I looked at the Wikipedia page for Ithkuil. It doesn't seem to be geared towards preventing cognitive errors, so much as packing as much information as possible into as few phonemes as possible. For most of them I can't see the point. In English I can say "Trees are green." in a few simple words. From the sound of it, in Ithkuil I'd have to pack in so much information about the trees that it would take me an hour to figure out how to write the sentence. Is the set of trees spatially contiguous, in a specific but unnamed forest? Or is this the set of all trees on earth, being denoted as members of an abstractly defined set? And how do I feel about the matter?

Also, the creator packed the phoneme set so tightly that I can't see how he's going to avoid a high rate of transmission errors. There comes a point where you've got so many vowel sounds that individual sounds are so close together in phonetic space that you can't reliably distinguish between them.

So far I've been going in the opposite direction. Rather than including multiple layers of meaning in each word by complicating the sounds and grammar, I've been planning to restrict meaning to one meaning per word, with each additional bit of information requiring more text to transmit, and no restrictions on what the user can leave out because it's irrelevant. It looks to me like Ithkuil goes in the opposite direction. Which isn't to say that there isn't some interesting material there. The creator seems to have broken down his informational overlays into unusual categories, such as "configuration", "affiliation", "perspective", and so forth. It would be interesting to read about why he selected the categories that he did. So there might be some interesting stuff there to borrow. But it seems that he gave up too much in the way of usability. What good is a language that decreases the frequency of common cognitive errors, if the only people who can use it are already so smart that they rarely make such errors?