Thanks Niraj. That looks a lot like Rapidminer, which I guess may be more used in Europe. But I don't think Rapidminer has the ability to use the combination of parts-of-speech with causality verbs (causes, leads to, drives, triggers, etc.) to then automatically generate bayesian networks and the likes. But I'll dig into Weka to find out whether it does! Thanks again

May be KNIME: http://www.knime.org/ ? With KNIME you can build also applications.

Automap (for visualization you can use ORA Casos): http://www.casos.cs.cmu.edu/projects/automap/

And just to be more specific - what I am looking for is sthg along the lines described here: https://netfiles.uiuc.edu/girju/publications/papers/rox-AAAISpring02.pdf. or http://cs.iupui.edu/~yxia/Publications/DASFAA11_AutoBayesian.pdf. I presume the general datamining programs Niraj and Dr. Petrič mention MAY be able to do this - they both have text-mining tools in them (as does RapidMIner), but I haven't found THIS particular functionality in any of those. Yet! The quest continues... And assistance from anybody who may have done sthg like this remains highly appreciated.

Bayesian lab in http://www.cs.ubc.ca/~murphyk/Bayes/bnsoft.html is a powerful tool. However they give only 1 month trial. Here where I learned on casual analysis. They give a powerful visual analyzer and nice tutorial.

Thanks! But does that also allow to learn from text? If so, do you know of anybody who has done that (with papers we can take a look at? Anyways - this DOES look interesting, and they so seem interested in research projects too, so I'll contact them.

Not sure, but I think WordStat (content analysis and text mining tool) can do this kind of analysis: http://provalisresearch.com/products/content-analysis-software/