Yeah, even as I was writing it I was thinking "this is far too brief of an explanation".
In the example you cited, why would Detect Magic count as affecting dice rolling? Because presumably if magic were detected it would impact dice rolling in the future. Maybe just because you don't need to roll a saving throw because you detected a magical trap. And if there's zero possibility of a connection between the detection of magic and future dice rolls, then maybe it doesn't actually require a mechanical casting of a spell?
I don't think that's an absurd way to go, but it's not the way that I would go.
Ron Edwards, following Jonathan Tweet, classifies resolution systems into
fortune, karma and drama. I think you are focusing on
fortune to the exclusion of the other modes. (D&D spell casting is a mixture of karma - spell slots, for instance - and drama - when I cast the Detect Magic spell, the GM comes under an obligation to tell me certain stuff.) When I think about
mechanics, I tend to think about the (more-or-less) binding processes for establishing the consequences of a player's action declaration for his/her PC. (There's also systems for character building, world building, etc, but they're not relevant to the ZoT issue.)
Because that's somewhat vague, that means the boundaries of "what's mechanics" is not clear-cut. Is it a
mechanical thing that, if the GM has narrated "The house is made of timber" and I then state "My guy fireballs the house", the GM has to at least
think about whether or not the house catches alight? Until you tell me what's at stake in answering one way or the other, I'm not sure that I care all that much.
In the case of ZoT, I think that the narration needed to make 5 INT Eloelle work is (obviously) non-standard. It requires drawing a distinction between at-the-table effects (
What knowledge does Eloelle's player have access to? and
Does Eloelle's player have freedom of action declaration - if the save is failed, then no, s/he doesn't!) and in-fiction states of affairs (
What does Eloelle know? and
Is Eloelle's mind controlled by the magic?) which the spell description doesn't itself draw.
Is that change in narration a mechanical change? I'm not sure what's at stake in answering one way or another. Clearly the Eloelle narration changes the shared fiction. That might be relevant to action declarations and framed scenes down the track (eg we might hope for some sort of big reveal about the patron's plans for Eloelle). What happens if Eloelle's PC gets an INT-boost item (say a
Gem of Insight or a
Tome of Clear Thought, to use the old AD&D item names)? Does this free her to some extent from the domination of her patron (given the names of those items, that doesn't seem an inappropriate outcome!)?
Does the Eloelle narration "break the game", or give a player an undue advantage? I can't see that either. The decision about whether to permit it or not strikes me as overwhelmingly an aethetic one. The ZoT problem is obviously an edge case; and there might be others. But if a group is into this sort of non-standard narrations, I think the edge cases are hardly going to be more than modest speed humps. They're not going to derail anything.
Not everything that falls exclusively under the rubric of "DM judgment call" is a house rule, right?
I want to say
obviously not!
RAW has only one possible definition. Rules as WRITTEN. If it's not written, it's not RAW. Call it a ruling, house rule, or whatever, but it's not RAW for a fireball to set things on fire.
The problem with this is that no set of rules covers all the cases it has to when taken strictly literally, with no entailment or extrapolation permitted.
Some of the entailments are obvious and uncontentious (eg the rules tell us that recovering 1 hp takes 1 day of rest; and so we extrapolate that if my PC is 7 hp down that wil take a week of rest to recover).
Some of the entailments are more contentious. For instance, the rules tell us that alchemist's fire, burning oil and a lighted torch all do
fire damage (SRD pp 66, 68), and also tell us (under the heading "Damage Types -") that "Red dragons breathe fire, and many spells conjure flames to deal fire damage" (SRD p 97). This seems to me to support extrapolation to "fire damage is the result of being burned by flames". It is pretty uncontentious that flames can set timber structures alight. Hence, I see an extrapolation to "fire damage can set timber structures alight". The extrapolations here are weaker than strict entailment, but they're much stronger than mere conjecture, or mere permissible selection from a range of feasible alternatives.
Is that a departure from RAW, or a "ruling", in the way that the
1 week's healing to recover 7 hp is not? Again, until someone tells me what's at stake in the distinction, I'm not going to express a view.
They are not examples at all. The rules do not call them illustrations or examples and use the following language.
Strength measures... Intelligence measures... Dexterity measures... There is no "might measure", or "measures this and other things", or "Sometimes measure", or any other language that is not absolute. Strength measures X, period.
They're instances of the things that the stats might measure. Nothing suggests that they are exhaustive. They're barely even
canonical, given that the items on the list changes from occurrence to occurrence. Just sticking to INT, the SRD on p 76 tells us that INT "measur[es] reasoning and memory", and then on p 81 tells us that INT "measures mental acuity, accuracy of recall, and the ability to reason", and then on the same page tells us that "An Intelligence check comes into play when you need to draw on logic, education, memory, or deductive reasoning."
What is the relationship between
mental acuity, which appears on only one of those lists, and the other components of INT? What about inductive or scientific reasoning, which isn't mentioned at all? Is that included under
reasoning ability? But then why specifically call out
logic and
deductive reasoning?
As I said, these formulations aren't even canonical, let alone exhaustive.