422 Prof. Forbes on the evidence for a Physical Connexion 



fusion between the expectation of a given event in the mind of 

 a person speculating about its occurrence, and an inherent im- 

 probability of an event happening in one particular way when 

 there are many ways equally possible. Secondly, a too limited 

 and arbitrary conception of the utterly vague premiss of stars 

 being " scattered by mere chance, as it might happen ;" — a 

 statement void of any condition whatever. 



39. Having thus stated broadly the errors of deductive 

 reasoning (as I conceive them to be) to which the sanction of 

 many eminent authorities has been given, I would guard 

 myself, once more, in conclusion, from being supposed to 

 doubt either the fact of the physical connexion of double stars 

 generally, or to deny that Mitchell was right in suspecting 

 that fact from the mere circumstance of the proximity of two 

 stars observed in so many cases. This was, on his part, no 

 more than the first stage of an induction (the nature of which 

 is now so well understood by all students of physical science), 

 being a mere collecting and arranging of a certain class of 

 facts, and assuming the probability of a common cause to be 

 afterwards investigated. Mitchell's appeal to the mathematical 

 theory of probabilities was a false step, tending to weaken not 

 confirm his argument. Still less do I object to the inference 

 drawn by some of Mitchell's distinguished successors in favour 

 of the physical connexion of binary stars from the following 

 fact, viz. that if we arrange such stars in groups, each of which 

 includes double stars whose angular distance is between cer- 

 tain limits, as for instance from 0" to 4", 4" to 8", 8" to 16", &c, 

 the class of close stars is actually far more numerous than 

 those contained within wider limits, and which we might there- 

 fore fancy ought to be of more frequent occurrence. This is 

 a very curious and a very striking fact, and carries the process 

 of induction one step higher than it was carried by Mitchell. 

 It shows a greater frequency of occurrence for small distances, 

 and furnishes something like a law connecting those two 

 things, the fact of duplication and the angular distance. Such 

 an empirical law, if well supported by facts, is undoubtedly a 

 good inductive argument. We are brought nearer to the 

 conception of a cause by this observation, and our conviction 

 is complete when we observe one such star revolving round 

 another with a clear orbital relation. 



40. I should hardly have thought it necessary to state so 

 much, had not my views been strangely misunderstood, par- 

 ticularly by an ingenious writer in the Edinburgh Review*, 



* Edinburgh Review, July 1850. 



