|
Introduction, p.1-10
p.1 This is not a how-to book about prediction... Instead, this book will look at the big picture. It seeks to explore the general lay of the predictive land by considering the fundamental principles of the predictive domain at large and examining the theoretical issues that arise within and about it... with a view to the prospect of developing, at least in outline, a general theory of prediction
[JLJ - Ok Nicholas, we are going to develop a general theory of prediction, without actually going into the details of how to predict. I get it. No, actually I don't. Rescher, however, later discusses predictive methodologies, p.86.]
p.2 We need to know about the future because we are always being pitched headlong into it... The predictive domain is clearly a minefield where much caution is required.
[JLJ - Alternatively, the future can be thought of as a concept which helps us figure out how to "go on." Disagree? Ok, well then do this for me. Specifically point to where the future is. I'm waiting. If you can't do that, to this or any other concept, then perhaps it is in your mind.]
p.3 A creature of our sort must make its way in a difficult world by the use of intelligence... Information regarding future developments is for us humans not just a matter of curiosity but one of survival... We are by nature forward looking and our view of the future shapes our view of the present and our activities within it... In many human decisions... people's present actions depend crucially on their beliefs about the future, and thus on the sort of predictions they are inclined to accept.
[JLJ - Rescher misses the importance of my concept of the scheme to go on. A belief is kind of pointless if we do use it to scheme how to proceed in our present predicament, and in the predicaments that follow.]
p.4 The future matters for us because it both takes away and gives. It unravels the present condition of things, replacing it with something else - be it good or bad.
[JLJ - Is the future a "thing" that "takes away and gives"? The future is more likely conceptualized as a name we give to a concept so that we can think about it in the present. The present moment soon turns to yesterday, being replaced by a modified version of the present that we are to some extent ready for, and to some extent not. It is our wise investments which give, our misfortunes which take away. For practical purposes, the future exists in the mind as a guiding factor for present actions, but we can think about "it" and prepare ourselves for "it", and "it" is the seed of the present. The me of yesterday (or last month or last year) decided that the me of today (for them - the future "me") would probably not object too strongly to the condition of things, or else "he" (they?) would have done something about it. To those who die today, the future is only a dream that comes to an end.]
p.5 Almost always, we gear our present actions and reactions to future expectations... Prediction is essential for us to exist and to function in this world as the sort of creature we are
p.6 Predicting the future to a greater or lesser extent is an inescapable part of our human destiny.
[JLJ - It is wise to concern ourselves with prediction, in order to 'go on' in our current predicament.]
p.6 We live in the present, but always do so in the face of an anticipation of the future... only the future will decide on the signification and significance of various present actions.
[JLJ - Alternatively, the future becomes important to us because of the present predicament we are in, and the driving forces which threaten us at every turn if/when we ignore the real and imagined future and its potential consequences. Consider the fictional Peter Pan and his friends living in Neverland, where they do not grow up. The future is kind of suspended, and so the present becomes nothing more than a 'neverending' game of adventure and intrigue. There is no managed growth of any kind. We non-Neverlanders, on the other hand, consider the future because it is wise to do so, and because the consequences of not making investments which pay off tomorrow in denominations of potential for adaptation - are large.]
p.7 The appropriate characterization of the thing we presently do is generally something that can only be determined on the basis of future outcomes... To characterize something in objective terms... is also to make a claim about its behavior in the future... Wherever we turn in the domain of human knowledge and human affairs, issues regarding the future loom up before us.
p.7 the book endeavors to give a rounded picture of the predictive project and to present a considered appraisal of its problems and prospects. Its main aim is to clarify the general processes and ideas at work so as to provide at least the rudiments of a theory of prediction, as it were - a synoptic overview of the basic principles at issue here.
[JLJ - ...which also might be useful for game theory.]
The Philosophical Anthropology of Forecasting, Chapter 1, p.11-18
p.11 The future is, for us, an object both of curiosity and of intense practical concern, and prediction is our only access to it.
[JLJ - Do we really need to predict "the" future, or can we instead base our present actions on future-leaning metrics which guide us in forming or assuming a posture or stance towards it? The company I work for has taken a loan out with a bank. The loan calls for the calculation of a common metric known as EBDITA - Earnings Before Depreciation, Interest, Taxes, and Amortization, and should this metric indicate that the cash flow of the company has started to suffer, the loan must be repaid in full at once. The bank cannot see into the future - it does not need to - it simply calculates a performance metric which strongly hints at good or poor future performance, and an indirect ability to service debt. That performance hint is practically good enough to set a reasonable threshold for a tripwire for the loan principal to be repaid. Flash forward to 2019 - the bank has now foreclosed on the loan in question, taken over my employer, and sold us off in order to repay the loan. All based on metrics which predict the future.]
p.11 Our lives are lived in a world whose eventuations all too often lie outside the range of our predictive foresight
[JLJ - ...so? With adaptive capacity and wise investments and multiple paths available to us, the 'future' can be tamed and wrestled to 'behave itself.']
p.12 Knowledge is power. And the fact that virtually all action is in some way future oriented endows our predictive knowledge with special practical potency.
[JLJ - Knowledge is power when used strategically and intelligently to guide action or posturing, but perhaps only in the sense that in the end it is useful only to the extent that it can help select among 'tricks that work' to find the one or ones that work better than the others. 'Knowledge' needs to be effectively grasped by an agent in a predicament, when scheming how to 'go on,' or else it loses its ability to 'empower.']
p.16 Foreseeing the shape of the future demands an insight into how things actually work in the world
p.17 while prediction is an endeavor of critical importance for us, it is one in which we can expect at best a very limited success... The most critical fact in the theory of prediction is accordingly that while there is much that we can successfully foresee, our powers in this direction, through real, are unavoidably and substantially limited.
[JLJ - Perhaps the future is best summarized and generalized as a premonition which guides action and posturing, and which guides intelligent scheming and planning efforts, in order to 'go on' within our current predicament.]
Historical Stagesetting, Chapter 2, p.19-36
p.22 Only within the sphere of established sciences phenomena are reliably predictable (to whatever extent science permits), but elsewhere they are not, because where science cannot operate, chance and contingency prevail to establish an uncertainty that frustrates prediction.
[JLJ - Yet 'tricks that work' in human nature and in competitive events allow actions and postures to be taken that produce predictable results, on average, in the long term. A team of con-men and women, like the kind I personally saw in 2014 working together on London bridge, do not know exactly who will fall into their trap and allow themselves to be swindled out of their money, but in the long run their *trick* works - or they would not be there doing it. We might not know *the* future, exactly, but that does not mean that we cannot prepare for it, make wise investments in the present, monitor and make wise adjustments, and eventually profit from this activity - from what emerges down the road. There is no 'science' to business - there are merely 'tricks that work,' which are put into practice by intelligent, enthusiastic, financially-resourced and scheming people, for example, as documented in publications like Harvard Business Review. There is no 'science' to warfare - one studies the practical concepts involved in putting an army in the field and in solving the problems that arise in strategically outmaneuvering an opponent out to do the same.]
p.25 Prediction is no big problem for those for whom the world is fundamentally stable
p.27 Increasing mastery over nature has at last rendered the future knowable because our scientifically guided capacity to shape and control the unfolding course of events makes us humans masters of our destiny to a greater degree than ever before.
p.27 In Henrik Ibsen's Hedda Gabler we find the following exchange: "
LOVBORG: This new effort is my real book - the one I have put my heart and soul into. ...It deals with the future.
TESMAN: With the future! But, good heavens! We know nothing of the future!
LOVBORG: No. But there is a thing or two to be said about it all the same."
[JLJ - We might know nothing of the future, yet that does not prevent us from crafting intelligent schemes to be ready for it, and for the opportunities it presents. A spider spins a web and waits patiently, comfortable with a 'trick that works' for supplying future meals.]
p.31 To what extent is the future... knowable at all? How desirable is it that people should obtain insight into their future?
[JLJ - To what extent do we practically need to know the exact future that will emerge as the outcome of the complex driving forces of our predicament? Perhaps we strategically prepare for a range of futures, and act accordingly, relying on our adaptive capacity when things do not go as planned. The question should be recast as, 'How much should we obsess over the future, when our inputs can only be premonitions, and much will remain unknown and even unknowable? How should we balance the constant demands on our time, so that we can at times, revel in the present, and manage our rage, rest from rage, and recovery?']
p.33 In a world pervaded by chance and chaos, predictive insight into those aspects of the future that interest us the most has proved to be far more problematic and difficult to come by than those enthusiastic post-war futurologists ever envisioned.
[JLJ - Yet look at any modern business seeking to manage itself in the current predicament it is in. It must look into the future to see the emerging trends and position itself for what will arrive. Customers must still come knocking on the door, seeking services and goods, or else the business must be fundamentally restructured. The future is never solely a curiosity or a trend - those who can manage the emerging trends as well as the things that will remain constant have a good chance to continue on as a business - those who miss out will struggle and perhaps fail.]
Conceptual Preliminaries, Chapter 3, p.37-52
p.37-38 To predict is, more or less by definition, to endeavor to provide warranted answers to detailed substantive questions about the world's future developments
[JLJ - Not so fast - prediction might also be a part of a scheme to prepare for an event. In our scheme - which calls for a prediction - we ask ourselves over and over: 'what might happen?' then, 'other than that, what else might happen?' The net effect is that we emerge from this exercise prepared for whatever happens or occurs in the interaction of networked objects, of our predicament. We did not predict exactly - we merely adopted an adjustable initial stance, followed by continuously modified stances, that allowed us to pivot with events as they unfolded. Prediction can be a part of a scheme or business plan rather than a "thing" unto itself.]
p.38 Any really appropriate prediction must address the question at issue... Any prediction worth bothering about must rest on some evidential basis... serious cognitive interest attaches not to predictions as such but rational predictions - those that are credible in the sense that there is good reason to accept them as correct then and there, before the fact.
p.39 for a claim to constitute an actual prediction there must be someone who makes or accepts it, who takes responsibility for endorsing its correctness.
[JLJ - Yet in the world of artificial intelligence, such a prediction is made without such a someone. Worse, such predictions can be relied on in ways that jeopardize human lives. The Boeing 737 Max Angle of Attack sensor failed, and the software reading and acting on the flawed sensor output caused the plane to act as if it was approaching a stall condition. Essentially no one person - a computer acting blindly - caused a plane to crash and people died. No one predicted - essentially - that an error in this sensor prediction could jeopardize human lives. We are a half step away from robots on the battlefield being programmed to kill, with no one ultimately being responsible if the killing is in error. Who speaks for these people - victims of a time, a place, a technology, and a prediction gone wrong?]
p.39 Prediction, in sum, is our instrument for resolving our meaningful questions about the future, or at least of endeavoring to resolve them in a rationally cogent manner.
p.40 Prediction deals in indicating what the future will be... To be sure, there is nothing wrong with scenario construction as such; it is unquestionably an interesting and instructive venture. However, the fact remains that scenarios are a matter of surveying possible courses of future developments. They are imaginative speculations about what might happen and not informative speculations attempting to preindicate what will happen. By their very nature, then, prediction and scenario construction are different sorts of enterprises.
[JLJ - Practical prediction deals in indicating what the future might be, in order to decide how to 'go on,' or to form a posture in the present. Prediction and scenario construction can be combined - for example, we can use scenario construction to predict how ready we are for the uncertain events of the future, then adopt a stance or posture which makes us the most prepared for whatever happens.]
p.41 Predictions can be either categorical or conditional. Categorical predictions have the form "E will happen," or "E will not happen," where E is some particular definite occurrence or outcome. Conditional predictions, by contrast, have the form "E will happen if F does"
p.41 Conditional prediction comes in two main types: specific (when-next) and general (whenever).
p.42 Prediction is a question-answering process that involves a commitment regarding what will or will not happen - be it categorically or under certain circumstances. Attaching a high probability to a future event is not to answer the predictive question of its occurrence but rather to characterize the status of such a prediction... To make a prediction is to move beyond anything that probability as such can do for us.
[JLJ - To make a useful prediction you must use a scheme of some kind that hinges on a profound understanding of the driving forces, the agents involved, and how interaction over time can unfold. I don't know Nicholas, about the question-answering part - I have played sports games of various kinds as a child and teenager where I had to predict the actions of an opponent, and I did so intuitively, without asking questions. When multiple things can happen in a sports game, you worry about being out of position, or where the action might be, and you act accordingly. The predictive scheme might call for asking the questions, 'how might things proceed?' and 'other than that, how might things proceed?' but the answers come from our experience and intuitive understanding of events. You can ask questions eternally but never predict anything. It is only when a practical scheme of informed action is used intelligently that prediction as such emerges from guesswork and speculative imagination. What matters is not the prediction itself, but the posture we take as a result of the prediction. The prediction as such will always remain a premonition that contributes to how we intelligently decide to 'go on.']
p.45 In matters of prediction, large-scale issues must be addressed by large-scale processes.
All the same, there is much - indeed virtually everything - to be said for adopting the practical policy of letting probability be our guide in matters of prediction.
[JLJ - Should a prediction take into account the fact that someone might take actions based on the prediction - actions that might affect the prediction itself?]
p.45 Predictions, properly speaking, must deal with contingent matters regarding the future.
p.47 Proper predictions must as such have a substantive bearing upon the contingent issues of the future. Where this is absent we will have, at most, pseudopredictions that only masquerade as the real thing.
Basic Epistemic Issues, Chapter 4, p.53-68
p.60 As we have seen, a prediction must, in order to be useful, be both informative and credible.
p.62 Clearly a great deal can be predicted if one is vague enough... the fact is that with any sort of estimate - prediction included - there is always a characteristic tradeoff relationship between the evidential security of prediction on the one hand (as determinable on the basis of its probability or degree of acceptability), and its contentual informativeness (definiteness, exactness, detail, precision, etc.) on the other.
[JLJ - Yes, including: any future book written by Nicholas Rescher will contain errors which could have been prevented by the hiring of a competent editor. How is that for a vague prediction?]
Some Ontological Issues, Chapter 5, p.69-84
p.76-77 For us humans the future is veiled, as it were, in a cloud of unknowing. Through our predictive efforts we peer into it much as we peer into a fog. Very little can be seen at a distance - and that little with but little clarity. But as things draw nearer, the fog of uncertainty dissipates somewhat and we can begin to make out some of the features of things with greater detail. And so it is with the future too.
[JLJ - Yet consider, what kind of life, or living, could be had if we could predict everything? The future must remain a premonition which guides actions in the present, combined with the wisdom of the past, and allows us to decide exactly when and how we revel, rest and rage. End of story.]
p.79 What sorts of conditions must obtain for rational prediction to be feasible?
[JLJ - The prediction is likely useless if it is not part of a tested and experienced scheme, itself composed of various 'tricks that work', and the triggers which 'decide' when to apply them. We execute a predictive scheme, and what emerges is and remains a premonition.]
p.80 To what extent, then, can we predict success for the predictive venture in this world of ours? What can we reasonably expect in this regard?
p.81 Prediction, in sum, is a cognitive venture whose successful pursuit is inseparably bound up with factual matters regarding the nature of the world's modus operandi. Only when this is duly benign - only if nature's detectable patterns can underwrite adequately informative estimates regarding the future - will prediction be feasible. And this is something we can never guarantee in advance on the basis of abstract general principles.
[JLJ - Yes, but do we truly need to predict perfectly? I would venture that we merely need to be ready for whatever does emerge. I get on I-66 every morning and I have no exact prediction for what I will see. But I am ready for it, whatever it is.]
Predictive Methods, Chapter 6, p.85-112
p.86 As indicated in the preceding chapter, it is only where the future is somehow foreshadowed in the discernible stability patterns of the past-&-present that rational prediction becomes possible. For all of the rational processes for validating predictions are in the final analysis based upon pattern fitting, since all of them proceed by conforming an envisioned future to the structure of the available data.
p.86 Any sort of rational prediction... will accordingly require informative input material that indicates that three conditions are satisfied:
- that relevant information about the past-&-present can be obtained in an adequately timely, accurate, and reliable way,
- that this body of data exhibits discernible patterns, and
- that the patterns so exhibited are stable, so that this structural feature manifests a consistency that also continues into the future.
We thus have the conditions of data availability, of pattern discernability, and of pattern stability as three indispensable preconditions for rational prediction.
p.86 Reasoned prediction... rests upon an insight into order and calls for detecting stable patterns in nature's modus operandi. All modes of rational prediction call for scanning the data at hand in order to seek out established temporal patterns, and then set about projecting such patterns into the future in the most efficient way possible.
p.87 Rational prediction pivots on the existence of some sort of appropriate linkage that connects our predictive claims with the input data that provide for their justification. This linkage can be based either on explicitly articulated principles... or on personal judgments that exploit a knowledgeable expert's tacit, unarticulated, and sometimes unarticulable background knowledge about the matter.
p.87 what is pivotable for the reliability of predictions is that a justifactory mechanism... have a good "track record" - a suitably favorable history of providing for predictive success.
p.88 A. N. Whitehead wisely remarked, "Foresight is the product of insight."
[JLJ - ...and hindsight.]
p.90 In 1943 Thomas Watson, the shrewd founder of IBM said, "I think there is a world market for about five computers."
p.92 To be sure, the accuracy of such estimates varies with the particular parameter at issue: economists perform far better with production and employment levels than with interest rates.
p.102 No predictive methodology - however carefully contrived it may be - can ever immunize forecasting against the prospect of failure.
[JLJ - ...yet we can play with such unfalsifiable predictions as "tomorrow always comes," because we can never be in a position to prove it wrong.]
p.102-103 One of the most common forecasting methods consists in employing predictive indicators that are based on an empirical finding that two (usually quantitative) factors are closely correlated in such a way that the behavior of the one foreshadows the behavior of the other... Predictive indicators presuppose the operation of causal linkages.
[JLJ - Perhaps this is a core concept of artificial intelligence. We intuitively understand the linkage and react when we see it. However, what if an opponent knows we will react accordingly, and prepares a trap of some kind?]
p.103 Predictive indicators, like predictions in general, can thus be either explanatorily rationalized or unexplained.
p.105 Predictive black boxes thus constitute a somewhat questionable resource... But there are times when we have to settle for the imperfect best that we can get
[JLJ - ...such as a piece of computer software 'playing' a complex game of strategy on a computer via the automated execution of code. A move is produced that is to be played in the game, without any justification or explanation of the internal process(es) involved. Is it in fact 'the imperfect best that we can get'?]
p.106 Every sort of predictive resource can and should be tested... only if it fares well in this regard does a predictive source deserve credence.
p.110 sophisticated mathematical models... predictive successes... two obstacles can arise:
- an insufficiency of data to support the operation of a workable model;
- analytical complications arising because the real-world phenomenology at issue is too complex for our inferential tools.
[JLJ - All the more reason to use 'tricks that work' in prediction, even in the presence of missing data or analytical complications. Prediction offers insight in deciding how to 'go on,' in the present predicament, with a clock ticking and things changing in front of us. If we predict wrong we recover from our resilient posture and capacity for such things. Rarely is prediction fatal when wrong - except of course in predicting whether a high voltage AC line is connected or not. Sure, predict if the line is connected, but ground yourself and test the line with a voltmeter to be sure. Why predict, unless we somehow use the results in deciding how to 'go on.']
p.110 a model is a potentially powerful instrument of prediction - though one that is also particularly liable to the limitations and malfunctions that beset all human contrivances.
p.111-112 Only by trial and error can one eventually assess in the hard school of actual experience the efficacy of a particular predictrive method with respect to a particular sort of predictive issue.
p.112 Predictive efficacy as such is something that can itself only be established by means of predictions.
The Evaluation of Predictions and Predictors, Chapter 7, p.113-132
p.113 The evaluation of predictions involves both assessing the merit of predictive questions, and assessing the merit of predictive answers.
p.113-114 the mission of the predictive venture is to obtain high-quality answers to high-quality questions... The importance of a predictive issue is evaluated in terms of how much is lost by not having a correct answer to it - the difference that this absence makes in the larger scheme of things... importance can be either theoretical or practical: the absence of an answer to a particular question can matter both for our practical affairs and for our theoretical understanding of things.
[JLJ - Isn't it a bit naive to think that our predictive model will outperform all other predictive models? We all walk around with a predictive model of sorts - it helps us to position ourself in the present and to make certain necessary investments in the future that provide a resourceful payoff. A predictive model is no more and no less that an intelligent attempt to position oneself in the present, in order to 'go on', and is as good or as bad as time, money, testing, and creative modification and complexity permit or allow it to be. To predict what year we will establish a moonbase, or will have the first woman president, will help few position themselves in the present, and is more of a curiosity than a critical need in order to determine how to 'go on'. Ideally, a prediction should be part of a strategy to position ourselves in the present, so that we can 'go on' wisely, productively and resourcefully, while still allowing time to 'revel' in the present. Show me someone who is looking to predict, and I will show you someone who is deciding how to position him/herself in the present, with an eye for an easier future-present which will soon arrive. When we predict, are we not in fact jumping to unproven conclusions from what are only intelligently-constructed forward looking metrics, premonitions, which themselves are grasped and elevated to a place of importance, because we have to grasp at something?]
p.115 The resolvability of a predictive question - its definitive verifiability or falsifiability in due course - poses significant problems because predictive questions often defy all prospects of practicable settlement.
p.118 The difficulty of a resolvable predictive question is a function of the amount of effort that will have to be expended on actually producing a resolution for it - of coming up with an appropriate answer. It is clear that this is going to depend, at least in part, on the existing state of the art... And sometimes nothing that we can actually do will provide sufficient. Predictive performance is going to be a function of questions: pretty much any predictor can perform impressively if one asks easy questions.
[JLJ - In playing a complex game of strategy, we might ask ourselves, "How might I proceed?" and "Given that, how might my opponent now proceed?". By examining possible lines of play, we can effectively (and indirectly) construct a diagnostic test of a proposed stance, aiming for one which is robust - that is, one that has the adaptive capacity to mobilize coercion. Perhaps the question we should be answering is, "What moves in the game are deserving of my attention, knowing what I know now, of course, and which moves can I wisely and realistically postpone investigation until later, or at least reduce the time spent?]
p.119 With prediction, everything pivots on the questions that we ask.
[JLJ - I would say that with prediction everything depends on the intelligently constructed predictive scheme that we adopt, and an implied practical sensitivity to critically important concepts or metrics.]
p.120-121 even in situations where we cannot confidently predict the specific course of developments, it may be feasible to delineate a limited number of alternatives that map out the range of plausible possibilities... A reliable prediction of the weaker sort may well satisfy our planning needs
p.123 To achieve credibility, a prediction must accordingly have a plausible grounding of some sort that is able to render it discernibly probable in advance of the fact.
p.125 the cardinal principle of the theory of prediction is that it is not just predictions we want, and not even just correct predictions, but predictions of high quality in the relation to the entire spectrum of relevant criteria.
p.127-128 What are the factors at issue in assessing predictive competence?...
- Versatility or Range as determined by the extent of the topical and thematic range over which the predictor is able to function effectively.
- Daring as determined by the ability to tackle (and to succeed with) difficult issues.
- Perceptiveness in terms of the detail and definiteness of its predictions.
- Foresight as determined by the temporal reach - the span of the future over which the predictor is able to function effectively.
- Consistency as determined by the uniformity of the predictor's performance over time.
- Self-criticism as determined by the accuracy of a predictor's self-appraisal - manifesting superior performance where the predictor indicates greater confidence and/or claims greater competency.
- Knowledgeability as determined by the predictor's cognitive competence with respect to nonpredictive issues in the sphere of its predictive domain.
- Coherence as determined by the compatibility and systematic harmony of its predictions.
p.131 Yet what if a credible (sensible, responsible) prediction misfires and turns out to be wrong? All one can say is "too bad!" That is simply the sort of world we live in - one in which even the best made predictions of mice and people can go awry.
[JLJ - Yet we all possess an incredible amount of adaptive capacity, which we use with our other cognitive resources, in order to "go on." If our prediction does not work, we shift positions, adapt accordingly, and go on. What other choice do we have? Rarely do we do anything without a fallback position, or even several.]
p.131 We have no real alternative but to expect that some of even our best predictions will go wrong. Indeed we can safely predict it.
Obstacles to Predictive Foreknowledge, Chapter 8, p.133-158
p.134 the project of rational prediction can be frustrated in various ways. Specifically, the principal impediments to predictability are:
- Anarchy: Literal lawlessness - the absence of lawful regularities to serve as linking mechanisms
- Volatility: The absence of nomic [JLJ - customary, ordinary] stability and thus of cognitively manageable laws
- Uncertainty: Law ignorance - the lack of nomic information due to an epistemic failures on our part in securing information about the operative linking mechanisms
- Haphazard: The lawful linking mechanisms (such as they are) do not permit the secure inference of particular conclusions but leave outcomes undetermined...
- Fuzziness: Data indetermination...
- Myopia: Data ignorance...
- Inferential capacity: the infeasibility of carrying out the needed reasoning... even when we have the requisite data and know the operative inferential linkages
p.135 There is no question that the present is replete with predictive portents of the future... But whenever one cannot... determine the present condition of things, then equally, one cannot manage to foretell the future state of things that will evolve from them... Where we lack adequate information about the specific processes at work - where the causal nature of the situation remains a mystery to us - rational prediction thereby becomes infeasible.
p.135 We cannot foresee what we cannot conceive.
[JLJ - This does not prevent us from developing and maintaining a general-purpose adaptive capacity, and an ability to improvise, which can be used to re-position ourselves once surprised. Its not so much predicting the future that we seek, but rather to be ready for whatever it is that emerges in the future-present.]
p.136 the most promising way of handling the sort of impredictability that is grounded in ignorance rather than chance is to pursue further inquiries to eliminate or reduce the uncertainty at issue.
[JLJ - I would agree, but the type of uncertainty that I would attempt to reduce - in matters of strategic conflict - would be specifically the uncertainty of my ability to re-organize (under specific or surprising circumstances), as intelligently explored through practical diagnostic tests. I do not care about the exact future which arrives - merely my ability to successfully adapt to it - when it comes.]
p.136 The situation where there are no operative laws... is clearly fatal for prediction... Next to outright anarchy there stands volatility and instability of patterns, which also functions as a major obstacle to prediction. As we have seen, rational predictability requires stable transtemporal coordination relationships - discernible patterns of lawful regularity that link that past-&-present condition of the phenomena at issue, where alone the observational determination is possible, with their future state.
[JLJ - Who would want to predict exactly in these types of conditions?]
p.139-140 in chess the players decide within the framework of rules, and this means that a great deal can be predicted... But not, of course, everything. The contingency introduced through the decisions of the players also comes into it.
p.141 the net effect of chance and uncertainty is the same either way as far as predictability goes - namely, incapacitation... despite this theoretical difference, there is often a cognitive (rather than optical) illusion at work to make ignorance look like chance.
[JLJ - Yet there are always practical rules of thumb, quick and dirty approximations, which often work, and are readily grasped at when in a predicament and there is the constant need to determine how to 'go on.' We as humans must manufacture how to go on out of information correct or incorrect, perceptions real or illusions false. We have plans but also fallback plans, and adaptive capacity, and continuously modify our plans as things emerge.]
p.141 when the laws governing the phenomena are too complex for ready apprehension, then it is too easy to ascribe eventuations to chance and thereby attribute to nature the effects of our own cognitive deficiencies.
[JLJ - I would say that it is not necessarily the "laws governing the phenomena" that are "too complex for ready apprehension," but rather instead that it is the complex interaction of the driving forces that cannot be resolved, to the point of determining how to 'go on,' in a current predicament. For example, the laws of chess are simple, yet the pieces in the game can interact to such a degree that it is difficult to determine how to 'go on,' in certain positions, when playing an actual game.]
p.143 Nothing is clearer than that where we are uninformed about the processes at work, we cannot foretell the results they bring to fruition
[JLJ - Yet we can hire a consultant who is so informed, and ask this person to foretell the results.]
p.143 where reliable information about how wires are connected is absent, we cannot foretell what happens when the switches are thrown.
[JLJ - Not entirely true, we can watch someone operate the switches, and guess how the wiring is connected - and if we are lucky or smart, we even can guess reliably - and use that information in order to 'go on'. We can also play with the switches and make an educated guess. Swing and a miss, Nicholas.]
p.151 For rational prediction to be feasible, sufficient data must be available - and available to the predictor and at the time... The prospects of prediction are often blocked by our inability to obtain the necessary data with the requisite completeness and precision.
[JLJ - Yet how is one to proceed if an opponent plants false evidence? We are so attuned to cues and to making extrapolations from evidence that we respond even to evidence planted by an opponent who cleverly predicts that we will respond rationally to it. Truly rational prediction instead must rely on tricks that work, which likely include data, but also a certain amount of verification and testing of uncovered evidence. Ultimately, in life or in playing a complex game of strategy, our practical tricks that work must work better than an opponent's practical tricks that work. What is needed is a scheme or plan which acquires the data and does a certain amount of verification and cross checking. The Angle of Attack (AOA) sensor of the Boeing 737 Max airplane is effectively a predictor of this angle which works most of the time when intelligently assessing when the plane is approaching a dangerous stall condition, yet needs to be coupled with a practical scheme which uses it intelligently when it is judged to be accurate, and which dismisses it when it is judged to be inaccurate or when actions based on it - intended to help a pilot fly a plane - are deemed instead to be likely to hinder a pilot's efforts to intelligently control the plane.]
p.152 prediction becomes easier with aggregate phenomena, since in statistical aggregation we often cancel out the vagaries of individual fluctuations to produce a stable overall result.
[JLJ - ...yet prediction in these cases does not rise to a level of certainty. Intelligently crafted prediction always remains a premonition, and is arguably useful as part of a practical scheme to decide how to 'go on.']
p.153 in contriving models of real-world processes... one sometimes needs to suppress potentially significant lesser-order effects in order to be able to carry out the inferential process at issue... Where logical inference and calculation fails us, estimation need not do so, and where we cannot make exact predictions we may yet be able quite plausibly to make rough or approximate ones.
p.154 Predictability thus runs into difficulty whenever the causality of the issue sensitively involves several difficult-to-determine factors, which themselves depend in their turn on others of a like nature.
[JLJ - This pretty much summarizes the problems of game theory when playing complex games of strategy.]
p.155 Once present for any reason - chance, choice, uncertainty, or whatever - impredictability always ramifies over a far wider domain. For the world's processes constitute a fabric of cause-and-effect interconnections within which all those impredictable occurrences themselves proliferate further causal consequences that are thereby also bound to be impredictable.
p.156 Rousseau wisely observed, "The ability to foresee that some things cannot be foreseen is a very necessary capacity."
p.156 Our predictive capacity may be severely limited, but it is not - and in the circumstances cannot be - too radically impoverished. For if we were not as good at prediction as we actually are... then we just would not be here to tell the tale.
[JLJ - Kind of garbage. Insects predict poorly, but are incredibly resilient creatures that recover quickly from their many mistakes and are quite adept at finding food, reproducing and fighting. Insects will likely be 'here' long after we are gone. Prediction is useful in a complex or changing world, but so is instinctively directed action or posturing, and overall simply being a good match to/for the environment one is operating in.]
Prediction in the Sciences, Chapter 9, p.159-174
p.161 Prediction is the very touchstone of science in that it affords our best and most effective test for adequacy of our scientific endeavors.
p.165 Only a complex, reciprocally interactive gearing of explanation, prediction, and control can in the final analysis provide a satisfactory standard of scientific adequacy.
p.169 The complex interweaving of fact, theory, and conjecture in scientific prediction means that even good theories sometimes yield poor predictions... And contrariwise, even where we make successful predictions this will not necessarily mean the basis of theory from which they emerge is scientifically appropriate.
p.172 Predictive feasibilities and infeasibilities penetrate pervasively into every sector of the world and cannot be neatly prized apart.
Predictions about Natural Science: The Problem of Future Knowledge, Chapter 10, p.175-190
p.176 Nature is predictable insofar as its phenomenon exhibit discernible patterns that reveal rulish lawfulness in its operations.
p.177 Any human creation - be it a house, a dam, or a knowledge claim - is designed to function under certain known (or surmised) conditions. But the processes of change that come with time always bring new, unforeseen and unforeseeable circumstances to the fore.
p.183 We can reliably estimate the amount of gold or oil yet to be discovered, because we know a priori the earth's extent and can thus establish a proportion between what we know and what we do not. But we cannot comparably estimate the amount of knowledge yet to be discovered, because we have and can have no way of relating what we know to what we do not... The long and short of it is that there is no cognitively adequate basis for maintaining the predictive completeness of science in a rationally satisfactory way.
p.188 the amount of significant knowledge (K) that can be extracted from a body of information (of size I) is given by the relation: K ≈ log I. This Law of Logarithmic Returns, as it might be called, constitutes a prime predictive instrument in the matter of volumetric forecasting about the cognitive domain.
Predictions in Human Affairs, Chapter 11, p.191-210
p.192 there is - happily - much that we can manage to foresee in human affairs... But... Within broad limits, the fate of individuals - and especially those of its aspects that particularly interest us about ourselves - are veiled in a fog of impredictability.
p.197 Yet another of the main problems for economic prediction lies in the inherent complexity (nonlinearity) of economic systems. In any system whose workings are subject to a very large number of intricately interacting factors, there is going to be a great sensitivity to parameter determination, so that even a small variation on input values will amplify into substantial variations in output values.
p.198 The actual behavior of a real-life economy is faced with contingencies of innumerable sorts... whose aggregate effect is that even the best-made economic models cannot engender accurate forecasts. Choice, chance, and contingency undermine the stability of economic processes, so that the prospect of precise and secure prediction is severely compromised... various influential theoreticians argue... that economics cannot succeed with prediction at the level of exact quantitative detail regarding specifics, but can only succeed at the level of generalities, tendencies, and probabilities.
[JLJ - Nevertheless, if we use these 'generalities, tendencies and probabilities' to construct scenarios, and determine from these scenarios a level of readiness, we can adopt a posture in the present which will keep us mostly out of harm's way, and position us for opportunities as they emerge.]
p.198 policy guidance is one of the main aims of the macroeconomic enterprise. And here effective operation does indeed not demand categorical predictions, since even merely probabilistic considerations can provide serviceable and perfectly cogent guidance to action.
p.198-199 For purposes of rational planning in economic policy matters actual prediction is not necessary; it suffices to have information about the ways of influencing the probabilities of outcomes, recognizing here as elsewhere there is no practicable way of turning statistical probabilities into individual-case actualities... When one manages to shift the probability of a certain desirable future eventuation from .2 to .4 one certainly cannot predict the outcome with any confidence, but one has nevertheless managed to realize an eminently useful degree of control in the matter.
p.201 people's actions are highly sensitive to the nature of their beliefs.
[JLJ - A belief is a short-cut to action, but that does not mean that we cannot delay action, ponder more deeply, and/or take later actions of a different kind.]
p.202 The reality of it is that insofar as developments in social affairs reflect human choices, and insofar as these are influenced by circumstances... that develop in a chaotic way, the course of development in human affairs can only be predicted to a small extent.
p.202 It is the nature of the phenomology of the domain - its volatility, instability, and susceptibility to chance and chaos - that is responsible for our predictive incapabilities here, rather than our imperfections as investigators... In this domain, where the causal phenomenology at issue is so enormously complex, volatile, and chaotic, there is only so much that can be done. The limits of social prediction here lie in the intractability of the issues, so that there is little reason to think that the relatively modest record of the past will be substantially improved upon in the future.
p.204 a doctrine of punctuatued chaos... sees the operation of human societies as a matter of periods of local stability within which prediction is possible, interrupted at random, unpredictable intervals by chaotic transitions to a new and unpredictably different but once again temporarily stable order. In such systems, short-term predictions are generally possible because within limited periods there are patterns that can be discerned and projected. But periodically there comes a cataclysm that upsets all the apple carts, so that long-term prediction becomes infeasible. There is good reason to think that social systems... actually function in exactly this way.
p.205 In the context of agents interacting in the setting of human societies, human actions can and generally do have unforeseeable consequences.
[JLJ - ...yet few consequences which we cannot prepare plans for in advance, or develop adaptive capacities which are effective in general-purpose recovery, or in seizing new opportunities as they arise. A bear positioned in a stream near a waterfall does not know exactly when the next fish will come splashing by, but it is nevertheless well-positioned, and when they do come by the opportunity is grasped as it arises.]
p.206-207 The main reasons for the long-term impredictability of historical developments are plain to see. For one thing, chance and chaos come into it: developments over the longer term... depend too much on subtle interactions which, while virtually indiscernible at present and negligible in the short term, can nevertheless make an enormous difference to what happens over the long term.
p.207-208 major social cataclysms... do not spring forth unexpectedly ex nihilo. They come to realization gradually... First their possibility is noticeable, then their predictability can be discerned, finally their inevitability becomes apparent.
p.208 By providing us with an ampler understanding of what can happen and what may well happen rather than what will happen, history places us in a better position to react with intelligence, balanced perspective, and good sense to whatever does happen - no matter what this may be.
Fundamental Limits on Predictors, Chapter 12, p.211-222
p.213 some predictive questions are unresolvable as a matter of practical reality.
[JLJ - Then again, others are easy to resolve: will Nicolas Rescher EVER hire a surrogate editor to help fix the text errors in his published works? NO.]
p.213 If it is simply an answer that we want, we can contrive predictors that will oblige us. But the problem, of course, is to have a predictor whose predictions are probable or, in the case of a putatively prefect predictor, correct - on a basis that renders the prediction credible in advance of the fact.
[JLJ - Do we want an answer, or instead do we want a suggestion on how we might proceed, which we can then ponder in further detail? Intelligence hinges at root on a generation of - then pondering - of the strategic options at the moment, the candidates for wise action, and a selection of the one that offers the best prospects, for the moment, the minute, the month, or the millennium, after an imagined simulation of the consequences, possible or improbable.]
p.214-215 The predictive inaccessibility of the future of the world... roots in the very nature of things. It is an aspect of reality as such
p.215 Our conceptions limit our cognitive horizons and conceal various regions of future occurrence in an unavoidable darkness of unknowing. We can only ask that which we can imagine, and our imagination is circumscribed by the cognitive resources and limitations of our time and setting.
[JLJ - Yet we don't have to understand something completely in order to 'go on,' we just need to place it in the correct action category, and then to respond accordingly, from our carefully prepared and trained action repertoire.]
p.216 Let us return for a moment to our hypothetical predictive machine Pythia.
[JLJ - ...which is a pointless conception. Such a device, unless it were magic or supernatural, would simply use competition-tested heuristics, of some kind or another, which simply work when executed at the right time and place. It would or could be composed of nothing more than tricks that work, and Rescher ignores the composition of such a device - instead playing games with how one would or could use it. It would or could have an artificial intelligence, or instead be practically sensitive, such as NOAA-developed software predicting the path of a hurricane.]
p.218 Robert Heilbroner has sensibly remarked that "Visions of the future express the ethos of the time." Often as not, a forecast will inform us rather more reliably about the times and circumstances of the predictor than about the times and circumstances of the prediction.
[JLJ - I would think this to be true of certain future visions, but other visions (such as necessary for corporate planning and with a lot riding on the outcome) are likely to have a more practical sense and likely reflect an intelligent reading of the interacting driving forces, such that they are, and the new trends as they develop. Every time and age produce clear-thinkers who are not afraid to call the balls and the strikes, to think clearly, regardless of how they are perceived by their peers or commonly held standards of judgement.]
p.219 We incline to view the future through a telescope, as it were, thereby magnifying and bringing nearer what we can manage to see.
[JLJ - Yet the "future" so seen is and remains a premonition.]
p.219 The fact is that predictors, like the rest of us, usually suffer from conservatism, lack of imagination, and intellectual lethargy.
[JLJ - Speak for yourself, Nicholas. Properly used, a predictor should be considered successful if it helps to properly position us in the present, for the future-present that will (continuously) arrive. Anything else is nice, but not truly essential. The problem with knocking prediction is that we have to do it anyway, in order to 'go on.' Second order prediction might consider revisions or tempering initial or first order predictions based on known or suspected weaknesses of the method, or for risk mitigation.]
p.222 Prediction is a domain where following "common sense" will often let us down
[JLJ - Hidden complexity and carefully constructed strategy cannot simply be brushed aside by an opponent using common sense alone - the trap has been laid and is sprung the minute the opponent reacts the way things "seem" to be. A more rational approach is to handle situations strategically, practically, with a competition-tested scheme full of wisdom, and with abundant adaptive capacity, for those situations when things are not as they seem to be, due to hidden complexity and carefully constructed plans and schemes of a wise opponent. "Common sense" is a useful trick that works, but not all the time, and it is not the only trick in the book.]
Predictive Incapacity and Rational Decision Problems, Chapter 13, p.223-230
p.227 As is clear from the existence of interaction games like "Rock, Paper, Scissors" in contrast to Tic Tac Toe, there sometimes just is no effective strategy of action apart from brute random choice.
[JLJ - What about using tricks that work? For example, what if Sam and Bob are playing a fun, social game of Rock, Paper, Scissors (we always called it Rock, Scissors, and Paper, but what do I know?). Sam is approached by someone and told in confidence that Bob - his opponent in the game - always plays for a few minutes with friends and then offers to wager money on the outcome. His opponent Bob, Sam is told, always tries Paper on his first bet, Scissors on his second bet, and then tries to trick his opponents by playing Scissors again on the third bet. Sure enough, after a few fun rounds, Bob announces to Sam that he would like to wager a nickel per game, and sure enough, tries Paper on the first play and Scissors on the second play. Sam, acting on his inside knowledge of Bob's actions, offers to raise the stakes of the third bet to $10, confident that Bob will play Scissors on his third play. But Bob has cleverly planted the information that he will play Scissors again on his third play, and instead plays Paper to Sam's Rock, and wins $10. So tell me again Nicholas that "there sometimes just is no effective strategy of action apart from brute random choice"? Tell me Nicholas that there are no tricks for sending me and getting me to click on spam e-mail, click-bait, or getting me to open unsolicited junk postal mail, or getting me to answer unsolicited phone calls about free gifts that I have won merely for sitting through a sales pitch for a vacation resort, or trying to get me to shop at your store by sending me a 30 cents off coupon. Tell me again Nicholas that Internet web sites have no effective strategy of action in serving me ads, apart from brute random choice. Everywhere you look, Nicholas, people are using tricks that work, large-scale schemes of influence and awareness-steering, even in situations where there are apparently no effective strategies. Furthermore, tricks that work are being countered by tricks that work better. People are installing adblocking software and using Virtual Private Networks to connect to the Internet and using effective spam filters and 'do not call' lists and 'nomorobo' to screen phone calls. These examples make Rescher's 'rationality' argument laughable and quaint, in a modern world full of complexity, motivated creative agents pursuing goals, and traps and tricks in various stages of working.]
p.229 predictive situations can arise in which the resources of rational reflection fail us.
The Shape of Things to Come: Facing the Future, Chapter 14, p.231-246
p.232 the actual future differs crucially from the contemplated future.
[JLJ - Again, could it be that the future is simply an intuitive concept that helps us to strategically position ourselves in the present, so that when subsequent presents arrive, we are somehow ready/prepared for them, that we do not in fact see the future, but experience a premonition? The future we are concerned with is the future present, which we can contemplate to some degree, and which (in the form of a premonition) is useful as a guide to action, perhaps ultimately as part of a scheme to 'go on.']
p.235 In general terms, control - full control - is the capacity to intervene in the course of events so as to be able to make something happen and to preclude it from happening, this result being produced in a way that is not only foreseen but intended or planned. Control thus calls for the possibility of causal participation ("intervention") in the course of events ("to make something happen or preclude it") with a power that can be exercised both positively ("to make happen") and negatively ("to preclude from happening"). Control is a matter of potential, of capability or capacity.
p.236 A future we cannot foresee is a fortiori a future we cannot control.
[JLJ - Yet general-purpose adaptive capacity, practical sensitivity, tricks that work, and improvization can be used strategically in schemes of control, without foreseeing *the* future in specific detail.]
p.237-238 Inquiry is perhaps our prime tool for the management of current impredictability... Insurance is also a tried and true resource for impredictability management... Planning is yet another prime instrument of impredictability management... For planning purposes it often suffices to deal in contingencies and probabilities. Rational planning and comportment does not demand actual forecasting of the future's course, and is, accordingly possible and profitable even in a world that is substantially impredictable.
p.245 It is surely of the essence of the condition of humanity as it actually exists that we should live with a mixture of knowing and ignorance that may change in its proportions with the condition of the times but always hovers well between the extremes.
p.245 Evolution has attuned us to the world as it is.
[JLJ - ...yet the world 'as it is' is in fact complex, and requires a great deal of imagination to position ourselves correctly in it, in each and every predicament we are in, with the driving forces which exist around us the way they are. Merely seeing things the way they are does not tell you the way they will be.]
p.246 a theory of prediction is possible because there are various general principles at work in the predictive realm... The predictor's task is undoubtedly a difficult one... one does not face these challenges entirely empty handed, without the aid of distinctions, principles, and other cognitive instrumentalities able to provide useful aid in the work.
Notes, p.247-282
p.254 Oliver Wendell Holmes's legal realism... holds that a lawyer's counsel in legal matters is nothing other than a prediction of what the judges to whom the matter is eventually presented will decide.
[JLJ - Very similarly, a machine 'playing' a complex game of strategy is merely using complicated heuristics to predict what will emerge in the game, in positions beyond a reasonable planning horizon, and using such predictions in a diagnostic way to estimate the capacity or potential of a position to coerce, in the unknown future.]
p.256 Outside the context of grammatical examples and imaginative fictions, neither statements nor predictions have any serious interest for us in the absence of reasons for seeing them as credible.
[JLJ - However, when playing a complex game of strategy, the implied questions, 'How might I proceed?' and 'Other than that, what reconfiguration of pieces, initially appearing to be unpromising, might quickly or eventually turn around to favor us?' are always on our mind, so to restrict out attention only to 'credible' predictions is in effect a strategic mistake. Those who attempt to solve the chessgames.com daily Monday puzzle, known to be 'easy', usually try initial moves which involve sacrificing the Queen, no matter how initially unlikely the move appears to be. Chances are, the puzzle solution is not far from the initial guess. Rescher does not adopt my view that in order to proceed in our current predicament, we ought to use a strategic scheme or plan of some kind that involves a portfolio of tricks that work, in a world where the complexity of agents and driving forces often hinders our ability to use rationality as our guide for action.]
p.260 What past performance does not enable one to do is to predict with failproof accuracy.
p.261 experts... become experts through learning, study and practice. But the formation process for predictive expertise does not in general consist in imparting a mechanical set of rules. Few and far between are the areas in which one can operate predictively successful "expert systems."
[JLJ - Rescher is apparently not a big fan of Artificial Intelligence. But backing off a level, there are clearly wise, clever and therefore useful rules of thumb which apply in estimation, and which can be used as a first-level guide to action, perhaps in a rough determination useful in sorting candidates, such as for a job interview or for a college entry. We must decide at every waking minute we are alive how exactly it is that we 'go on,' and rules of thumb are critical and time saving approaches that operate to reduce the complexity of our current predicament in order for us to take postures and actions that are necessary and appropriate for future success - in a world where we only partially foresee what future will arrive.]
p.263 In areas where guesswork must be relied on, there is much to be said for making it more explicit and freeing it from individual vagueness and biases
p.263 Under the general heading of "judgmental processes" it is also proper to list the construction of scenarios. This methodology... is unquestionably a useful technique for futuristic studies and provides a useful tool for planners... However, since scenario construction addresses the issue of possible or probable courses of future developments and is not concerned with prediction as such, it falls outside the scope of our present deliberations.
[JLJ - Too bad for that. You, Mr. Rescher, may go about pondering "the" future, while I will go about pondering possible futures, and hence position myself and my adaptive capacity to be "ready" for whatever it is that does emerge. We concern ourselves with "the" future in order to strategically manage the present, whatever it is, which emerges at every moment, in our current predicament. "The" future will always remain a premonition, no matter how you approach "it."]
p.266 "Combination of Forecasts," ...The question is whether in combining methods the weaker are on balance offset by the more effective or whether the efficacy of the latter is simply diluted. The available evidence, though meager, speaks for the former prospect.
[JLJ - How so? Does Rescher ignore the fact that - in certain cases - a sophisticated forecasting tool can be competition-tested to the point that it adjusts itself for all reasonable situations where it otherwise would be inaccurate, and hence is in no need of 'modification' by lesser methods? In other cases - such as hurricane forecasting - the best model prediction-wise is a combination of other models. Essentially, tricks that work can be complicated and require careful modification in order to work effectively, approaching the situation faced by economists as they try to see into a void cloudy and dark by nature.]
p.270 [Pierre Duhem] "[T]he highest test, therefore, of our holding a classification [theory] as a natural [reality-characterizing] one is to ask it to indicate in advance things which the future alone will reveal. And when the experiment is made and confirms the predictions obtained from our theory, we feel strengthened in our conviction"
[JLJ - continues "...that the relations established by our reason among abstract notions truly correspond to relations among things."
p.273-274 The role of unforseeable innovations in science forms a key part of Popper's case against the impredictability of man's social affairs - given that new science engenders new technologies, which in turn make for new modes of social organization.
[JLJ - And so it goes in game theory. In complex games of strategy, the emergence of the unforeseen prevents us from seeing the future, but that does not prevent us from positioning ourselves in situations where good moves ought to emerge, like a bear positioning itself in front of a waterfall - it cannot foresee exactly when and where a fish will jump up, but the position is good and a fish will eventually jump up within range of catching and eating.]
|
