There are a lot of metaphors used to talk about op amps, and none of them should be taken literally, and rarely are they meant to be used in wider contexts.
The metaphor of the virtual short between the inverting and non-inverting pins is just that - a metaphor. That's your plausible reason. The pins aren't actually shorted, so this idea does not carry over to positive-feedback situations.
Another metaphor often used, which is in my opinion slightly better (as talking about things in terms of interconnections like shorts could be mistaken to mean there is actually such a connection, and that it is present in other situations), is that an op amp with negative feed back is 'trying to keep the two input terminals the same'.
In fact, I am just going to say that the virtual short metaphor is flat out incorrect. This makes it seem like a passive effect, and one that would carry over to other situations, when neither is the case. Op amps are still just amplifiers, and all they really do is take the difference of potential between the input terminals and amplifies it to a larger one at the output.
With negative feedback, this amplified difference of potential has the effect of reducing the very input signal being amplified, or rather, the difference between the input terminals, so when the gain is very high, this means the op amp is actively reducing the difference between its two inputs to almost nothing. There is no virtual short, we are just using the amplified output to effect the input, and negative feedback always helps reduce the input.
In positive feedback, the amplified output is going to increase the difference of potential at the input, which makes the output even larger, which makes the input even larger, and in a magic ideal op amp, everything diverges to \$\infty\$ .