The article is written by S N Balagangadhara, on his own blog. I thought it worth sharing.
I have discovered that there is a fundamental difference between arguments and theory-building. As a philosopher, I have come to think that one could argue almost any position, within reasonable limits. Mostly, they consist of putting across plausible, or even logically possible considerations in order to show that either some point is plausible or that it could be true. Up to a point, arguments have a function: they force you to reason, check for inconsistencies, train your thinking process, etc. However, you must remember that ultimately all you have done is make a sentence or a set of sentences sound plausible or shown it to be logically consistent and possible. More often than not, it has a psychological purpose as well: that of shooting down some person, demonstrate intelligence, exhibit stupidity and so on. However exhilarating it might be at times, this is a very unproductive occupation, if carried on too far: one does not advance knowledge a great deal.
The second too involves reasoning, logic, etc but it tries to build some kind of a theory. Such an activity advances hypotheses, rejects or reformulates them, tries to solve problems and so on. Even where proved wrong, a fruitful hypothesis tells us something about the world. I find this an entirely different kettle of fish: it is far more difficult; it is subject to many more constraints than the first one; but it is even more exhilarating than the first one.