Spurious wake up??

From: Min,Lee (abraxsus_at_yonsei.ac.kr.NOSPAM)
Date: 03/31/04 

Date: Thu, 1 Apr 2004 00:20:55 +0900

Hi, folks.
I'm reading Robert Love's book "Linux Kernel development".
On the page 45, he said..

"there are spurious wake ups. Just because a task is woken up
does not mean the event it is waiting for has occurred; sleeping
should always be handled in a loop that ensures the condition
the task is waiting for has indeed occurred."

Could you explain in detail about this "spurious wake up??"
I don't think it means "a signal". then, what else can wake up the
sleeping kernel thread??

TIA