Actually, I like the three-body problem in regards to the confusion here.
If you examine each member of a three body system individually, without regard to the presence of the other two, then yes, it won't work. However, this is not reductionism (except in some people's absurb sense); this is ignoring the problem. The properties of the three-body interaction are still derived from the basic case. What this means is that the gravitational constant doesn't suddenly become 124.3 if three bodys interact...the basic equations are the same. Gravity doesn't suddenyl change to follow a inverse cube law, it's still square. Etc, etc, etc. The first principles from individual elements and properties still apply to the larger grouping, and in fact these elements interacting lead to the behavior of the larger grouping.
Protein folding is another good example. No, you can't look at a single molecule and descibe various protein interactions. That's not reductionist, that's ignorance and self-induced blindness. However, the compelx actions of the proteins all rely upon the interacting properties of the amino acids involved and environment of the protien, which in turn rely upon various molecular interactions and laws for their structure and properties, which in turn come fromt he atomic interactions and principles, which in turn come from quantum effects and micro-physics. That's all reductionsism is...it is not the idea that one can work out the properties of the entire universe by looking at a single electron, but that a single electron follows specific laws of nature whether floating by itself, orbiting a nucleus, involved in a plasma cloud, beaming through a TV tube, or being compressed into a neutron star. The individual elements interact to produce effects on a larger scale, which then interact to produce larger effects, which then interact to produce larger effects, etc.
Using Ian's ridiculous definitions, the property of distance implies dualism and is non-reduceable. Looking at a single element, concepts of distance and position are unknown and, in fact, unknowable. However, the properties of an object do not change radically because another object exists; the laws for the single object apply equally to the two, although new behavior may emerge by the interaction fo these objects.