Re: [lojban] Lemma and conjecture

[Pierre Abbat <phma@webjockey.net>]

On Saturday 21 December 2002 19:57, Jorge Llambias wrote:
> Assume that the conjecture is false.
>
> Then we must have two lerpoi: S1=CVVRC... and S2=CVVKC...
> where R is one of r,n,l and K is some consonant other
> than R and the two lerpoi have the same pattern of
> vowels, consonants and clusters, and such that S1 is
> a lujvo and S2 is not a valid word. (In no other position
> could a difference in consonants have an effect on the
> validity of the lerpoi as a brivla.)
>
> The initial CVV can't fall off from S2 because KC must
> be an impermissible initial, to match the form of S1.
> It is clearly not a possible slinku'i either. It is a
> valid brivla as far as permissible clusters because S1
> is one, and it is not a lujvo because it would have to break
> as CVV-KC... but the second part can't be a lujvo. So it
> is a valid fu'ivla. So our assumption is wrong and the
> conjecture must be true.
>
> Or am I missing something?

You still have to prove that, if S1 is a lujvo and S2 a fu'ivla, there are no 
S3 related to S1 and S4 related in the same way to S2 such that one is valid 
and the other is invalid.

phma