This became harder and harder to accept as a theory, though, through the 18th century. Particularly problematic were the cases of materials -- magnesium is the obvious one -- which actually gain weight as they burn. So, phlogiston proponents started to amend the theory, claiming that phlogiston was either lighter than air or actually had negative mass, or arguing that phlogiston was more of a general term, not precisely a substance.
Note what's happening here. The theory is proposed -- that intuitive leap I talked about earlier. The theory runs into trouble -- it is inconsistent with observations. As a strict matter of logic, one could just reject the observations. After all, any observation may be mistaken. Perception isn't perfect, and when we're dealing with precise measurements of chemicals, metals, and gases, it can become easy to attribute contradictory observations to observer error rather than problems with the theory. However, once observations start to stack up, it looks more and more like the theory is wrong. So, to seek the truth, we amend the theory, as little as possible in order to accommodate the observations.
Unfortunately, Antoine Lavoisier argued decisively that even the amended theory of phlogiston had to be wrong. Lavoisier heated a material -- tin -- in a set of closed vessels. The point of using closed vessels was to ensure that nothing could enter or leave the experimental setup during burning. If the phlogiston theory, as amended, was correct, then we would expect that heating would cause the tin to lose phlogiston to the air; and further heating should cause it to lose more and more phlogiston, until it was entirely dephlogisticated (seriously, that was the word). However. Heating the tin initially seemed to prove phlogiston theory right -- the tin became a mercury residue, and the volume of air measurably decreased. That's consistent with phlogiston theory, as the air is absorbing the emitted phlogiston. However, further heating caused the residue to revert back to tin, and the volume of air to increase. According to phlogiston theory, that's just impossible, and no jimmying around with weights is going to make the theory work.
Here's the argument:
- Phlogiston theory predicts that tin, when further heated, will remain a mercury residue.
- Tin, when further heated, becomes tin again.
- (1) and (2) contradict.
- Therefore, either phlogiston theory is wrong or the observation is incorrect.
- The observation is not incorrect (because repeated).
- Therefore, phlogiston theory is wrong.
Better reasoning through logical technique gets us closer to the right answers. One more example.