Why not just replace a coordinate by its adjacent force?
We will have a closer look on a 1D example of a function
.
Let
|
| (1.19) |
and
|
| (1.20) |
We find
|
| (1.21) |
consequently
|
| (1.22) |
respectively
|
| (1.23) |
This differential equation has the solution ; but there exist more solutions
|
| (1.24) |
Simply substituting a coordinate by the adjacent force leads to a loss of information, since the original function is known only up to a constant. For a function with several parameters this is even more critical since the ”constant” may by a function of all the remaining parameters.