To enact linguistic behaviour designed to produce
assent in one or more language users. Engagement in this behaviour by one or two or more language users is termed an
argument:
There are (it would seem) two main senses:
- 'X, in his seminal paper Y, argues persuasively for Z'
- Before, we were having a rational discussion, but now you're just arguing.
But that's not to say it's always clear which sense is appropriate in a given case. One could argue that we choose sense 1. when we agree with the arguer and 2. when we
disagree.
A question, therefore, arises as to what means are available for settling an argument.
Any offered criterion for settling an argument can be contested; the meaning of any term may be questioned; the argument may even degenerate into a pre-linguistic means of determining the outcome.
Because of these, and other reasons, much thought has been devoted to devising incontestable means of settling arguments, resulting in mathematics, the sciences, and, arguably, politics, law, religion, and other military technology.
Quite early on, people realised that an ad-hominem argument (or the pre-linguistic equivalent) though occasionally effective in individual cases, does not ultimately prevail. But the exact difference between a valid and an invalid argument is still an open question. What little progress has been made has largely consisted of attempts to restrict the syntax, terms and referents of the language being used (and, indeed, creating whole restricted languages in which to argue.)
This question is perhaps the proper subject of philosophy,
on the grounds that philosophical questions manifest as arguments, and philosophy has no purpose other than to answer (or otherwise disperse) these questions.
Or, if that doesn't convince you: any counter-argument to the above view (a philosophical one) would necessarily also represent a philosophical position.