Are there logics with fairly natural semantics that don't have disjunction introduction? Is there a good way of thinking about what "flavor of system" you're in when you don't take disjunction introduction? Systems without double negation elimination seem have a "constructive" feel to them. Is there a nice "family resemblance" of sorts between logical systems lacking disjunction introduction?

Here's a somewhat artificial example of a three-valued logic constructed not to admit disjunction introduction while still having fairly symmetrical truth tables and agreeing with classical logic when the intermediate truth value U is ignored.

Here's a three-valued logic that doesn't admit disjunction introduction, and has the connectives K (conjunction/koniunkcja), A (disjunction/alternatywa), and C (implication/implikacja). We pick both T and U as our designated truth-y truth values.

Kab      b        Aab       b             Cab    b
      F  U  T            F  U  T              F  U  T
  (F) F  F  F        (F) F  F  T          (F) T  T  T 
a (U) F  U  U      a (U) F  F  T        a (U) F  T  T
  (T) F  U  T        (T) T  T  T          (T) F  T  T

So, we can see that disjunction introduction doesn't hold by demonstrating that CaAab is a non-tautology. Let's assign both a and b the intermediate truth value U.

  • It seems to me that disjunction introduction does hold according to your truth-table, since whenever a is true or b is true, you have Aab true. But because your truth-table for the conditional is not standard, this isn't reflected as a tautology expressible with the conditional. – Ben Dec 12 '18 at 6:57
  • @Ben ... U is a designated truth value. If I take a as a premise, then it can either have the truth value T or U. a=T would allow me to conclude that Aab, but a=U would not. In this section of the article on three-valued logic: en.wikipedia.org/wiki/… , LP and Kleene's logic have the same truth tables but different designated truth values. – Gregory Nisbet Dec 12 '18 at 7:10
  • I'm sorry I assumed U was undesignated. That makes your truth-table for conjunction the same as Priest's in LP, but the disjunction table different - and it does indeed invalidate disjunction introduction (but not I think disjunctive syllogism?). – Ben Dec 12 '18 at 7:15
  • 1
    @Ben ... Yes, I think disjunctive syllogism still works. If Aab is T/U but b is false, a has to be T. Interestingly b being U is also enough to eliminate the disjunction. – Gregory Nisbet Dec 12 '18 at 7:28

In some paraconsistent and relevant logics, there's a distinction between intensional and extensional disjunction. Disjunction introduction fails in those logics for intensional disjunction.

The motivation is that these logics reject the principle according to which A, ~ A entails an arbitrary sentence B. The argument for this principle involves assuming A, using disjunction introduction to get A v B, then using disjunctive syllogism with ~A to get B.

One way to resist this argument is to reject disjunctive syllogism (this is the course most relevant and paraconsistent logics take). But another way is to reject disjunction introduction instead - this is the way that gets you "intensional disjunction".

There is a brief discussion of this in the following section of the SEP article on disjunction: https://plato.stanford.edu/entries/disjunction/#DisjSyllAddi

There are probably helpful points in the rest of that entry too.

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.