Natural Language Processing with Python … and Existentialism
After O’Reilly offered me a discount on their stuff I went off and bought some ebooks – including Getting Started with PyParsing and Natural Language Processing with Python. Both of them look very interesting and relevant to my work (ultimately, processing legal texts). While I haven’t read either of them in depth, both are intimately connected with parsing texts. From the looks of it, neither will be an exact match for me, although both look like they can be coaxed to where I want them.
I hadn’t realised that part of the aim of NLPP is to assist on the path of constructing ‘artificial intelligence’/understanding meaning in texts. The NLPP text talks a little about the philosophy of language but fails to mention one of the major philosophical criticisms of artificial intelligence – ie, that “substance ontology” is an inadequate way of representing reality. Substance ontology is from memory that things are subjects (chair) which have predicates (ie properties – short, red) and the criticism is that things are not like this, rather, what they are is determined by context. For example, the Coke Bottle in The Gods Must be Crazy wasn’t anything to the bushmen, least of all was it a coke bottle. Indeed, something has no meaning without being interpreted in the context of a whole heap of other things – eg hammerness has no meaning without nails, beams, houses etc.
If anyone is interested in the concepts I refer you to these sets of podcasts (the first especially, and it is quite entertaining – all of Professor Dreyfus’ podcasts are entertaining, the Coffeen lectures I gave up on half way through):
Dreyfus: Philosophy 185 Heidegger <- this one probably talks to the issue more directly, but is also pretty hard going