The title is nonsense. PATH is the name of an environment variable (a Real Thing(TM)) which lists a set of directories to search for an executable. It is used by shells (including those running on Linux) to locate an executable when the full path to the executable is not supplied by the user.
This is needed because the exece()/execve() [2] kernel system call is unaware of things like environment variables so it will not have any idea how or where to execute a program 'cat' unless it is given the full path to 'cat', so the shell has to look it up (again if the user doesn't pass the full path). It's the same on every POSIX system and the original UNIXes. It's been this way for at least 50 years. (edit 60 years, it's from Multics [1])
Kids today really need to learn the fundamentals of computer operating systems. Or do that boring old-person thing we did before StackOverflow, and read all the manual pages, which tell you all this [3] [4].
The fact that the Linux kernel does not track environment variables of the processes is not a "fundamental". The setenv/getenv could very well have been syscalls, it's simply a design decision that they are not. One can make a kernel with such tracking, and it'd still be POSIX compliant as long as you supply setenv(3)/getenv(3) wrappers with expected signatures in your system libc.
It's fine to understand what the code is doing in a shallow way. But this leaves out a lot of important information. Information that isn't immediately obvious just by reading code, that can help you avoid problems and understand the system in-depth, without years of trial and error. Which is why they wrote a manual. You can also both read the code and the manual, but the manual will give you much more knowledge in a smaller amount of time.