Most people believe that it is permissible for an 
agent to kill an innocent threat—that is, someone 
who threatens his life without exercising agency.  
Call this the “commonsense view.”  I offer two 
arguments against this view.  First, I show that 
despite their efforts, recent defenders of the 
commonsense view have been unable to account for 
the wrongness of killing certain bystanders.  In 
making this argument, I criticize the claim that 
one who kills a bystander, unlike one who kills a 
threat, uses another person as a mere means.  
Second, I argue that the commonsense view faces a 
dilemma in cases involving stationary innocent 
threats: it either draws arbitrary distinctions or 
else yields implications that its proponents would 
disavow.  If I am correct, innocent threats pose no 
problem for theories that tie an agent’s defensive 
privileges to the other party’s culpability.