior than by a desire to propose information. An advertisement may give information, but this is not its primary purpose. Furthermore they may influence behavior in ways that would not be useful to the actor. Clarification of any information such a statement might be proposing can help counter a tendency to be influenced by it. Again focusing on plausibility attitudes is one way to clarify the information.
A statement does not even have to propose information in order to influence behavior. It only needs to seem as if it does. Thus vague statements can influence behavior, and as with advertising this may be their main purpose. When a statement is about remote matters such as values or social policies, it may often be more about influencing behavior than about providing information. Since many statements about manifest situations are clear enough for the purposes at hand, we have a tendency to also treat unclear statements about remote matters as if they were propositions. Cultivating a habit of thinking about their plausibility is one way to counter this tendency, especially if you consider whether your attitude is warranted. Doing so is more likely to provide the flexibility needed to bring a statement into better focus than merely than marshalling reasons for defending a fixed stance on the statement.
For a vague statement C about remote matters, the only way some people may react to C is by asserting it and arguing about it. The desire to maintain or refute C may interfere with even considering that C might be vague. I know that I make many statements about remote matters without even realizing that they are vague to me. Exploring differences in plausibility attitudes can help me see why my statement was vague. This could help me clarify it, at least to myself. Likewise people often make statements about remote matters that seem vague to me, and I usually suspect that they are vague even to the people making them. Given a willingness to acknowledge vague thinking, considering plausibility attitudes could be a useful in obtaining greater clarity. Even when it is not, I find it useful to spot that what is being said is vague to me. This tends to be the case with a multitude of claims about social policies and values. Unless a person is willing to search for clarity and to focus on plausibility, I have little interest in discussing such statements.
Example 1B: Let C be the statement that the war on drugs is a disaster. The libertarian party would like us to find this so plausible that would be inclined to vote libertarian. Of course how a person acts on a proposition also depends on a variety of attitudes. An alternative way to act as if C was plausible would be to vote for candidates who want to put more effort the interdiction of the drug traffic. Using the word disaster instead of failure is one way to counter that tendency. The concept of a disaster may be vaguer than the concept of a failure in this context, but it suggests that a radical change is needed. When challenging an entrenched policy, a vague but powerful statement can rally the opponents and possibly instill doubts in some who have passively accepted the policy. Reflecting on my plausibility attitude towards C, I see that I have treated C as true altho C is not clear enough to me to be a proposition. I would not deliberately treat a vague statement as if it was true, nor even act as if it was totally plausible. Spotting that a statement is vague, there are several options; try to clarify it, simply refrain from acting on it in any way, deliberately leave it vague but also adopt a plausibility attitude towards it. This last option is what I chose to do with C, acting as if my plausibility attitude towards C was [70, 100]. I do this because C represents a complex of propositions that I have not formulated, but which I would expect to find in the fairly plausible or highly plausible ranges. Altho I could formulate this complex, it would be tedious. Since the only way I act on C is to use it as a summary of implicit propositions, as long as the vagueness of C in focus, this seems like an appropriate way of dealing with it.
Total Plausibility and Truth: Saying that C is totally plausible means that C is plausible beyond sensible doubt, i.e. only by some flight of fantasy could we currently imagine otherwise. For instance I find the claim that I am using a word processor as I edit this paper totally plausible, altho I can imagine the word processor has been disabled and that what appears on the screen does because some strange virus has infected my computer. The concept of totally implausible is treated in a similar manner. An attitude of totally plausible towards C is both conceptually and pragmatically different from believing that C is true. For one thing, to explicitly acknowledge such an attitude is to remain at least somewhat open to changing it. However regarding C as true or as totally plausible usually has the same influence on action, and it is often more convenient to think of C as being true. This is especially the case for our basic reliable knowledge, as well as a multitude of thing we take for granted in our daily activity. I get up in the morning with some plans for what I might do. All of these presuppose thing I do not explicitly consider. I act as if it there is hot water in our water heater, as if there is a spoon in the kitchen, as if the roof is on our cabin, etc. Later most of these are verified to the extent that they need to be. Of course, some things I regarded as true turn out to be false. This is usually at most a minor nuisance. I was sure there was a can of peas in storage, but as there was not, we had spinach instead. On reflection I would say that most of these things that I took for granted were merely things that I would have considered totally plausible, had it occurred to me to think this way. In such ordinary matters I easily accept mistakes and see no reason not to think primarily in terms of true or false. I merely use the very ordinary criteria that in many matters thinking of what I take for granted as being true is not only convenient but also justified by the competence I have in evaluating such matters. It is when I want to reflect more deeply on my understanding of epistemic concepts that I focus on the fact that I much of what I take as true I might instead regard as totally plausible.
Side Remark: I often use [99,100] to indicate totally plausible. Since an attitude relates to action, it is more appropriate to say that C is totally plausible means I would not hesitate to act as if C was true in any situation where C was relevant. If [99,100] seems weak for the concept of totally plausible, consider betting. I would risk 99 to win 1, but not 100 to win 0. This is certainly the case with anything I consider totally plausible. Had I allowed rational numbers for interval endpoints, I could have used a narrower interval for the concept of totally plausible. However I do not want to suggest they plausibility attitudes are actually measurable.
Epistemic Utility: Achieving greater consensus and achieving more clarity are fairly manifest ordinary reasons for focusing on plausibility attitudes. I also have a more remote epistemic reason, namely that it helps me reflect on my epistemic concepts. One aspect of understanding epistemic concepts is to realize that even when reasonably challenged, I may act as if my claim was true. As a mundane example, I was sure that there was a can of peas in storage, but Charmayne says she cannot find any. I still claim that we have some. This is partially justified by the fact that I organized our storage areas and that I remember seeing them there, as well as by the fact that I often find items that she overlooks. However given what I know at the time, I should at best regard my claim as totally plausible. Plausibility easily becomes relevant as I search in vain, but is quickly forgotten as I finally decide my claim was false. Reflecting on plausibility helps me realize the utility of cultivating the habit of giving more tentative responses when ordinary claims are challenged. My ideal is to have only a prudential interest in being right, not only about such trivialities, but about any claims I make. Cultivating a more tentative attitude towards propositions about manifest situations carries over to those about more remote ones. In addition I frequently remind myself that evidence for remote propositions tends to be complex. Thus in claiming something like “last years tax cut had effect X.”, I would find discussions involving plausibility more useful than arguments that try to convince. It seems to be difficult for most people to regard cherished positions on complex issue in terms of plausibility, and for that reason I find myself slipping back into a truth oriented thinking when such matters or being discussed. It is only by being vigilant that I can take what I consider a more appropriate plausibility perspective.
SECTION 2 DEVELOPING PLAUSIBILITY CONCEPTS
The Betting Interpretation: Betting involves a willingness to take a chance at certain odds, and thus involves an attitude. For a simple case consider some proposition C and suppose P would consider making a reasonable size bet that C will turn out to be true. Odds are usually stated as ratio, such as 3:2. To accept a bet with 3:2 odds is risk an amount 2x to gain an amount 3x. Likewise to take a bet at 2:3 odds means to risk 3x to win 2x. For easy comparison, think of 3:2 as 60:40, 4:1 as 80:20, 9:11 as 45:55, i.e. as ratios whose parts sum to 100. This way, each can be coded as a single risk number: 40, 20, 55. The second part of the ratio is used because it indicates the amount to be risked in order to gain the other part. Thus a higher number indicates a greater confidence in the proposition being considered. A plausibility attitude is represented as an interval rather than a single number because one important aspect of such an attitude is the extent to which it is wide or narrow. To say that my plausibility attitude is represented by [20,55] indicates that I would gladly bet on C if given odds of 80:20 odds or better but would definitely refuse to bet given odds of 45:55 or worse. This betting type of interpretation can be taken primarily as an analogy, but there are situations in which it seems reasonable to identify a plausibility attitude with a betting attitude. The use of numbers is merely a convenience for indicating differences in attitudes. They are intended to indicate rather than measure attitudes, and for many purposes I see no need further qualify any distinction between a betting attitude and a plausibility attitude. However qualifications can always be imagined. In the next example I have implicitly assumed that the reason P would not bet at 4:1odds is related to his attitude about the plausibility of his team winning. If he needed $80000 to pay a loan shark and only had $20000, then his actual betting attitude could differ from his plausibility attitude.
Example 2A Let C be the proposition that the Celtics will win the NBA championship next year. Suppose Sal would not even consider betting on C if given odds of 4:1, would not hesitate to bet on C given 9:1 odds, and would be undecided for odds between 9:1 and 4:1. Since these odds are 90:10 and 80:20, his attitude towards betting on C can then be represented by the interval [10,20]. With certain qualifications, this interval also represents his plausibility attitude towards C. End points indicate personal comfort levels at a particular point in time. It may seem unlikely that Sal would really jump at a chance to bet at 90:10 odds but hesitate at 89:11 odds. Perhaps, but his reasons may be fairly arbitrary. Sal may merely find 90:10 easier to understand because it converts to 9:1.
Example 2B: Let C be the proposition that in the next hard rain the top mantle over my fireplace will remain dry. Had I made this prediction a month ago my plausibility attitude would have been something like [1,50]. In terms of betting this means I would gladly risk 1 to win 99, i.e. bet at 99:1 odds. However I would definitely not bet at even odds. Yesterday we found cracks in the roof tar around the chimney and patched them. Now my attitude towards C is in the fairly-to-highly plausible range say [80,95].
Present Plausibility: Altho the concept of a plausibility attitude interval sometimes applies to something that might happen in the future, it also involves relevant things that could reasonably be known about the present. In the Example 2B the plausibility attitudes focused on what would happen if it rained. I could just as easily substitute my attitude toward the proposition about the roof tar, which is actually more relevant to my concerns. Having observed water on the mantle was what motivated me to inspect the tar, and filling the cracks was intended to prevent the leaking. My [80,95] attitude was more about having fixed the cause of a leak than about the future effect of a rain. I focused on rain primarily because this was why I was worried about the leak. In general, as long as chance is not a significant factor, most plausibility attitudes about the future would be reducible to plausibility attitudes about the present. It is the fact that we act on what we tend to believe about the present and the past that motivated me to formulate plausibility concepts. This applies even to betting. Altho betting usually relates to future events, the analogy to betting does not depend on temporal considerations. Given 9:1 odds, I would gladly bet that a coin already fairly flipped but not yet examine is tails up.
Example 2C: Let C be the proposition that there is life on Mars. My initial reaction is that C is barely plausible, perhaps expressing this as plausibility [0,5]. As a betting attitude, suppose that the truth-value of C could be determined in the near future. Then I would bet on C being true at 100:0 odds but I would not bet at 19:1 odds. Since a plausibility attitude is a relation between a person and a proposition occurring at a particular time, this may change. Given a new probe finding fairly strong evidence of life on Mars I might change my attitude to highly plausible, say [90,95]. Which attitude I take might affect the strategy I would use in advocating a manned mission to Mars. The above statement about life on Mars was not vague for me. I implicitly knew that I was thinking of life as including microscopic life. Instead of barely plausible, Marge said totally implausible because she was thinking of life as a complex ecosystem involving higher order plants and animals. In explaining my plausibility attitude this difference became apparent. We easily saw that our attitudes only appeared to differ because we were using this statement for different propositions. It is fairly easy to indicate the propositions associated with statements about life on Mars. The difference in the concepts suggested by the words is minor and the kind of information that might be relevant is fairly clear-cut. With statements for which this is not the case, it could take more effort to see that at least part of the difference in attitudes is due to the difference in concepts.
Utility of the Betting Interpretation: When chasing the dog from the porch I had given no thought to any plausibility attitudes involved. In fact I did not even formulate the proposition DG until several days after the incident. Since plausibility attitude involve a tendency to act, rather than to reflect on what-to-do, this is typical of the way we usually treat most of our attitudes. I act on many plausibility attitudes, but I formulate few of them and consider plausibility interval only rarely. Otherwise little would get done. Formulating DG and reflecting on how to think about my plausibility attitude towards DG in terms of betting can have no effect on what I did. Nor do I intend it to help me decide how to act the next time a dog appears on our porch. Why then do I find the concept of a plausibility attitude useful, and why do I find it useful to relate it to betting attitudes? My answer may seem simplistic. Reflecting on my attitudes helps me understand my self and my world, and as a result it increases my options. Perhaps reflecting on DG may have only some influence on my plausibility attitudes towards dog attacks, but this is not its purpose. Its main impact has been on my overall understanding of how I might think more clearly about plausibility attitudes and how this way of thinking could have some utility.
It is the betting interpretation that provides the clarity I need. Using it in several examples has made me more sensitive the conjectural nature of the propositions that implicitly influence what I do and say, altho not yet to the extent that I would like this to happen. This interpretation is also a tool I can use to communicate about my plausibility attitudes, at least to anyone willing to take the effort to understand them. I hope it is a tool that other could use to better understand and communicate their attitudes towards various propositions. Altho this tool is not directly useful for attitudes that are not towards propositions, most attitudes relate to some propositions that could be formulated. Plausibility attitudes towards these propositions could provide insights into other attitudes. Since the effort in formulating statements with sufficient clarity to be propositions can be considerable, this may not seem worth the effort if the only purpose is to understand a specific attitude. Since I enjoy the effort, this is not deterrence. Furthermore my purpose in formulating such propositions is always broader than this. Understanding a specific attitude influences my understanding of my epistemic concepts.
Whether person would actually bet could depend on factors other than odds. So without some qualifications a person’s attitude towards betting on a proposition may differ from that person’s plausibility attitude towards it. The discussion of qualifications below is motivated by a desire for greater clarity than may be called for by considerations of direct utility. It may seem somewhat tedious and artificial, so skip it if it seems uninteresting.
Qualifications: A person’s attitude towards betting often depends on the amount of the bet rather than just on the odds. For instance at Stage A2 in Example 0, I gave my plausibility attitude as [10,99]. As a betting attitude, this means I would definitely bet at odds of 90:10 and would definitely not bet at odds of 1:99. However I qualified this by indicating a reasonable wager. This is because my attitude toward betting also depends on the amount of the wager. I certainly would not have bet $100000 at even better odds. Even a small risk of loosing this outweighs prospects of becoming a millionaire. Likewise I would not actually wager $1 because the winning would be too small to compensate for placing the bet, and in addition to this I have a bias against petty gambling that has little to do with the odds. In order to make the betting interpretation work for me, I interpret my betting attitude in terms an imaginary world in which I can effortlessly make imaginary bets. With a playful goal of increasing my financial status in this world, my plausibility attitudes are likely to coincide with my betting attitudes.
Another qualification about the relationship between betting and plausibility attitudes can be illustrated by looking back at the proposition DG mentioned earlier. Being at the low end of slightly plausible, I could represent my attitude by [10,20]. As a betting attitude this means I would risk 10 to win 90 but would not risk 20 to win 80. On reflection, this seems about right, altho precision is hardly called for. However my plausibility attitude differs in a major way from the way I would act in the betting interpretation. The relation of risk to payoff was not very clear. Furthermore not acting on C did not mean that I could not get the payoff. I could hedge my bet by choosing a better one. In getting a large stick I decided to bet instead on the proposition that the dog could not successfully attack me if had this stick. This was a proposition that I high considered highly plausibly [95,99], so obtaining the stick seemed worth the extra effort. The effort was minimal and the apparent risk was considerably reduced.
Plausibility vs. Probability: In a game such as roulette the odds have a fairly direct relationship to the likely hood of what might actually happen, altho people playing roulette accept poorer odds than the probabilities involved. People also bet when there is no way to even calculate probabilities. Suppose we pretend that the odds in Example b are probabilities. Jumping at 9:1 odds is to act as if the probability of the team winning was better than 1 out of 10 or 10%. Being unwilling to bet on 4:1 odds, is to act as if the probability of their winning is certainly not as great as 1out of 5 or 20%. To be unsure of 17:3 odds is to be unsure about a probability of 3 out of 20 or 15%. In many betting games there is a clear sample space from which to calculate the odds for a fair game. However in Example b I have no idea of what sample space to use, so acting as mathematical probabilities were involved is merely a heuristic device. Thus Sal’s attitude towards this bet does not represent the probability of the Celtics winning the championship. Instead it at most represents the plausibility that he associates with the proposition C. Unlike a situation for which probability can be used to assign a definite value to uncertainty, plausibility usually involves a more amorphous uncertainty. To account for this, I use an interval for his plausibility attitude towards C, namely [10%,20%] or more briefly [10,20]. Unfortunately, persons often talk in terms of probability when plausibility would be more appropriate. For instance if the treasury secretary is asked about the probability of an economic recovery by the end of the year, it seems unlikely the answer will involve a conceptual claim about some sample space. I strongly suspect the answer to be more indicative of a plausibility attitude. This is not to denigrate the intent of the question or his response. His plausibility attitude should be of the type that I will later classify as evidential, and some of his actions may be influenced by this attitude.
I want to stress that the even when related to betting attitudes, the concept of a plausibility attitude in no way depends on the mathematical concept of probability. Betting involves a degree of willingness to take a chance, which may be related to probability, but which often occurs when there is no clear way to figure probabilities. Thus the concept of betting does not depend on the concept of probability. The mathematical concept of probability was not even formulated until the 16th century. Clearly people were making bets long before this. Furthermore many bets are still being placed with little or no regard to mathematical probabilities. In fact the work of Pascal and Fermat in probability was at least partially motivated by a question posed by a gambler about the division of stakes in a game of dice.
Warranted Plausibility Attitudes: Since an attitude relates to choosing actions, I am not interested in merely identifying my willingness to act. I also want my plausibility attitudes to be warranted enough for some purpose. I represented my attitude towards DG as [10,20]. Since the dog on our porch turned out to be extremely docile, I think my attitude towards DG was warranted enough to act as I did. Likewise [11, 25] and a variety of other intervals would have been just as warranted. Had it turned out otherwise, such attitudes might have seemed unwarranted. However outcomes alone are not sufficient to determine whether or not an attitude is warranted. DG turning out to be false once tested does not suggest that a plausibility attitude of [0,1] would have been warranted. Given what I could reasonably be expected to know before confronting the dog, only a much more open attitude would have been warranted. Furthermore altho this particular dog turned out to be docile, perhaps my attitude was a matter of false confidence and not being harmed was just a matter of luck on my part. I claim my attitude was warranted because of my competence, i.e. because of what I know about dogs and my ability judge this dog’s reactions when we arrived. I have several ways of thinking about a plausibility attitude being warranted. All of them are somewhat vague, and there is no sharp line dividing warranted and unwarranted attitudes.
A personal plausibility attitude A towards C is warranted for P to extent that it involves an adequate understanding and interpretation of the information that P could reasonably be expected to consider about whether C is true or will turn out to be true.
The ‘reasonably be expected to consider’ qualification is one of several reasons that the concept of being warranted can be vague, especially when taken out of context. However given the context it can be easy to apply, and this qualification is essential to the concept I want. Altho P might be able to use bribery to obtain insider information about the prospects of a stock, it is easy to imagine that this would not be something P could be reasonably expected to consider. A warranted attitude for P would differ from a warranted attitude for a person willing to break the law. In general, P can only be expected to consider information that P can reasonably be expected to obtain, given P’s characteristics and the state of affairs involved. Because different people could reasonably know different things and could also have different analytic and synthetic abilities, widely different attitudes toward a proposition could be warranted for different people. Furthermore what a single person could reasonably know is not static and plausibility attitudes are not precise. There would often be more than one warranted attitude for a person towards a proposition and this could change as things evolve.
Consider the initial attitudes in Example 1A from Section 1. With the information given and the condition about expressing their initial attitudes, it is possible that all of these attitudes were warranted. However an attitude of [1,99] or of [20,80] would also have been warranted for Sue. As they obtain more information none of their initial attitudes would be warranted. However unless Bob has good reasons not to trust Jan’s judgement that are relevant to the issue, his [10,30] attitude seems less warranted to me. Without more information about these persons it would be impossible to say to what extent his final attitude is warranted, but I could easily imagine that it might be warranted.
Pragmatically Warranted: The most manifest concept of warranted that I consider involves being warranted in relation to some specific purpose or set of purposes. I refer to this as being pragmatically warranted, and this means that for the purposes at hand, obtaining additional relevant knowledge would involve doing something that would not have positive utility for P. Altho my attitudes in Example 0 shifted rapidly as I passively gained information, I regard each of them as pragmatically warranted in relation to mild curiosity. Altho I could have interrupted to ask about the closing, this would not have served a useful purpose. Had I a more significant purpose, only A4 would have been warranted, since I could reasonably be expected to ask immediately about the closing if anything major was at stake.
Example 2D: While walking, Charmayne and I were discussing plausibility concepts when we heard something in the woods some distance behind us. Turning we got a glimpse of a tail and movement of foliage. We both concluded that we were fairly certain that it was not a squirrel. Other than our fleeting visual and auditory impressions, and what we know about the size of squirrels, there was little else that we could have known that would have been relevant. Yet what we did know was enough to say that our attitudes of barely plausible (or highly implausible) to it being a squirrel were warranted. We also noted that an attitude of totally implausible to it being as an elephant would be warranted. We had different attitude towards it being a deer, but we agreed that our perceptual differences warranted different attitudes. My first reaction was fairly implausible, but after a little discussion I decided that a much more open attitude might be more warranted, especially after considering what else it might have been. Again only mild curiosity was involved in our conjectures, and so these attitudes were at least pragmatically warranted. Altho we might have obtained more information by going back to where we saw the movement, given only mild curiosity, whatever we might have found that would have been relevant is not something we could reasonably be expected to consider. In fact given our limited abilities in regard to animal signs, what we might have observed was unlikely to be anything we could reasonably expected to consider. For an experienced tracker P, an attitude similar to ours might have been pragmatically warranted in relation to mild curiosity, but it might not be more generally warranted.
Narrow Attitudes: A variety of attitudes towards DG could have been warranted enough for purposes at hand, but not one as narrow or as low as [0,1]. Such an attitude would have been warranted if it had been our neighbor’s dog, but extreme attitudes such as [0,1] and [99,100] are warranted only when know enough that plausibility would not normally be considered. Other narrow plausibility attitudes can be warranted under fairly special circumstances. For example if a fair coin has been flipped and not examined, then an attitude of [49,51] toward it being heads is certainly warranted for many purposes.
Impersonally Warranted: So far I have only considered plausibility intervals as being personally warranted. I also want to be able to say that even if P’s plausibility attitude was warranted for P, it was not really warranted. Perhaps P was not competent to evaluate the proposition. This suggests a concept of a plausibility interval for a proposition being impersonally warranted.
A plausibility interval for C is impersonally warranted to the extent that the persons who are qualified enough to evaluate the plausibility of C would consider this interval as a warranted plausibility attitude towards C, if they considered what they could reasonably know about C.
Since this is a consensus concept involving a degree of competence, there are some caveats involved in using it. First note that just as various compatible intervals for C can be personally warranted, various compatible intervals for C can be impersonally warranted. However, it may be that no interval is impersonally warranted for C, and even if there are impersonally warranted intervals for C they may be difficult or even impossible to determine. For a more precise concept we would need to specify the degree of consensus that is demanded.
Example 2E: Consider again the proposition DG about the dog not attacking. Is the interval [10,20] impersonally warranted? Since Charmayne and I where the only people to observe the dog at the relevant time, let us suppose that we were the only people who could have been qualified to evaluate the plausibility of DG. Charmayne and I agreed that I was better qualified. However we might be wrong about this. Perhaps she was qualified enough for her attitude to be considered, in which case [10,20] would not be impersonally warranted. Perhaps neither of us was qualified, in which case no plausibility interval would be impersonally warranted. In saying this, I supposed being qualified as being actually qualified. There are clearly many others who were potentially qualified, i.e. would have been qualified if they had the opportunity to observe the dog. Including them might mean that there was one or more impersonally warranted attitude, altho there might be no way to determine what these were. While both Charmayne and I agreed that potentially qualified persons might have found [10,20] as warranted, this is not something either of us knew how to verify.
Example 2F: Let G be the claim that the earth revolves around the sun. Altho dispute about G did not focus on plausibility, suppose Galileo’s plausibility attitude towards G was [95,100] after he observed the moons of Jupiter. This seems not only warranted for Galileo, but also a warranted impersonal plausibility attitude. Many critics of Galileo had an adequate understanding and interpretation of the information they could reasonably know about G, but lacked the background to understand the principles on which a telescope is based. For such critics a personal plausibility attitude of [10,90] might have been warranted, but since they lacked the competence to evaluate the plausibility of G, their attitudes are not relevant to whether an impersonal plausibility interval was warranted. After the work done by Galileo the only warranted plausibility intervals for G were those in the range of highly plausible. The extent to which a plausibility interval is warranted can vary with time. For the last few centuries the only warranted plausibility attitude for G has been [99,100], so much so that we usually merely think of G as true. I do not have enough information to consider what impersonal plausibility intervals would have been warranted before his telescopic observations and his studies of motion, but perhaps something as open as [60,100] might have been impersonally warranted at that time. Prior to being proposed, no attitude towards G is conceptually possible, and the statement that the earth revolves around the sun would not even propose the same information as G in a time prior to Greek astronomy.
Evidential Plausibility Attitudes: One factor that provides a reason to consider whether an attitude is warranted is the extent that it based on relevant evidence that can be articulated and examined. This suggests the concept of an evidential attitude.
A personal
plausibility attitude towards C is evidential for P to the extent that it is
based on the
information relevant to C that P could reasonably know and articulate and
examine.
I am fairly confident that my attitude towards DG was pragmatically evidential. Of course there are things I could have reasonably done to make it more evidential. Looking for studies about dog attacks was not one of these things. However I could have made a more careful observation of this dog, but implicitly believing that my attitude was pragmatically warranted, this did not occur to me. Altho being evidential and being warranted are related there is one major difference. Being evidential is not a sufficient condition for being warranted. Being warranted also involves competence in adequately applying the evidence. Nor is being evidential a conceptually necessary condition for being warranted. As conceptualized, a warranted attitude cannot ignore evidence that can be examined, but this leaves open the extent to which it might involve other factors. I do not believe that any of my warranted attitudes involve mystical powers, but I cannot rule this out by purely conceptual consideration.
Being evidential also applies to impersonal plausibility intervals, and it is one of the factors providing a reason to consider whether a plausibility interval is impersonally warranted is the extent to which it is impersonally evidential. A plausibility interval for a proposition C is impersonally evidential to the extent that it would be an acceptable evidential plausibility attitude for most persons who would be competent to evaluate the relevant evidence.
Objective Plausibility: In thinking about whether or not a plausibility interval for a proposition C is impersonally evidential or impersonally warranted, I am implicitly thinking about the plausibility of C rather than about a personal attitude towards C. Thus when I say that C is highly plausible, and I say this as if this is an attribute of C rather than an expression of my attitude, I am tempted to say that what I mean by this is that an interval in this range is impersonally warranted. However I am reluctant to conceptualize it this way. The conceptualization of being warranted is at least partially intersubjective. I have yet to decide if it would be useful to try to formulate a concept of plausibility that is a totally objective attribute of a proposition. Nor do I know how I would do so. Currently I use a vague concept of this type, as if a plausibility attitude is some how objectively grounded in what could reasonably known but does in any way depend on the considerations of qualified persons.
A heuristic version of a correct objective plausibility is easy to give, altho I have doubts about its utility. Imagine an infinitely competent person objectively evaluating what could be reasonably known at the time and deciding the extent of the impersonal plausibility of C. Also pretend that this person has a much more precise concept of how to assign intervals to attitudes than any concept that we have been able to imagine. The plausibility interval this person would assign to C could be conceptualized as the correct objective one.
Not having infinite competence, there may be no way to determine a correct objective plausibility of C at a time T. However I am still willing say that the claim G about the earth and sun is currently totally plausible in an objective sense. I am also willing to say that [49,51] is a correct objective plausibility interval for the fair coin proposition mentioned earlier. That I could be wrong about such matters is no more important to me than that I could be wrong about a multitude of other matters that I have no reason to doubt. Furthermore not knowing a correct objective plausibility interval for a claim does not prevent me from being fairly certain about some intervals that do not qualify. Whatever the objective plausibility of G was when Galileo proposed it, it not in the range of barely plausible. In spite of these considerations I have little commitment to the utility of the concept objective plausibility. However when I use this concept, the extent to which an impersonal plausibility interval for C is warranted is at least one factor in deciding the extent to that C is objectively plausible.