Where did I tell you to compare a 12 and 20? I've double checked my post, and I don't think I did. I don't see those numbers anywhere in my post. Also, in PF2's system, I don't think that that is a reasonable range to compare for two characters of the same race, class, and background combination. Do you think that two druid characters of the same race, class, and background combination could end up with a 12 and 20 Wisdom?Sigh. Yes, I think they could? Mechanically, it is possible as long as the background and race do not grant a Wisdom bonus. However, I do not think that is a point that matters.
Your premise is that stats are not a large definition of character.
Therefore largely different stats should not have a large impact on how a character plays, otherwise your premise is flawed.
The only way to test the premise is to compare largely different stats. If largely different stats do not result in largely different play experiences, then your premise is correct.
Except that your "example" to prove me wrong is disingenuous. If you expect me to take your examples seriously, they should be serious examples. If you were comparing a druid with an Wis 16, and a Wis 18, I would buy the plausibility of the example. That is very possible, and could be an actual variation in how the class is built. The premise of your 12 and 20 is ridiculous at face value.
And yes, I would agree that large gaps in ability scores should be roleplayed differently. But an 8 point gap is a big gap, which usually accompanies other major character differences. Like you'll probably never have an 8 point gap in Strength scores for two strength based fighters. You might have that gap between a Strength Fighter and a Dexterity Fighter, but already you're looking at variations between the character which will lead to significant choices (armor, weapons, feats, probably races and backgrounds as well) if that is your goal.
If someone is purposely playing against type for a class, they will obviously play their character different than someone who is playing on type, which is what an 8 point gap in a key ability represents. The ability score isn't the source of that roleplaying difference, but more a consequence of that different roleplaying goal.
All other things being equal, are you going to play a druid with a 14 Con, 16 Wis significantly different than a druid with a 12 Con, 18 Wis (assuming race, background, alignment and backstory are all identical)?