What are some suboptimal design decisions of Esperanto?

I really don’t like this kind of question. As many on Quora and outside Quora have said, optimal design is not how a language prevails; language typology is not about better and worse but different; and the defensive Esperanto assertion that Esperanto has an internal cohesion is quite true.

*sigh*

That said, if there is a point in debating Esperanto design (which I dispute), the most questionable features are:

• The diacritics. It’s a solution that makes sense if you live in Poland, but it’s been a burden.
• The accusative. I *love* the free word order of Greek, and I love that it shows up in Esperanto. But it’s an unnecessary burden for learners of caseless languages.
• The neurotic avoidance of polysemy, which has led to some quite artificial differentiations of stems (sento/senso/senco from the beginning, but it got much worse).
• Inventing aspect distinctions, and not having the linguistic sophistication to realise it and document it. Three decades of argument between the atistoj and the itistoj could have been avoided.

I’m not perturbed by definiteness or number-marking, or the deprecation of ci “thou” (which happened at the last minute before publication). I’m very happy Esperanto didn’t go the way of Occidental, and prioritised schematism over naturalness: if I want Italian, I know where to find it.

There’s been much moaning over the mal– words; Orwell lampooned them in Newspeak; Ido got rid of them; and the dirty secret of the “poetic” replacement words in Esperanto is that they are Ido words, introduced by former Idists. They don’t bother me either: once you learn them, you internalise them.

And Zamenhof dealt as well as anyone could with the deluge of reform proposals. Much of this had already been brought up in 1894.