Life can be defined as the search for a balance between specialization and generalization. We can understand easy that life is a result of a specialization process. The cells of the human body are an example of it, the different species in nature too. Specialization tends to increase the effectiveness of the tasks that support life. But, sometimes we forget something important: Specialization without coordination is chaotic.
In mathematical terms, when we are analyzing a system, specialization means higher dimension of the state space and coordination means mathematical relationships among the state variables that limit the possible states. In the physical world, the dimension of the state space and the relations among the state variables are defined by physical laws.
If life is a process of specialization, we can consider it as a natural mechanism that is searching for a higher number of dimensions, but the process of evolution is related to a process of cohesion. Namely, there is evolution neither without diversity nor without cohesion.
Why is it necessary a specialization process? The reason is adaptability. Think in an industrial robot. A robot with higher number of free degrees is adaptable to a higher number of situations and it can perform more tasks. The higher the number of free degrees is, the higher the possible states of pose (position and orientation). The possible states are limited by the physical structure of the robot providing cohesion, the additional stability, required to preserve a certain pose, is provided by the control devices of the engines in every joint. Robots as live beings are the result of an evolution process. The design of the next generation usually proceeds from the previous one. A strong innovation can be seen as a mutation process.
A free system in any environment requires these three properties: adaptability, cohesion and stability.
There is a different process that can be seen in nature to acquire more adaptability: grouping. Simple organisms can cooperate to perform complex tasks. This is very common with insects. A single bee is not as intelligent as a man, but a set of them can act more effectively to preserve the life making a lot of task with a high degree of coordination. They can perform some tasks as search for flowers showing a certain degree of strategic behavior, in a similar way as a single neuron cannot make complex decisions but a human brain can. The difference here is the degree of cohesion and stability of both systems. If the individuals can communicate, then, there is some cohesion, but different communication systems have different level of noise.
Where is the limit of the specialization? There are some physical limits to the evolution of a system. One of them is the energy required to maintain the system, a robot with two joints only needs that two engines are fed. The consumed energy grows with the dimension of the state space. If the environment cannot provide infinite energy, efficiency instead of effectiveness will drive the evolution process.
A way to find efficiency is generalization. The process of generalization is related to the acquisition of information about the environment and response. It can be seen as follows: Many situations do not require a precision of 100 %, then, if we reduce the possible states we can reach a useful pose nearby the target. For instance, survival is got finding a proper food, and it is not necessary to find the optimal one. An animal can survive without analyzing where the best place for hunting is, it only needs to know that he is hungry and that something is moving nearby. Human beings have overtaken this behavior searching for effectiveness and efficiency simultaneously. Their adaptability is huge and they have needed to take advantage of it in order to increase the possibilities for survival.
It is not true that effectiveness is equal to specialization and efficiency to generalization. Specialization and generalization are intrinsic processes of a live system but effectiveness and efficiency depend on the environment.