Freakzilla wrote:Mentat thinking is on the other end of the spectrum from mechanical thinking.
Mentats were valued for their ability to get accurate conclusions from incomplete data, they are trained to get as much from what is missing as the data they have. The work with patterns. This was a big theme too.
Could a computer ever be taught to recognize a pattern it has never seen before?
I think a big difference between computer AI and a human brain is that a computer is restricted to data in ---> data out, and the data is what's fed to it (like a mentat). Once the AI processes the data it will come to a result and will then be idle as its work is done. The AI's data will be fixed quantities, and once an operation is complete it will need new data for a new computation once the old one is considered obsolete. Its computations will always be about external data sources fed to it, but never about itself, since computers don't 'want' anything and don't involve themselves in their own thinking process.
Now, a person can think "How can milk be obtained?" and also "I want milk!", and those two queries can become correlated as the desire for milk leads to wanting to know how to get it. A computer can also assess how milk is gotten, and even can be fed the premise that it is to collect some. But here's what a computer cannot think: "Wait - does my knowledge of how to get milk lead to my desire for it, or was the desire that which led me to explore how to get it? And now that I know this, does it change either my desire or my decision of the best way to get it? And now that I know THIS, does it change...etc etc." The human brain can constantly modify thinking patterns based on recursive and repeating interaction with the information and desires it possesses. It's a never-ending modifying system, never able to just 'know the answer' and have done with it. Knowing something changes the person doing the knowing, and so the information gained itself becomes changed as the person does, which then changes the relationship between observer and observed on a constant basis. This is the big difference between a mentat and the KH, as the KH includes himself and his knowledge among the data in the universe to study and this relationship leads to a constantly changing perspective and truth-as-relationship. The 'infinite regress' of search for the truth and forecasting future actions and decisions is the speciality of the KH, and is impossible for a mentat, who is reliant on fixed data for projection and who is obliged to eliminate himself from the data being observed. The best mentat is the one most able to bypass his own humanity and to focus on the facts fed to him, and the KH really does the opposite of this.
Frank was concerned, I think, with the fact that 'knowing' and thinking about how to live are a process, not a solution to be found, and that changing how we think is just as important as deciding what to think about. I think mentats are presented as being the human equivalent of thinking machines, as we can see here:
Dune Messiah wrote:"A mentat!" Mohiam muttered.
Scytale glanced at the old Reverend Mother, seeing the ancient hates which
colored her responses. From the days of the Butlerian Jihad when "thinking
machines" had been wiped from most of the universe, computers had inspired
distrust. Old emotions colored the human computer as well.
I would say this sounds more than just a prejudice, but more like a suspicion that mentats utilize the kind
of thinking that machines used, even though the users are human; and the jihad was against machine-logic (ref: Dune Appendix II), not machines. I can certainly see someone astute like Mohiam being wary of humans adopting machine-logic, and indeed, Leto II banned mentat training outright.
In the context of human development, while the KH was a 'super-being' of a sort, his important attribute was the understanding of the ever-changing future and the frail relationship between the thinker and the objects of thought, and therefore between the activity of the present and its relationship to the vision of the future. In real life any of us could work on getting more attuned to this relationship, and in trying to avoid machine-logic - the attempt to make simple answers and finite solutions, as if reality were a logic puzzle to be solved.
In the books the KH's power is eventually nullified by the 'Siona gene', but that doesn't mean the KH's lesson is lost; in the end power over others (which properly ought
to be nullified as time goes on) isn't the point, but rather coming to realize what real understanding is: it's not an answer, it's a question!