Mathematica use MachinePrecision as 16 digits. I know that we can specify the WorkingPrecision and PrecisionGoal in many of the functions. However, mostly it will generate a warning message like "...the argument used has a precision lower than the WorkingPrecision...". It is reasonable because if we don't use an argument having a precision higher than the working precision, the high working precision is nonsense. It is especially important if we are applying some system defined functions to some self-defined functions. So, my question is, can we specify the precision of any calculations from the very begin, so that all the calculations following on will have at least that precision?