Logic or dialectic (for these two terms have often been employed synonymously) has been defined in many different ways. Indeed, the definition which a logician gives of his science will usually indicate to what school he belongs. Some of these variations arise from the different ways in which the sciences have been classified without importing any difference either in the subject-matter of logic or the method of treating it; for since a definition usually refers the word or thing defined to a class, disagreement concerning classification will result in disagreement concerning definitions. Other variations have arisen from different opinions in regard to the method in which Logic ought to be investigated. But there is also a great diversity of opinion as to what ought and what ought not to be treated of in a book upon logic.
It is a historical fact that logic originated in an attempt to discover a method of investigating truth. Moreover, the doctrines of logic, as they exist, centre about the forms of inference. Sebastianus Contus has acutely observed that although Scotus and his school profess to regard logic as a purely speculative science yet in their whole method of treatment of it, they show that they really consider it in a practical point of view; and the same may be said of most of the other writers who term it a purely speculative science. On the whole, therefore, we cannot utterly contemn that definition with which Petrus Hispanus opens his celebrated Summulae, the classical work upon logic of the middle ages, "Dialectica est ars artium scientia scientiarum, ad omnium methodorum principia viam habens. Sola enim dialectica probabiliter disputat de principiis omnium aliarum scientiarum." In short, we may state it as a historical fact that logic has been essentially the science of the structure of arguments, whereby we can distinguish good arguments from bad ones, can estimate the value of an argument, can determine upon what conditions it is valid, how it needs to be modified, and what can be inferred from a given state of facts.