hmm I returned to PTb to algebraic representation task.

I decided to do it like: 

hmm  

I , K and G groups 

I is subgroup of K and K is also subgroup of G

then G/I = g

and G/ K  = 2 

and to have a homomorphism from G/I to G/K and  lets call it h 

then Kernel h be  being K/I  representing the We part and there () (0 element of G) represents I and others represents in K/ I  you

and then G - K/I represents they and elements of it represents he/she/it or all pronouns other than I / you

I then selected the subgroups as: G being dihedral group

I = () of G

K =  being rotation subgroup 

G = the dihedral group 

hmm 

then also e.g. go literal could had been represented as:

 a union of homomorphisms where each homomorphims is from cyclic permutation 0 to n to dihedral groups either r or sr generator.  

hmm for to like prepositions i think or other Ptb types hmm i might need to study rings for geometry ring based representation. 

for cause and effect representations, I had not figured out how yet./ e.g. why literal. 

hmm so lets continue tomorrow then. 

hmm maybe initially I could omit geometric representations since I dont think  I can fastly learn rings based geometry representations yet. but I could have prologue like rule based definition sets for geometry definitions to at least some additional semantics there with patching missing algebra representation alike with prologue definitionsets. hmm seems logical approach for first version where i wont yet be any expert in group theoretic topics yet nor geometry's ring based definitions etc.

hmm I think this could be started like this initially at least for 0.1 ai version.

hmm idea of TF /IDF method for this time ontology signature support system could give an additional support system to dependency graph information that can be generated by transformer net.