"But how do you distill and analyse these huge volumes of raw data to uncover business critical patterns and relationships especially hidden, subtle relationships and trends no one has ever thought to look for?"

Text data mining
    - text is different from other data which is often relational or has some other well defined structure
            - countless combinations of subtle abstract relationships among concepts
                    - intentionality
                            - highly redundant and susceptible to simple algorithms for coarse tasks
***concept hierarchies***


"Traditional query based analysis involves posing specific questions to the data. The trouble is, the answers are limited by the assumptions the questions are based on... So patterns which are counter intuitive or that relate to questions no one thought to ask remain hidden."

Artificial neural networks, machine learning and natural language research refine the question and answer procedure.

Intriguing parallels with one of the seminal theological texts of the western tradition, the Dialogus, by William of Ockham, written in the early 15th century in the form of a dialogue between master and student. Here the student puts questions to the master, who retains a neutral voice, giving different possible examples in answer to the questions without providing an opinion or judgment. The Dialogus is concerned with heresy, how it can be identified and by whom, and what should be done about it. It draws a distinction between canonical and theological "science" (the Latin scientia) - top down and bottom up approaches to Christian truth and dogma, and talks about the different methods used to "decide" on an issue. It's a treatise on interpretation and institutional power.