... and other forms of the two: Inferencer v Reasoner, etc. Interested in both theory and practice.
Reasoning is a more general term which generally applies to somehow drawing new conclusions from a set of premises. There are three categories of reasoning which are
- deductive reasoning which is based on combining premises to make conclusions, such that if the premises are true, the conclusion must also be true (all Toyotas have four wheels, Freds car is a Toyota, therefore Fred's car has four wheels).
- inductive reasoning which is based on application of generalisations to specific scenarios, is not always certain, and is typically what us humans do (almost all cars I've seen have four wheels, therefore Fred's car likely has four wheels)
- abductive reasoning involves trying to explain a conclusion by guessing a premise (Fred bought something with four wheels... it's probably a car).
Deductive reasoning can be thought of as deriving a conclusion from a known rule and a known set of premises.
Inductive reasoning can be thought of as deriving a general rule from a known set of premises and their conclusions.
Abductive reasoning can be thought of as deriving a possible premise from a known set of rules and a conclusion.
Inference is moreso the means or mechanisms by which reasoning occurs. Reasoning is the "goal" whereas inferencing is the "implementation".
One of these questions that may easily lead to long philosophical debates... but, anyway: :-)
Inferencing is clearly a kind of reasoning. I think one might argue that certain reasoning tasks such as satisfiability detection for classes and ontologies do not really count as inferencing, since there are no inferences as the result. Or the detection of non-entailments. However, I would say that the opposite, the detection of entailments, is inferencing. After all, it's a quite nit-picking discussion.
On a higher level, one could say that a reasoner does more than just reasoning (and, thus, more than just inferencing), since it also loads, parses and preprocesses ontologies and does other jobs that a complete software system does.
Ok, not more on this, back to real work! (Oh wait, what am I doing here right now: working on reasoning or on inferencing? ;-))