Joel showed me this great website
It is quite incredible what it can do in terms of calculating bi-grams, 3-grams etc for you.
I also was told by my friend Hema about the voices in the NeoSpeech Lab
NeoSpeech Text-to-Speech voices | Interactive TTS Demo | TTS SDK
Besides being amazingly realistic they can be used as a proof reading tool for your paper,
if you have the patience to paste it into their sample box, paragraph by paragraph...
There is also the XLE online that the Norwegians made available here INESS :: XLE-Web.
But my favourites are the OpenMultiLingual WordNet Multilingual Wordnet 1.0
and the Portuguese corpus Corpus do Português. Go NEH (National Endowment for the Humanities)!