Normally if a station receives a corrupted PLCP header, the corruption is apparent and the station simply resumes looking for preamble patterns among any arriving RF energy.
If a station receives a corrupted PLCP header that by chance looks correct, the receiving station is required to honor the length field in the header and defer from transmitting on the channel for that time regardless of the success or failure of receiving the PSDU promised by the PLCP header.
If the transmitter actually transmitted a shorter frame than what is stated in the corrupted but apparently correct PLCP header, the transmitter could then retransmit fruitlessly while the deceived receiver was unnecessarily in deferral.
If while deferring the receiver detected a subsequent preamble arriving, I want to think that the deferral would be cut short, but I do not find this in the standard.
I hope this helps. Thanks. /criss