Quoted By:
If they ever reference in canon, it would probably akin to how science handles it:
"We originally lumped Pokemon like Gardevoir and Azumarril with the other types simply because there wasn't much clear evidence showing they had a difference in the quality and attribute of their attacks. However, it didn't escape us that they never meshed well with those types--either partly or completely. This situation was completely underlined by the great number of Pokemon that the Kalos Region has introduced to us which we simply couldn't--with ANY confidence--apply any type of the previous types to.
Therefore, we feel extremely confident that we have discovered a completely new type, and we have thus pooled all the previous Pokemon that shared those similar attributes together as part of that new type.
Bear in mind, other Pokemon also don't exactly seem to fit in the types we've attributed to them--such as Whismur (and its evolutions) and Kricketune--but we aren't in any was confident that they belong to a completely new type; the evidence doesn't seem to suggest as much. But like all good scientists, should the evidence prove otherwise, we'd be more than willing to introduce yet another type."
> TL;DR: It would be like how real world scientists try to sort out which genus a specific species would belong to.
