sion proved by a number of alternative pairs of Premisses; and it may therefore be argued that it is impossible for the truth of the Consequent to imply that of any one Antecedent, or of the Conclusion to imply that of any one pair of Premisses. But this argument forgets that the appearance of non-reciprocal implication is here due to superfluous elements—bare conjunctions—in the numerous alternative Antecedents and pairs of Premisses, and that there most certainly is in every case some Antecedent, common to all the number, and some pair of Premisses, underlying all the alternative pairs, the truth of which is implied in the truth of the Consequent or Conclusion. In "If he is poisoned he is dead"—death does not imply poisoning, but it most certainly implies the features of poisoning which are essential to death, and common to all its modes. The appearance here of non-reciprocal implication is simply due to the fact that we take our rules from unscientific thought. In every Conclusion there is some pair of Premisses implied, in every Consequent some Antecedent; and reciprocally, as we saw, there is a strictly conditioned Conclusion implied in every explicit deduction in respect of every actual Premiss that can enter into it, and if there seems to be an opening for falsehood from true Premisses, it is because bare conjunction, that is, facts inadequately conditioned, have been admitted into the chain of reasoning. I therefore agree with Mr. Spaulding's account of the matter (pp. 225-6) except that what he takes as occasional fact, appears to me to be a truth of principle.
The strange contradiction between these rules
