Subject: How do we know light not transmitted in a spectrophotometer is absorbed?

To calculate the light absorbance of a specimen in a 
spectrophotometer, we take the negative log of the transmittance.  The 
assumption seems to be that light not transmitted through the specimen 
(from the light source on one side of the specimen to the 
photoreceptor on the other) is absorbed by the specimen.  But how do 
we know that some or all of that light is not being REFLECTED rather 
than ABSORBED by the specimen?