|
p.2 Explanation is sometimes said to consist in "reduction to the familiar." ...But familiarity is by no means a necessary feature of explanations... Comprehensibility or understandability is the key factor in explanation, not "familiarity."
[JLJ - Yet consider: if someone is teaching another an unfamiliar skill, perhaps the teaching begins with "Look, just do this, and then do this, and then when this happens do this. You don't have to understand it now." The steps the teacher or master uses with the novice are the familiar.]
p.6 The idea of "explaining" is thus basically a dialectical one, involving a question and an answer.
p.8 Scientific explanations have their starting point in the occurrences of nature - all of nature, including man and his works.
p.9 The "scientific spirit" calls for the use in explanations of tested and confirmed generalizations that have been qualified by the evidence for acceptance as laws.
p.16 It is a convenient but unwarranted step to condemn the unfamiliar as unscientific
p.17 Every useful instrument can be misapplied.
p.23 There can never be a complete explanation of anything for the reverse of the reason that there can never be a complete description of anything.
[JLJ - I would think that practically we would never need a complete explanation of anything. At some point, whatever it is, falls into a category which we can address through some kind of scheme.]
p.24 Like description, explanation is an inherently incomplete process.
p.25-26 We shall now concern ourselves exclusively with physical systems that, at any given moment of time, exhibit some definite, specifiable state of affairs, a state that is not instantaneous but lasts for some interval of time (however short)... A system of this type, which exhibits some particular state for each of the - perhaps very short - time periods (intervals) at issue, will be termed a discrete state system, a DS-system for short.
[JLJ - Perhaps a finite discrete state system is the appropriate starting point for game theory.]
p.32 A potential prediction of the supposed fact that a system will exhibit the characteristic Q at time t is an argument whose conclusion is the statement that the system exhibits Q at t, and whose premisses consist of two types of statements: (1) general laws, L1, L2, ..., Lm, governing (i.e., known or assumed to govern) the behavior of the system, and (2) data statements to the effect that the system exhibits the characteristics C1 at t1, C2 at t2, ..., Cn at tn, where all of the times involved are anterior to t, i.e., all ti < t.
[JLJ - In a complex environment, do we absolutely need to predict in order to 'go on,' or instead (wisely, and practically) can we aim be ready for whatever emerges? A smartly constructed diagnostic test involving evolutionary experimentation with the driving forces, can help us select a current (and consequential) dynamic posture, and can show us ideally stretching to adapt to the emergent possibilities, or alternately buckling under the likely pressures.]
p.37 In general, a potentially explanatory argument can fall into one of two categories. It may be either deductive... or probabilistic... With a deductive explanation, the explanatory premisses world, if true, provide conclusive evidence for the conclusion, constituting a totally sufficient guarantee of the explanatory conclusion. With a probabilistic explanation, the explanatory premisses do not provide a guarantee of the conclusion, but merely render it relatively likely, and so endow it with a relatively substantial (conditional) probability
p.38 Prediction with DS-Systems. In general, a potentially predictive argument can also fall into the deductive or the probabilistic type. With a deductive prediction (D-prediction), the predictive premisses must, if assumed to be true, provide conclusive evidence for the predicted conclusion. With a probabilistic prediction, the predictive premisses do not provide a conclusive guarantee of the conclusion, but merely render it relatively likely, and so endow it with a relatively substantial probability.
[JLJ - And so it seems that in playing a complex game of strategy, we must make probabilistic predictions (since we cannot usually provide the conclusive evidence needed for deductive predictions) in order to establish the win-ability of our position or posture. Perhaps we must establish, though some scheme of exploring the likely, promising, unlikely and even unusual moves, the capacity of the position to evolve further in our favor in the uncertain future, perhaps by looking at how established pressures, threats and posturings typically turn out, in order to allocate our time for further trial-and-error evolutionary explorations.]
p.46 Aristotle's conception that the fundamental task of science is to concern itself with laws, with the general... and not with the idiosyncratic particularities of individual cases
p.72-73 The evidence concept... includes all of the special relationships that obtain when a body of discourse "supports" some proposition in any of the numerous appropriate senses of this term, ranging from the most demanding species of evidence which calls for establishment of a conclusion "beyond the shadow of a doubt," to the most provisional and tentative modes of argument, such as analogy or "circumstantial" evidence. The evidential relation holds whenever we must give some weight or credence to the conclusion upon some given statements (the evidence) as hypothesis.
p.73 The logical nature of the concept of evidence is a matter greatly in need of theoretical analysis and clarification.
[JLJ - "evidence" is just the support or reasoning used in an argument or dispute for jumping to a conclusion. It is still jumping to a conclusion, but it is the basis for the argument that the jump is appropriate. An argument proceeds by producing supporting evidence and casting doubt on other kinds - even the losing side of an argument can be supported by evidence of some kind.]
p.74 by adducing evidence in support of some proposition we will not, at least in general, go so far as to establish this proposition. It is only necessary for evidence to render its conclusion more tenable or more likely than before
p.76 "Probable evidence," as Bishop Joseph Butler wrote in The Analogy of Religion (1736) "in its very nature, affords but an imperfect kind of information."
p.76 In the present section, two distinct albeit related conceptions of evidence will be explicated and analyzed: confirming evidence by means of which a thesis is established, and supporting evidence which does not establish its thesis but merely renders it more tenable.
p.79 the measure of degree of confirmation as treated in the extensive literature on confirmation is based upon the mode of evidential backing at issue in the concept of evidential presumption.
p.80 By evidential presumption a statement is rendered more likely than not. Supporting evidence, on the other hand, renders a statement more likely than before
p.84 Thus confirming evidence must at once render its hypothesis more likely than before and more likely than not.
p.94 in domains in which one operates with evidence proper, leaving behind the secure ground of proof, the possibility of error must be accepted. Indeed in such fields a valid distinction may be drawn between error (drawing a conclusion, which, though false, actually derives from soundly conducted inquiry) and mistake (falsehood owing to a fallacy in the inquiry).
[JLJ - Yes, and we should operate with a wide array of mitigating factors to deal effectively with error and mistake, when they happen.]
p.94 The looser the evidence concept appropriate to an area of inquiry, the more will the system of reasoning take the form of a cluster of interlocking propositions lending mutual strength and support for one another... no system of this type presents the aspect of a collection of chains of deduction. Rather, they are akin to crossword puzzles, each piece bolstering and interlocking with every other.
[JLJ - Yes, this is perhaps how the loose heuristic 'evidence' is developed and processed in Artificial Intelligence, forming a loose supporting structure that can either stand as a 'guesstimate,' or point to further areas of deeper exploration, if time is available.]
p.104 [R. S. Walters]
To the extent that a law enables prediction about unobserved instances, it enables justifiable claims about unrealized possibilities.
[JLJ - Perhaps things that threaten to become real ought to demand our attention as well. Preparing for the real of today does not necessarily prepare you for the real of tomorrow. If you pay attention only to the observed real, you could be unprepared for unforeseen developments, or when others spring schemes upon you, or you could miss trends that others easily see and respond to. You will possibly become blindsided by inevitable change. We ought to develop a posture to the real - and what might become real in the future - in order to 'go on.']
p.110 laws are not just discovered; they are, strictly speaking, made... Lawfulness is the product of the well-founded imputation to empirical generalizations of nomic necessity and hypothetical force.
p.111 Various writers have long argued that the very idea of lawfulness is at bottom anthropomorphic. The basic idea is that lawful phenomena are rule-governed
[JLJ - Business is less so, operating instead on a business plan and perceived critical success factors, essentially the central portion of which is the scheme of the chief executive officer and his deputies - the vice presidents, to make money by generating sales, by luring the customer to voluntarily select your brand over that of the competitors, and examining the successes and failures along that route in order to keep the business plan viable in the present and future markets.]
p.116 lawfulness "lies in the eyes of the beholder," since the lawfulness of a generalization consists in its being regarded and treated and classified and used in a certain way.
p.120 Lawfulness can thus never be wholly based upon an observational foundation. Rather, it represents an imputation that is (or should be) well founded upon evidential grounds.
p.133 The consideration of stochastic systems forces us to the realization that scientific understanding can be present despite an impotence to explain (predict, etc.) even in principle certain particular occurrences... As regards prediction, it becomes apparent that it is not requisite for scientific understanding to be able concretely to foretell the future, but only that, by invoking natural laws, we be able to bring to bear a concept of possibility and impossibility in terms of which to canalize our expectations, to be able to say what can go on.
p.135 Nothing is in principle placed outside the purview of science... All the properties and states of things, any and all occurrences and events, the behavior and doings of people, in short, every facet of "what goes on in the world," can be regarded as appropriate objects of scientific explanation.
[JLJ - Yet the deliberations of the editors of scientific Journals themselves are private. The actual entities which "create" science arguably exempt themselves from outside observation. To an outsider such as myself, such Journals can take on the appearance of private clubs, publishing the thoughts of the select, and exempting the thoughts of outsiders. Science, contrary to the opinions of its practitioners, is simply the published writings of members of exclusive clubs.]
p.137 An explanatory framework is thus a family of basic concepts and principles that furnish the machinery needed for an entire range of applications in the study of some aspect of nature... Such explanatory frameworks can, moreover, exist at various levels of generality... In each case, the explanatory rationale within which the explanation of the occurrence in question proceeds is developed with reference to a certain group of fundamental processes to which explanatory efficacy is imputed
p.138 Can science actually explain everything?
[JLJ - No. It cannot explain why you, Mr. Rescher, continue to publish texts riddled with transcription errors, and do not hire an editor to better prepare your material for publication.]
p.140 Why does anything exist at all? Why is the nature of existence as it is?
[JLJ - That things exist at all is the opinion of a living being. The nature of existence is unexplained and possibly even unexplainable. Perhaps consciousness is an evolutionary-perfected mirage of a certain kind that leads or guides us to effectively perform tasks and behaviors which are related to competitions for reproduction and survival, as well as effective memberships/role playing in social groups.]
p.141 by what standards are explanatory frameworks to be evaluated?
p.142 the superiority of the scientific framework comes about not so much in point of explanation but in point of prediction and control. It is in this regard, not with respect to explanation at all, that the scientific framework is in the final analysis able to make good its claims to predominance... In the final analysis, the credentials of science derive from strictly practical rather than from purely theoretical considerations.
p.143 The choice of a theoretical framework of explanation... is not only limited, it is guided - principally by the criteria of pragmatic success in the areas of prediction and control.
appendix I: are historical explanations different?, p.147-162
p.148 in a significant sense every particular event whatsoever is unique... Events are all unique in actuality; they are to be rendered non-unique in thought only, by choosing to use them as examples of a conceptualized type or class.
p.154 Since generalizations must, in the nature of things, deal with types or classes of events, it follows that they can have pertinence to specific, particular events only insofar as these are typical and classifiable
p.161 categories, classes, and generalizations... taken together, they constitute the framework and structure of history, the setting in which the recital of particulars unfolds. They constitute the hard core of explanation and interpretation
appendix II: on the epistemology of the inexact sciences, p.163-208
p.164 It is a fiction of long standing that there are two classes of sciences, the exact and the inexact, and that the social sciences by and large are members of the second class - unless and until... they "mature" to the point where admission to the first class may be granted... For an enterprise to be classified as scientific it must have as its purpose the explanation and prediction of phenomena within its subject-matter domain and it must provide such explanation and prediction in a reasoned, and therefore intersubjective, fashion... While precise predictions are indeed to be preferred over vague ones, a discipline that provides predictions of a less precise character, but makes them correctly and in a systematic and reasoned way, must be classified as a science.
p.165 As for exactness, this qualification, far from being attributable to all of the so-called natural sciences, applies only to a small section of them, in particular to certain subfields of physics... unformalized expertise... becomes more dominant as we move away from the precise and usually highly abstract core of an exact discipline and toward its applications to the complexities of the real world.
p.165-166 architecture and medicine... must... properly be called "sciences," but they are largely inexact since in actual operation they rely extensively on informal reasoning processes.
p.166 it should be obvious that there is at present no clear-cut dichotomy between exact and inexact sciences, and, in particular, that inexactness is not an attribute of only the social sciences.
p.166-167 it is objectivity... that distinguishes science from intuitive guesswork, however brilliant.
p.169 some suggestions will be made as to how the application of relatively unorthodox methods (such as the use of experts and gaming), tried successfully in other inexact sciences, might aid the applied social sciences, especially in exercising their predictive function with regard to decision-making processes.
p.169-170 A historical law may be regarded as a well-confirmed statement concerning the actions of an organized group of men under certain restrictive conditions... Such statements share three features of epistemological importance and interest: they are lawful, spatio-temporally restricted, and loose... To consider lawfulness... Unlike a mere description, it can serve to explain developments in cases to which it makes no reference... Nevertheless, such historical generalizations are not unrestricted or universal in the manner in which the laws of the physical sciences are; they are not valid for all times and places. A historical law is limited... to applicability within specific geographic and temporal bounds.
[JLJ - Business schools typically side-step this issue of "laws" by dealing instead with "critical success factors," "case studies," and the business plan. ]
p.171 an important characteristic of historical laws lies in their being "loose." ...they can often not be spelled out fully and completely... the conditions that are operative in the formulation of a historical law may only be indicated in a general way and are not necessarily... exhaustively articulated. This characteristic of such laws is here designated as looseness.
[JLJ - Perhaps this was a reason why early computer programs written to play complex games of strategy were not very strong - the programs themselves were not able to replicate the human genius for grasping and applying loose laws that themselves were critical success factors for the successful contesting of such a complex strategic game played at a high level.]
p.172 A consequence of the looseness of historical laws is that they are not universal, but merely quasi-general, in that they admit exceptions... The laws may be taken to contain a tacit caveat of "usually" or "other things being equal" ...Such a "law" we will term a quasi-law.
p.174-175 Throughout applied physics in particular, when we move (say, in engineering applications) from the realm of idealized abstraction ("ideal" gases, "homogeneous" media) to the complexities of the real world, reliance on generalizations that are, in effect, quasi-laws becomes pronounced. (Engineering practice in general is based on "rules of thumb" to an extent undreamed of in current theories of scientific method.)
p.175 let us now turn to a closer examination of the role played by laws - or quasi-laws - in explanation and prediction.
p.182 Personal, or subjective, probability is a measure of a person's confidence in, or subjective conviction of, the truth of some hypothesis. According to Savage, it is measured behavioristically in terms of the person's betting behavior.
[JLJ - See also p.199.]
p.186-187 The consideration of such underlying regularities is of special importance for the inexact sciences, particularly (but not exclusively) the social sciences, because in this sphere we are constantly faced with situations in which statistical information matters less than knowledge of regularities in the behavior of people or in the character of institutions
p.187 quasi-laws... their use in specific circumstances presupposes the exercise of sound judgment as to their applicability to the case at hand. The informed expert, with his resources of background knowledge and his cultivated sense of the relevance and bearing of generalities in particular cases, is best able to carry out the application of quasi-laws necessary for reasoned prediction in this field.
p.191-192 The decisions that professional decision-makers - governmental administrators, company presidents, military commanders, and so on - are called on to make inevitably turn on the question of future developments... Thus a reliance on predictive ability is nowhere more overt and more pronounced than in the area of policy formation, and decision-making in general.
p.192 The use of experts for prediction does not constitute a line of demarcation between the social and the physical sciences, but rather between the exact and the inexact sciences.
p.193 For the decision-supporting uses of predictive expertise, there is in general no necessity for an anticipation of particular future occurrences. It suffices that the expert be able to sketch out adequately the general directions of future developments, to anticipate - as we have already have suggested - some of the major critical junctures ("branch points") on which the course of these developments will hinge, and to make contingency predictions with regard to the alternatives associated with them.
p.193 available evidence suggests that significant improvements are possible in the predictive instruments available to the decision-maker.
p.195 the expert's knowledge is not enough; he must be able to bring it to bear effectively on the predictive problem at hand, and this not every expert is able to do... The simplest way to score an expert's performance is in terms of "reliability": his degree of reliability is the relative frequency of cases in which, when confronted with several alternative hypotheses, he ascribed to the eventually correct alternative among them a greater personal probability than to the others.
p.197 In domains in which the flux of events is subject to gradual transitions and constant regularities... a high degree of predictive expertise is possible... In those fields, however, in which the processes of transition admit of sharp jolts and discontinuities, which can in turn be the effects of causal processes so complex and intricate as to be "chance" occurrences for all practical purposes, predictive expertise is inherently less feasible.
p.198 In the inexact sciences, particularly in the social sciences, the critical causal importance of such chance events makes predictive expertise in an absolute sense difficult and sometimes impossible, and it is this, rather than the quality of his theoretical machinery, that places the social scientist in a poor competitive position relative, say, to the astronomer.
However, when the expert is unable to make precise predictions because of the influence of chance factors, he can at very least indicate the major contingencies on which future developments will hinge. Even though the expert cannot predict the specific course of future events in an unstable country, he should be able to specify the major branch points of future contingencies, and to provide personal probabilities conditionally with respect to these.
p.199 the most important consideration is that even in subject-matter fields in which the possibility of prediction is very limited, the exercise of expertise, instead of being applied to the determination of absolute personal probabilities with respect to certain hypotheses, ought rather more profitably be concentrated on the identification of the relevant branch points and the associated problem of the relative personal probabilities for the hypothesis in question, relative, that is, to the alternatives arising at these branch points.
p.199 Even in predictively very "difficult" fields... the major branch points of future contingencies are frequently few enough for actual enumeration, and although outright prediction cannot be expected, relative predictions hinging on these principal alternative contingencies can in many instances serve the same purposes for which absolute predictions are ordinarily employed.
p.202-203 in situations concerned with complicated practical problems, no clear-cut hypothesis to which probability values could be meaningfully attached may be immediately discernible... before even a single predictive expert can be used intrinsically, some at least rudimentary theoretical framework must be constructed within which predictive hypotheses can be stated... Generally, the process involved is somewhat as follows. The situation at hand... is analyzed - that is, it is stated in terms of certain specific and, it is hoped, well-defined concepts. This step usually involves a certain amount of abstraction, in that some aspects of the situation that are judged irrelevant are deliberately omitted from the description. Then either some specific action is proposed, and a hypothesis stated as to its effects on the situation in question or, more typically, a law or quasi-law is formulated, stating that in situations of the kind at hand, actions of a certain kind will have such-and-such consequences.
[JLJ - Maybe this should also represent how we proceed in game theory, in constructing a computer program of some kind to play a complex game of strategy.]
p.203 A variant of this is of considerable epistemological significance. Instead of describing the situation directly, a model of it is constructed... in which each element making up the real situation is simulated by a mathematical or physical object, and its relevant properties and relations to other elements are mirrored by corresponding simulative properties and relations... Now, instead of formulating hypotheses and predictions directly about the real world, it is possible to do the same thing about the model. Any results obtained from an analysis of the model, to the extent that it truly simulates the real world, can then later be translated back into the corresponding statements about the latter. This injection of a model has the advantage that it admits of what may be called "pseudo-experimentation"... Pseudo-experimentation is nothing but the systematic use of the classical idea of a hypothetical experiment
p.204 a model furnishes the experts with an artificial, simulated environment. Within the environment they can jointly and simultaneously experiment, responding to the changes induced by their actions and acquiring through feedback the insights necessary to make successful predictions within the model and thus indirectly about the real world.
p.205 A particular case of simulation involving role-playing by the intrinsic experts is known as operational gaming, a special case of which is war gaming... In operational gaming, the simulated environment is particularly effective in reminding the expert, in his role as the player, to take all the potentially relevant factors into account in making his predictions; for if he does not, and chooses a tactic or strategy that overlooks an essential factor, an astute "opponent" will soon enough teach him not to make such an omission again.
[JLJ - This applies to game theory. Is not a game in part a competition, and in part a teaching exercise, where in the exertion to win the stronger player(s) will teach the weaker player(s) a lesson?]
p.207 Partly because of the absence of mathematically formalized theories, explanations throughout the area of the inexact sciences - within the physical- and the social-science settings alike - are apt to be given by means of the restricted generalizations that we have called "quasi-laws."
p.208 We have stressed the importance in the social sciences of limited generalizations (quasi-laws), which cannot necessarily be used in a simple and mechanical way, but whose very application requires the exercise of expert judgment.
|
