Fourth Condition (JTB+G) approachesIn general, exceptions ruin explanations.
The most common direction for this sort of response to take is what might be called a "JTB+G" analysis: that is, an analysis based on finding some fourthcondition "” a "no-Gettier-problem" condition "” which, when added to the conditions of justification, truth, and belief, will yield a set of necessary and jointly sufficient conditions.
For example, if I propose that when I jump off a building I will float and not fall, this is an exception to the theory of gravity. But it's not an itty, bitty thing that has consequences isolated to this one day, one building, one person. It completely ruins the theory of gravity for all cases. Why should the explanations about curved space-time (or the more common sense knowledge about how heavy stuff falls to the earth) not apply to this particular case? *No reason*. So, that's a disaster. When do they apply, and when don't they? Why those times and not the others? We have no idea what's going on! Just because this one exception messed things up.
If it really happened -- I jumped off the building and floated *just once* -- this one exception would interest scientists worldwide and make them rethink gravity in general, and other related physics too. It's consequences could not be contained by suggesting "the theory of gravity plus the theory of Elliot floating on March 7"
So getting back to the article:
They have a theory that K=JTB (knowledge is justified true belief). This theory, and the *reasons for it*, are intended to apply to all knowledge. Just like the theory of gravity's explanations apply to all people and all places.
Then they find an exception. So someone proposes a new theory: K=JTB+G. It's the old theory with an exception tacked on for the Gettier problem. It's exactly the theory we found ridiculous in the gravity case. You can't contain things like this. Epistemologists worldwide should be going, "Oh my God, we can't have exceptions! Something is very wrong here!"
Though I have to say, K=JTB was never a reasonable explanatory theory in the first place. It did not contain reasons it should apply to all knowledge, they just guessed that it did because they couldn't think of more exceptions. In fact, they seem to have started with the idea that whatever we think of is knowledge, and then eliminated three common exceptions. What's the big deal in throwing out one more? Only that checklists are the wrong approach to epistemology and finding that they are incomplete should be considered a hint that you've got the wrong approach.
PS I realize there were other responses given. If you think any of them are good, and say why, I'll be happy to comment.
PPS This is mostly explained in