In this paper, we describe TAKE_ONE, a parallel method of specifying human motion animation by a controlled mixture of values from three kinds of simulation: kinematic, dynamic, and constraint. In addition, tools to assist an animator to define qualities such as realism, individuality and expressiveness are developed. The issues in comprehensive animation methods are explosion of complexity, difficulty in determining values of input parameters, and lack of ease in fine-tuning an animation. We discuss the advantages and issues involved in structuring the specification of an animation. We provide a structured method to use or to convert to, a parameterized motion definition. We introduce a method of specification to allow the development of the essential and minimal definitional qualities of an action. The result is that a reduction in run-time complexity and user-specification is effected and groundwork for an action database is done. We develop techniques to refine kinematic animation specification so that it is more representative of actual positional goals and so that it is compatible with the use of the other methods. We provide a structure to systematically merge animations from the three methods, through user or program control, and provide an interface to an iterative method of definition and fine-tuning. Examples are provided to show the power of the TAKE_ONE method, including: an object placement example whose implementation is explained in detail, a wheel-turning task, and finally, a classical ballet pirouette which will serve as a goal example for our completed work.