Ex researcher at Bell Labs (Holmdel, vision processing) in 70's the various other jobs. Also simultaneous interpreter (Justice Dept., SDNY, State Dept., etc). "Semi-retired in La Jolla, Ca. My interest in graph databases based on interest in graph theory in general, and also in how it applies to language. As interpreter, I know that current Computer T\translation schemes do not work. Only amateur interpreters and translators go directly from string to string. I want to build a graph based general scheme. Also, Graph theory should be the basis of how data is stored, to approximate memory in living things. Finally, interested in connection of graph theory to category theory.
Interesting topic (language translation). Interested in your findings
Language translation is indeed interesting. When most multilinguals say that "they think in 'that language' when speaking 'that language', and "that lang" when speaking "that lang", what they are saying is that unlike computer programs, they are not directly manipulating strings, even with great look-aheads (context??). There is another level. That level is how the referenced is stored. So, in a way, it is about embedded recursive graphs- graphs that contain nodes that may be graphs themselves. These graphs may be modelled after Voronoi graphs. (You know, when we say, "Yes, this word is closer to what I mean ( either in the original language (I would say that this color is orange, I felt (sad synonym), or in a translation (simpatico is ...)).
Anyhow, this is the general gist of what I am doing, and would be interested in collaborating with all who might be like to investigate with me.