It's an adage/heuristic not a hard and fast rule of logic... So what do you expect?
Like the point of a heuristic is to give you reasonably good odds in a situation where you lack the time, information or resources to make an informed decision, so instead you take the gamble by going with statistics, "folk wisdoms" and "experiences".
Now the general problem with evidence is, that apart from very abstract concepts confined to special conditions, you'll hardly ever have enough evidence to confirm "the truth" for just about ANYTHING. Science even goes so far to abandon the search for "the truth", but rather tries to model nature and is content if the actual values are in a small margin of error around the predicted value (instead of being the predicted "true value").
So for example to check if Adam watered the plants, you could:
- have a look at whether the plants are still alive
- whether the soil is still wet
- whether it rained and they were accidentally watered rather than by Adam
- you could check DNA samples of Adam to see if he touched the vessel
- you can take water samples to check if it's tap water or rain water
- ...
Seriously you could apply ridiculous levels of rigor, but likely the 1st bullet point is all you care about. So that would already suffice as evidence.
So if you applied the same level of rigor to all claims, big or small you'd waste tons of resources most of the time and still end up with insufficient evidence where it actually matters.
So the heuristic of extraordinarity urges you to prioritize. Which most of the time is a good advice, like the extraordinary claims are often those that have major consequences which prompt you to action, so checking on whether that action is warranted or if someone tries to manipulate you is usually reasonable.
Now of course there are edge cases where the devil is in minute little details or where you are on the brink of change, where something extraordinary is about to become ordinary and vice versa. Like if you jump from a plane without a parachute, then for 99.999..% of the fall you'll be fine, it's that that edge case of hitting the ground that kills you. So there's a reason you should only go with prejudices if you have no other option, instead of always relying on them, because heuristics ought to work for a majority or plurality of cases they are NOT expected to hold all the time.
Also if someone put a gun to your head, then it's no longer about Adam or his plants or if he watered them. It's not even about the truth. It's about the gun to your head and what is more or less likely to make the gunman pull the trigger. Like even if you told the gunman the truth, the full truth and nothing but the truth and provided a mountain of evidence they might nevertheless call you a liar and shoot you because "their truth" doesn't match with "your truth" and what they were interested in might not have been "the truth" (whatever that is) but confirmation.
This example, to me, seems to highlight that testimony is still not enough evidence even for an ordinary claim, but rather that you have nothing to lose by believing in it. Would this imply that the extraordinary claims being the only claims that require extraordinary evidence is wrong?
Yeah exactly it's less about how much evidence is enough evidence (to establish the truth) and more concerned with the question of "does the truth matter". And when it matters you want to investigate while in other situations you might comfortably live with the ambiguity of being wrong. No it's not a hard and fast rule, it's more of a rule of thumb that expects the status quo to last (ordinary) and thus demands evidence for change (extraordinary).
Which is often enough a good approximation as change usually happens gradual which is why we have words like "everyday life", while sudden rapid changes are usually scarce. So if you were to roll a die on the question of whether the next day will be an "everyday"-day or an "extraordinary"-day you'd be right more often if you'd go with the everyday. But you'd also be caught completely off guard on all days where it isn't...
