Here is a useful resource comparing equal temperament (often abbreviated ET or 12-TET for the 12-tone scale) to just temperament.
"Just intonation" is the result of ensuring the ratios between waves are such that the waves line up without creating extra aural beating. This is achieved by simple ratios like 1:2, 2:3, 3:4, etc. These ratios are sometimes called "perfect" or "just." If you picture two sine waves superimposed, a 1:2 ratio will feature peaks that line up every time the longer wave reaches its period - this is what makes octaves sound so consonant. A 2:3 ratio will feature peaks that line up every
other time the longer wave reaches its period. This kind of tuning is only possible with relation to a fixed point - for example, you can achieve a pure fifth between C and G, and a perfect third between C and E, and a perfect fifth between E and B. But then you will not have a perfect third between G and B. This is what makes just intonation sound different in different keys, and is the reason that the keys in modal music have associated "characters" or feelings.
There is another problem that is often misunderstood or misrepresented: even for a diatonic instrument that plays in only one key, perfect just intonation is not possible across multiple octaves. This is due to the fact that the scale of frequencies and their corresponding pitches is not linear but exponential. The difference between the tuning possible by calibrating the fifths perfectly, and that possible when calibrating the octaves, is called the "Pythagorean Comma" - the deviation between 7 octaves and 12 perfect fifths (which "should" theoretically be the same note). One cannot tune all the octaves to 1:2 ratios, and then achieve pure fifths, nor can one tune all the fifths to perfect 2:3 ratios and then get the octaves to match.
en.wikipedia.org
This is part of what makes tuning a Hard Problem, and one which demands compromises - even for just-tempered instruments! These compromises are sometimes called "compromise tunings," or "well temperaments." They are various approaches to mitigating the damage done by ET. "Well-tempered" does not indicate a single type of temperament; there are many well temperaments. To add to the confusion, the term "well temperament" is sometimes mistakenly used when "equal temperament" is what's meant.
en.wikipedia.org
In my view, one of the ways the accordion (at least, a wet-tuned one) mitigates or obfuscates these problems is to embrace aural beating. When each note actually plays two pitches, separated by roughly the difference between a just major third and an equal-tempered third (14-15c), there will be an interval that is close to a perfect third in the aural "mix" when one plays the ET third.
[Mistakenly wrote "fifth" here earlier, good catch Paul] And even when the just interval is nowhere to be found, the resulting aural beating is masked by the beating of each individual note. That said, a wet-tuned instrument that is tempered to a compromise tuning (such as an Italian organetto) has its own unique character and the pleasure of the consonant 3rds and 5ths can still be felt despite the complexity of the soundwave interactions.
Sorry to stir the mud! I think it's a really interesting subject and one with many approaches.