I appreciate the answer thus far--I've learned a lot about MATLAB exception handling in the process. To illustrate the problem, suppose you have a file reader of some kind that is called by an analysis routine that is launched by a graphical interface callback.
GUI > analysis > reader
For modularity, I would normally create each of these function independently so they can easily be transferred to other programs (especially the reader).
Let's say that for whatever reason, the user provides an invalid file. Maybe it's the wrong format, or perhaps there is simply something wrong that can't be found until the reader begins to scan the file. Good practice dictates that the reader function should throw an error to let the user know something is wrong, but end users don't need to know where in the code the error was thrown. That information can be misleading (the code isn't faulty, the user made a mistake), frustrating (the user may have p-code, so they can't view the source anyway), or dangerous (the user may try to make changes to "fix" the problem).
MATLAB has ways to add information to the exception, but I haven't found a way to reduce the output to only the developed-provided message. I like the ability to identify the exception for use with lasterr, but something like a "stop" command would be handy. If can break out of a particular function using a return statement in the catch block, but that will lead to more errors in the calling function(s).