[ Pobierz całość w formacie PDF ]
.Performance-rules in general seem to fall into two main classes which Ishall, for brevity, label  Procrustean rules and  canons.Procrustean rulesare those which enable us to grade a given performance as correct orincorrect, legitimate or illegitimate.The rules of grammar and spellingare, with a few exceptions, Procrustean; so are the rules of the road, ofchess, cricket, syllogistic reasoning and rifle-range practice.Quite different from these are the rules or canons of style, strategy,prudence, skill and taste.A good chess player does not only not breakthe rules of chess, he observes a number of tactical and strategical prin-ciples as well.He is not playing chess at all if he does not observe its Pro-crustean rules; he is not playing efficiently unless he also applies maximsof generalship.Similarly Gibbon did not just write correctly, he wrotewith power and elegance.He observed canons of style as well as rules ofgrammar.Procrustean rules can generally be expressed in brief formulae or terseorders.These can be memorised and breaches of them can be promptlydetected.Observance of them can be inculcated by sheer drill and sobecome, at least in normal circumstances, automatic or habitual.We donot now have to wonder how to spell correctly or talk grammatically.Canonical rules, on the other hand, commonly resist codification.Theycannot be memorised nor can observance of them become an habitualroutine.We learn them by practice, but not by sheer drill.They are taughtby criticism rather than by rote, and we have never finished learning them.CHAPTER 16: CALCULUSES OF LOGIC AND ARITHMETIC 241All methods, techniques, crafts, skills, etc.are subject to canonical rules.The faults we find with performances are for breaches of such canons justas much as for breaches of Procrustean rules.There are canons of experimental method.A good experimentalist hasto be careful, accurate, neat, ingenious, patient, self-critical, self-confident,exploratory and so on.There are also canons of purely theoretic method.Agood mathematician does not merely avoid misreckonings and fallaciousdemonstrations, he also has a sense of direction, a nose for elegance and acapacity for recognising the fertility or infertility of new theorems, andthe relative powerfulness or powerlessness of new methods of proof.Theformal logician himself in selecting, ordering and proving his Procrusteanrules of inference is guided by similar non-Procrustean canons.Now the principles of induction seem to resemble canons rather thanProcrustean rules.They are akin rather to the tactical and strategic maximsthan to the rules of chess.Experts find fault with indifferent researchers,not on the score that any of their operations have been incorrect but, forexample, on the score that they were rather aimless, or that they wererather rash or rather over-cautious or that they were muddled orunoriginal.There is no decalogue of Inductive Fallacies and no CodeNapoleon of the Rules of Induction.So some rules of logic are notProcrustean legitimacy-rules.However this makes no difference to my earlier assertion that it isnonsense to ask how or why rules of logic apply to the world.Both theProcrustean and the canonical rules of logic are performance-rules.Onlyperformances can be or fail to be in accordance with them.If they areapplied, that is a fact about the efficiency and intelligence of theorists, nota fact about any radical docility of the world.IIIWhat is it like for someone to operate in accordance with a performance-rule? It is sometimes supposed that he has to go through three or fourstages: (1) to consider and/or accept an open-rule formula; (2) toconsider and/or accept that specification of this rule-formula which isappropriate to his situation; (3) to construct in theory a plan of operationconforming to the rule and to its specification; (4) to put this plan intooperation.(These stages are sometimes quaintly assimilated to the steps ofa syllogism.Doing something according to a rule is then described as242 COLLECTED PAPERS: VOLUME 2going through a Practical Syllogism.But drawing a conclusion in Barbarais a special case of operating in accordance with a rule.Operating inaccordance with a rule cannot therefore be described as a special case ofdrawing a peculiar conclusion in Barbara.)Now it is easy to show that this four-stage process is a myth.For to gothrough these four stages would itself be a process with a correct procedure,and so be subject to a higher-order rule.Moreover of the first twosupposed stages the agent would have to pick on the appropriate rule andthe appropriate specification of it, yet his application of the requiredcriteria of appropriateness would be just another piece of operatingaccording to a rule [ Pobierz całość w formacie PDF ]