The theory presented here has implications for philosophy in respect of how concepts and words can be mechanically defined. For neuroscience the paper at least sets out a problem that has received little consideration and offers a possible solution. Also the theory may be relevant to robotics in terms of object manipulation. Concepts need to be separated from each other in the brain in order for an animal to act on one object in isolation. A possible solution is to inhibit (...) common feature neurons that are shared by two concepts. There are some problems with doing so, including the fact that each feature of every concept is shared with some others so all would have to be inhibited. This can be resolved if common feature neurons are only switched off while two concepts are active at the same time. Another problem is that while inhibiting the common features, the two active concepts cannot be related to each other and understood in each other's context. This can be addressed by division of the brain into relational and isolational areas. Associative thinking, proposed to be the only memory mechanism in the right hemisphere, produces poorly defined but well related concepts. In fact all concepts have associations to others and so in a purely associative brain region, everything forms one giant concept. In the left hemisphere it is speculated that differential or oppositional thinking cancels out the common features of live concepts, leaving only the differences, resulting in clear definition but lack of relationships to other concepts. (shrink)