Dissonance and conflict

 From David Schum, The Evidential Foundations of Probabilistic 
Reasoning, pages 121-123:

The different conceptions of evidential force, strength, or weight we 
consider... all suppose that relevant evidence has vectorlike 
properties.  Evidence points in one direction or another and with a 
certain gradation of force. The direction in which an item of 
evidence points refers to the hypothesis we believe it favors over 
other hypotheses being considered.  When we have a body of evidence 
to consider, we may observe that some evidence favors one hypothesis 
and another evidence favors another. In such cases, we may say that 
our evidence is dissonant to some degree.

...contradictory evidence... one person or sensor reports... that 
event E occurred, and another person or sensor reports... that this 
same event E did not occur... we naturally look to the credibility of 
our sources... clearly one of the sources or sensors is wrong.

...conflicting evidence... a person or sensor reports the occurrence 
of event F, which we believe to favor hypothesis H.  Then another 
person or sensor reports the occurrence of event G, which we believe 
to favor hypothesis [not-H]... What separates conflicting and 
contradictory evidence is the fact that the events reported in the 
conflicting case may both have happened... events reported in 
contradictory evidence cannot have occurred jointly.