Hi Jim, thanks for your questions concerning my grammar!

> What specific phenomena in colloquial Esperanto are you talking about
here?  From your later message, I infer that using bare roots as
derived prepositions is one of them; I'm not sure what else you're
referring to.  (Those bare derived prepositions are pretty rare in my
experience, but my experience is limited.)

There are several indeed, and it's a shame they're probably completely unknown to the linguistic community. 

To begin with the described preposition method is sometimes attributed to William Auld who established 'far' (made by). Such prepositions are limited in use not only because they are 'anti-Fundamento' but also because there's a limit to how many prepositions are needed in a language. My theoretical point is that there could logically be an infinite number, but the ecology of language would keep them to a minimum.

Some of the most well documented pheneomena are the so-called 'tiom kiom-ismo','kia'-ismo' and 'senpera verbigo de adjektivoj' that de Hoog published books about. The first two deal with the logic of comparison; tiom kiom –ismo is quite common today although the Akademio tried to eradicate it in the 1950s and 1960s.

The most heated grammatical schism, itismo vs atismo occurred in the 1930s. I believe particularly the speakers of Slavic languages noticed a logical fault in the participle system of Esperanto; there's a brief explanation here Although the debate can be said to have ended half a century ago, there's still common disagreement about the usage of participled words, such as "Fumado estas malpermesITA" vs "Fumado estas malpermesATA" (Smoking is forbidden).

One later development is the use of the accusative preposition 'na' instead of the object ending –n. This handy with names, for instance. Here's an article

The Academy of Esperanto normally condemns all new 'isms' of which there have been many. This has been a central strategy since the Fundamento that was written and published in 1905 to solidify Esperanto. 

But there was a certain political development that caused one innovation to be accepted. The President of the Academy in the 1930s to 1950s was heavily criticised by the community after his attack against tiom-kiom-ismo. Self-made grammatician De Hoog helped him win the dispute with his book, and was allowed to publish another book praising the usage of stative verbs (such as 'beli' - to appear beautiful) which was consequently accepted to Fundamento although Zamenhof had specifically spoken against it (this was ruled as a misunderstanding).

Anyway, I'm definitely not a 'reformisto'. Although my study takes inspiration from the living language community of Esperanto, my grammar is not Esperanto; it's a theoretical approach that I hope could be used use as a tool for linguistics.

> Also, in your later message outlining your Esperanto-based engelang,
you say you're using the x-convention to represent circumflexed
letters, but you also have a verb ending "-ax"; does that represent
"â", and if so how is it pronounced?

Well, the thing is it may or may not represent "â", and it could be pronounced anyhow. I don't really have a fully fledged engelang, only a grammar sketch. I made up the ax-ending quickly when I was writing my message; the point is that ideally it shouldn't collide with anything that's already in use, but keeping that in mind, all endings can (will?) be changed.

> There are several other engelangs and schematic artlangs that have a
way to derive adpositions from open-class root words; my gjâ-zym-byn,
Rick Harrison's Vorlin, the collaborative project Voksigid, and a
number of others that don't come to mind at the moment.  

Thanks, I'll check them out :) I noticed gjâ-zym-byn is in Arika Okrent's list of 500 conlangs. I'm interested in what you say about "other schematic artlangs" - are they completely systematic?

> It's a useful
feature for making a flexible, human-speakable language but I'm not
sure it really produces a simpler grammar than the phonologization of
predicate logic that Logan and And are recommending.

Well, obviously that specific feature doesn't produce a (simpler) grammar. As you can see it's only one feature, and it didn't affect the number of rules. Do the languages you mention use the same formula and root words to create conjunctions?

At this point I don't know precisely how the demarkation problem between formal and human processable notation is going to be solved, but there's theory about narrow syntax etc. so I'll have a look at that. Anyway, as I mentioned earlier, numbering shouldn't be used (I could add bracketing).

I wish there were a lot more journal articles written on Lojban, but I think there are enough for references. When we're looking at so called logical languages that are both people and AI usable it still seems that an optimisation based on classical linguistics produces a grammar that is easier and simpler for both people and AI. The currently available loglans have a theoretically infinite set of grammar rules. This means that even the most complex finite grammar is simpler. How would you argue against that?

I've been thinking the simplest human processable grammar could be something like a "police report syntax": Offence: theft; Perpetrator: Ms Virtanen; Target: handbag; Victim: Vuitton Shop; Time: yesterday at 16:28; Wittnesses: shop-keeper, customer. Next page...

This is very precise and easy to understand even if it didn't make a great detective novel. But this approach needs grammatical rules too, and when the expressions get more complicated, it seems the syntax could become too difficult to follow.