reduced logical form Step 3

statement = statement-1 | prenex statement

statement-1 = statement-2 [[I joik-jek [statement-2]] ...

statement-2 = statement-3 [[I [jek joik|jek joik]] stag BO # statement-2

statement-3 = sentence | tag TUhE # text-1 /TUhU#/

subsentence = sentence | prenex subsentence

text-1 has already been reduced in Step 1 to paragraphs.

If there is a paragraphs, proceed to Step 4 and reduce it to statement.

Then:

tag TUhE # statement /TUhU#/: is reduced to "tag ku zo'u statement"

which is a statement.

Now, starting from the innermost statement-2, and using the appropriate

gek= SE GA NAI # | joik GI # | stag gik:

sentence [[I [jek joik|jek joik]] stag BO # sentence: is reduced to "gek sentence gik sentence".

• Note: the above transformation can't be done if both "(jek | joik)" and "stag" are present, because gek can only cover one of them. For example {broda i ja ba bo brode} has to be somehow both {ga broda gi brode} and {ba gi broda gi brode}. One way to deal with this would be to connect those two with {ge ... gi ...}, but that involves repeating the sentences, which means things like anaphora have to be dealt with first, as they can't be simply repeated.

Once all statement-2 have been reduced to sentence, we proceeed with statement-1:

sentence [[I joik-jek [sentence|sentence]]: is reduced to "gek sentence gik sentence" (a form of sentence).

That leaves:

statement = sentence | prenex statement

which means that statement has been reduced to subsentence.