The assumption we submit, because macroeconomic forcasts would be unperfect, is that behavioral equations doesn't enough describe economic behaviours through the capacity of reaction opposite to environment. Further, the forcaster belongs to his search-system, so that, may be, we must now integrate forcasting way in the system. How to proceed ? We examined microeconomics which is advanced to this purpose, and social sciences which analyses behaviors. Hence the new formulation of behavior equations would have intentional and potential components, which we have now to specify.