Term for the "extreme-extension" version of a straw man fallacy?
Solution 1:
An appeal to extremes is an often fallacious application of reductio ad absurdum where one takes an argument to an extreme and neglects the actual circumstances or implications of the initial statement. As the website Logically Fallacious describes it:
If X is true, then Y must also be true (where Y is the extreme of X).
There is no way those Girl Scouts could have sold all those cases of cookies in one hour. If they did, they would have to make $500 in one hour, which, based on an 8 hour day is over a million dollars a year. That is more than most lawyers, doctors, and successful business people make!
As the site then points out, the extreme version of this neglects that (a) Girl Scouts don't actually work 8 hours a day over the course of a year, and (b) the output of several Girl Scouts (not one) in a temporary operation actually could sell this many cookies.
The appeal to extremes relies on hyperbole or exaggeration to the exclusion of other logical constraints. In contrast, a reductio ad absurdum is valid when the circumstances and context are not exaggerated but nonetheless the original statement would lead to an absurd conclusion. An example from Wikipedia:
There is no smallest positive rational number because, if there were, then it could be divided by two to get a smaller one.
That statement uses the definition of what a rational number is (its ability to be divided by an integer) to critique the idea that a smallest positive rational number exists.
Solution 2:
You may be talking about an "Appeal to Extremes"
Description: Erroneously attempting to make a reasonable argument into an absurd one, by taking the argument to the extremes.
https://www.logicallyfallacious.com/tools/lp/Bo/LogicalFallacies/30/Appeal_to_Extremes
Solution 3:
Your example is similar to an appeal to extremes. But the problem may be more one of vagueness in the original statement.
If someone says "Drinking water is good for humans." and we are discussing formal logic or we are in a situation where we must cover all cases, I could reasonably argue that the statement is neither true nor false because it is too vague. I could properly illustrate this by providing an extreme. In fact, it can be refuted to a degree with an example far less extreme than you provide. It is hardly unheard of for Marathon runners to become sick, or even die, from hyponatremia..
Drinking water is not always good for humans. Flatly stating that your original statement was false because it is not true in all cases would be engaging in a straw man argument, which is a fallacy. But the example is ample to show that your statement is too imprecise to have a proper truth value in formal logic.
Of course, in casual conversation most people would interpret your statement to mean "Drinking water is generally good for humans" or something similar. But if we are trying to apply formal logic or are in engaged in some other discussion in which the existence of edge cases is important then responding by pointing out those edge cases is non-fallacious and proper.
In fact, as a practical matter knowing that the statement has limits may be significant. I don't have the reference handy, but I recently listened to a podcast where advice on avoiding hyponatremia during endurance sports was a major portion of the show and it was pointed out that for an average runner over-drinking was more dangerous and more likely than becoming dehydrated during a marathon...
Solution 4:
Consider reductio ad absurdum:
A mode of argumentation or a form of argument in which a proposition is disproven by following its implications logically to an absurd conclusion. - Logically Fallacious
This is talking about taking a position to its logical ends. When paired with valid reasoning, it can be used in rigorous mathematical proofs.
However, when used merely as a rhetorical device and when paired with unsound logic, it can be abused to take a position that is itself sound (like it is good to drink water) to an unreasonable end (like see what happens when you drink it to excess).
Solution 5:
This is an example of a motte-and-bailey fallacy (although, inversed from its typical use).
This fallacy states that a person will use a motte (an easy to defend position), as a proxy for a bailey (a hard to defend position), implicitly trying to draw an equivalence between the two.
At the same time, another way to approach this would be to argue that it is not a fallacy. It is true that too much water can kill you. In the absence of a further claim, this is both true and irrelevant to the discussion. Therefore it is less a case of being fallacious, and more a case of trying to employ what Thouless might call 'deceptive argumentation techniques' (in this case, trying to shift the conversation to a different topic).
So, you could easily continue on from your original position dismissing that claim.:
- Water is good for you
- Ah, but too much water is bad for you.
- Yes, this is true. Tautologically so, in fact, since that is the definition of too much, but this is irrelevant here. Anyway, since water is good for you, I propose that [...]
Looking at the other answers, I would also add that I do not consider this an 'appeal to the extreme' fallacy. The key word here is the word too much. As long as the (unstated assumption) is understood by your opponent, that is to say, "excluding extreme circumstances", then making a claim about those extreme circumstances is valid, but irrelevant. By its very construction, your opponents argument is effectively saying "Your argument is false in the special circumstance where it is false". You can accept this at face value, without having to adjust your position that your argument is valid outside that special circumstance which was not part of the conversation.
In my mind, for something to be an 'appeal to extremes' fallacy, the extreme position should still be valid in the domain of discourse. Naturally this comes down to the unstated assumptions. These may need to be further defined for fruitful conversation when it comes to blurred topics, like ethics, but in this instance it is clearly outside of the domain.