>Has anyone here done or seen work on a language relying on a grammar  
>that does not require the noun/verb distinction? 

I was faced with this issue when I tried to reinvent grammar from scratch.
Before I would allow word types in the grammar I had to have it justified
in some logical scheme. 

Are verbs and nouns required?
Is the distinction required?

I discovered in the analysis, that most definitely, verbs and nouns must 
and are distinct. HOWEVER, I found that there were two different issues.
This is obfuscated by the fact that the words "verb" and "noun" in English
have multiple meanings. In English "verb" means a word which semantically
indicates an action (something that would be recognized in any culture and
by speakers of any language). It also means a word which has a specific 
in a sentence. It turns out, of course, that both of these things are
required universally. That is, a language needs a way to express actions and
it also needs a way to express the verbal function in a sentence. (Consider,
for example, in Loglan where a "function" and an "argument" are required.)
However, I realized that these two requirements were independent of each 
The reason that so many languages utilize the verb-noun model was because
of a natural convenience. It is convenient for the action word to play the
role of a "function" and a noun to play the role of an "argument". But even
natural languages realize that this relation is not absolute as they employ
verbal nouns, thus using action words as nouns.

But here is the real mind-boggler!

The verbal noun is NOT just the action in the role of a noun. It also 
the verbal function as part of the noun. That is, it is a noun that is
actually referencing a sentence* (a sentence with just one word, the verb).

This was a real difficulty and it took me some time to figure out that
the action noun without the verbal function was merely the agent of the 
This solution was verified in Ankanian** when consistent conversions 
between nouns
and verbs in either direction were re-entrant. That is, if you convert a 
to a verb and then you convert the verb into a noun, you end up with the
same form and the same meaning. This would not work if the conversion of the
verb to the noun resulted in a verbal noun/infinitive (e.g. Esperanto).

>As I understand the matter, many  
>Native American languages modify their “action-words” (I decline to  
>call them verbs) in terms of the degree of validity with which the  
>speaker speaks.

I also thought of employing this, but decided it would add unnecessary 
since this validity can be shown by other means as it is done in the 
of languages, and since one of the goals of Ankanian was to be simple.
Even though ankanian was essentially an independent construction, general
practises of natural languages were taken into consideration, the theory 
that, if certain grammatical categories were present in a majority of
languages, it might indicate a natural inclination of human thinking.
This is why Ankanian has a basic plural form, even though, as some 
demonstrate, it is not technically required.***


*Seeing this was difficult, since some English sentences seem to distinguish
between the "fact" of the sentence and the verbal noun. But it seemed that 
were actually artificial differences caused by English structure.


***Some might argue that I disobeyed the language parameters, since Ankanian
has some grammatical elements that are unique. But these actually ended up 
required to avoid other problems. For example, the "preexpectant"
verbal form came about because I could not reduce the English 
adverb "already"
to a derivative of a basic verb/noun type. This happened with a few adverbs,
but not enough to require the creation of a whole new class of words.