Maybe it is just as difficult, and that's precisely the problem. Just because some people can learn a particular set of rules and become proficient with them doesn't mean they constitute a good system; consider the Roman numerals, for example.
For the most part I have become sufficiently accustomed to Perl's context scheme that I am not particularly inconvenienced by it, but I know many who struggle with it, and I can't help but notice the irony.
Esprit de l'escalier: Another objection to the subject-verb agreement errors analogy is that such errors are benign, in the sense that they rarely, if ever, lead to serious misunderstandings (if they did they would have been weeded out of the language long before the appearance of the first grammarians). In contrast, context-related errors in Perl are not benign.