Not necessarily. It wouldn't even take magic for such a thing to make sense.
Destroying is easier than creating, as a general rule of nature, so a potential rule of casting times might be that countering a spell takes half as long as the spell itself took to cast. If it takes two seconds to cast a Fireball, then it might only take one second to cast a Counterspell against that Fireball
I have no problem at all with Counterspell having the shortest casting time in the book, to allow it to work against any other spell...
in which case it might only take half a second to cast a Counterspell against a Counterspell against a Fireball.
...except itself. I do have a problem with it being able to be shorter than itself, as this implies that not all Counterspells are created equal. And it would only take one petulant player to say "But *I* cast the shorter version" and you've got a big-time headache....
If you then consider that spells take effect instantly after they are cast, then the casting time would logically also include flight time for that spell energy to reach its target; and if Counterspells are faster than the spells they are countering, then an easy way to visualize this is that Counterspell is literally intercepting a spell in-flight, and a Counter-Counterspell is intercepting the Counterspell before it reaches the first spell that it was targeting.
Again I'm behind this all the way except for one Counterspell being, in effect, beginning-to-end faster than another. I think the designers must have been looking at Magic: the Gathering when they dreamed this up, as that game does work on a last-in first-out model. But in D&D it's always been that any spell* takes 'x' amount of time to cast (e.g. 3 segments, 1 action, whatever; depending on edition) and that for any particular spell the 'x' time value is always the same e.g. a Fireball in 1e is always going to take 3 segments and in 5e will always take 1 action. This means that two spells that take the same amount of time to cast (which will [or should!] always be the case when it's the same spell being cast twice) will logically always go first-in first-out.
* - certainly any spell that affects combat. Some rituals (or ritual-like spells from older editions) have variable casting times; but I think we're talking only about combat here.
Now, the way 5e RAW reads they've in fact made Counterspell's casting time variable; but with the variance not chooseable by the caster as it's locked in by the sequence in which it is cast with relation to other spells. They also seem to make it legal to cast while in process of casting another spell, which I think is an all-time D&D first. Why is this a fail? Because it...
- defies internal (game-world) logic as I've mentioned above
- is inconsistent with the rest of the game's design aesthetic (they've tried to reduce or eliminate exceptions and oddities, this is both)
- is inconsistent with how spell interruption otherwise works (usually a spell is lost if casting is interrupted [yeah yeah, combat casting, whatever] but here the caster can interrupt herself without problem then resume as if nothing happened)
This is one where I'd really like to ask someone like Mearls what they were thinking that led them to come up with this...or whether this was an intended result at all.
Lanefan