I'm curious as to how rules of inference are established. Is this an empirical act? For example if I know the premises "If I jump, then I fall" and "I jump" is it truly valid to conclude that "I fall"? If rules of inference are empirical observations is it possible that there are many unknown rules of inference?

  • See philosophy.stackexchange.com/questions/6913/…
    – That Guy
    Commented Aug 12, 2014 at 1:40
  • 1
    i dunno, your question could be clearer. established by whom? IMHO there's ample opportunity for proving basic axioms like the law of identity IRL. try it ;-) !
    – user6917
    Commented Aug 12, 2014 at 21:31
  • 1
    that's a little facetious but i was kinda suggesting that standards of proof etc. are contingent
    – user6917
    Commented Aug 12, 2014 at 21:42
  • 1
    You can refer to Georg Kreisel's paper : Informal Rigour and Completeness Proofs [in Imre Lakatos (ed), Problems in the Philosophy of Mathematics. North-Holland. 138--157 (1972)] for an analysis of the relations between : "informal" validity and its rigorous mathematical logic counterpart. Commented Aug 13, 2014 at 9:53

2 Answers 2


This question in the philosophy of logic is studied as the nature of the relation of Logical Consequence (SEP Article). Broadly speaking, a construction relating a number of premises to a conclusion is a valid argument if and only if for every interpretation in which the premises hold, the conclusion/s hold. But what kinds of interpretation are we interested in? Perhaps an argument that seems logically valid to one person who only accepts a certain model of acceptable interpretations might not to someone who accepts a more narrow delineation of the possible ways things might be. (perhaps some of our core concepts are Intensional ones, which invites a much more complex connection between how things are and how things might be in our interpretive understanding)

Formal work tries to use mathematical fields like Model theory and Proof theory to help explore how we might split up our domains and our arguments in a structured way so as to help explore what a logical consequence might look like. Intuitively we do want to use certain connectives with well-established behaviour (like And, Or, Not, All, Some etc.) to help standardise some of what we do in fact appeal to in our practice of logical reasoning and inference. The mathematics of these connectives can be very informative as to what makes them tick, but this is more of a scaffolding than an explanation of what, exactly, you might mean in a given validity claim.

  • 1
    hey i am just asking for clarification Paul! when you say "if and only if for every interpretation in which the premises hold, the conclusion/s hold" are you using 'interpretation' in the intuitive sense such that a mystic could deny sound arguments simply by asserting that their experiences say otherwise?
    – user6917
    Commented Aug 12, 2014 at 21:39
  • 1
    @user3293056, I think a logician could fairly accuse a mystic trying to weasel out of a logical argument by adding new interpretations of engaging in an equivocation - the mystic isn't undermining the validity of the argument but challenging the semantics of the language; agreement on which has to be assumed for logical argument. The thing is, there is nothing illogical as such about arguing for alternative interpretations of the language in which a debate is phrased - it's just that this isn't any kind of challenge to the validity of the original argument, and is out of place mid-debate.
    – Paul Ross
    Commented Aug 12, 2014 at 21:47
  • hmm, ok cheers [you very nearly lost me - but not quite, i think]. hope this doesn't sound rude, but i find this thesis pretty funny - can you comment somethiing quotable for me to post up on my facebook lol ?
    – user6917
    Commented Aug 12, 2014 at 21:50
  • Hm. It's not rude, but I'm not quite sure what you'd be looking for. Maybe we can argue later whether ghosts exist or not, but right now I want to reason through the ways the murderer might have got into the locked room?
    – Paul Ross
    Commented Aug 12, 2014 at 22:01

From Wikipedia (so take with a pinch of salt but this one seems accurate)

Inference is the act or process of deriving logical conclusions from premises known or assumed to be true.

Greek philosophers defined a number of syllogisms, correct three part inferences, that can be used as building blocks for more complex reasoning. We begin with the most famous of them all:

All men are mortal
Socrates is a man
Therefore, Socrates is mortal.

The reader can check that the premises and conclusion are true, but Logic is concerned with inference: does the truth of the conclusion follow from that of the premises?

To infer truth from an argument it must be both valid and sound. A valid argument is one where it is impossible for the premises to be true and the conclusion to be false in the above example, this is the case so the syllogism is valid and sound.

A sound argument is one where all the premises are actually true. This prevents the garbage in/garbage out effect that computer programmers will be familiar with where a valid piece of code gives an incorrect response because the variable are wrong.

Here are a couple of examples:

  • All men are mortal
  • Socrates is a mortal
  • Therefore Socrates is a man

This is invalid but sound. All the premises are true but the conclusion is false. Socrates could be an animal or a strangely named woman.

  • All men are immortal
  • Socrates is a woman
  • Therefore Socrates is mortal

This is unsound and invalid. At least one of the premises is false (P1) Socrates may be a woman (we only have Plato's word otherwise) and even if both premises were correct, there is nothing to infer that being a woman makes one mortal.

This is an entire field of study in itself with a lot of bearing on Mathematics. If you are interested, check out Matt dilahunty's YouTube video Foundations 1.0 on the Logical Absolutes for a beginners guide.

You must log in to answer this question.

Not the answer you're looking for? Browse other questions tagged .