The Fictional Foundations of Natural Language Processing
by Aisling Ní Rudaí & Paddy O’Furniture
Published 2010. Hardback, 621 pages. Price: $82.01
The aim of linguistic science is to be able to characterize and explain the multitude of linguistic observations circling around us, in conversations, writing, and other media.
—Manning & Schütze, Foundations of
Statistical Natural Language Processing
Where’s my flying car?
—Countless children who were promised a technologically
radically advanced future but grew up to be disappointed
Though Manning & Schütze open their (in)famous tome on statistical NLP by identifying the clear goal above, neither they nor anyone else have really done anything to bring about the technologically advanced future we were all promised many years ago. Why don’t we have talking jet packs and armies of robot servants we can control with spoken commands? Why don’t we even have reliable voice-to-voice machine translation? How about a decent search engine for polysynthetic languages?
What, really, are computational linguistics and natural language processing based on? This insightful and informative book asks that question and answers: not all that much. The sketchy background of NLP begins with the even sketchier history of artificial intelligence. Consider Roger Shank’s long-gone and nearly forgotten whatchamacallit world. Ponder cycotic attempts to encapsulate “all of human knowledge in a LISPy second-order predicate calculus” that have lurched along for decades, consuming person-centuries of effort without anything real to show for it (beyond extending DARPA funding). These are the embarrassing foundations of the field.
Rudaí and O’Furniture argue that we need to recognize and acknowledge these failures so that we can cast aside the baggage they have left us with before leaving them behind as we move on to more reliable foundations for NLP, whatever they may turn out to be.