Earlier quoted context omitted.
There's a UK comedy show from a few years ago called Little Britain that has this same thing happening in a recurring skit except with English. There was a British woman who claimed to never understand what an old Indian woman wearing a sari with an Indian accent was saying despite it being quite clear. I wonder whether this is some low level passive aggressive behaviour around the world due annoyance of dealing with…
The difference in Japan is you can have zero accent and speak fluently and just by looks some Japanese will have some kind of mental block understanding and/or acknowledging your Japanese. Examples: I can call some business on the phone and people won't know I'm a foreigner. No accent. I've gotten in a cab late at night, talked to the driver for 10 minutes, and then had him freak out went he turned around to collect…
I think it's because languages are all about expectations. If some piece of information has primed me to hear one version of a homophone the other meanings never even pop into my head.
I've noticed a related effect while watching videos on youtube recently. I'm a native British English speaker. When I click on a random youtube clip I don't know what accent I'm going to get. If it's a strong accent I don't hear often it usually takes 5 seconds or so for my brain to 'work out' the accent. During that time I don't really understand what is being said. As soon as it clicks I can rewind the clip and hear the exact same audio with perfect understanding.
I think when the Japanese people don't understand me a similar thing is happening. They see my white face so are trying very hard to hear English. The sounds come out and are run through the 'process the English' part of their brain. Of course, they can't make heads or tails of it so end up not understanding anything.
I'm not a linguist but I'd love to know if anyone had any references on this effect.