Ripping through this. I have spent a lifetime where my science was computational cognitive neuroscience, but my engineering was AI/NLP/Vision/ML. I disagree deeply that somehow science and engineering are the same thing. Or even close. Science brings in the role of philosophy quite directly, but there are different kinds of scientists. The biggest split is between theoreticians and empiricists (or experimentalists). To be good at one is usually to be bad at another. The latter quite commonly is great at generating new facts about nature, while having very little clue about how the facts actually fit together, and the former bringing facts together not just into systems of who we are, where we are, and what we should be doing but what nature tells us might be the profound regularities within nature. Nature here includes ourselves as manifestations of nature which are also the subject of science (both theoretical and experimental).

My field of science (I first published on it in 1973) has had a popular and engineering revolution in recent years. Engineering serves the role of math, but math by computational modelling which is decidedly not computer engeering. This field is almost purely Cartesian, which is why I call Descartes the father of Cognitive Science. It seeks scientific answers to how and what humans can know, and why. Our knowledge in this area has grown exponentially and our theories have simplified. My complaint is, as a theorist, empirical facts from other scholarly domains such as philosophy (logic) and linguistic (grammars) that can be brought to the goals of cognitive science are often not taught so the science too often drifts down experimental paths one questions for which we already have the theoretical answers, or at least understanding.

Here are five short theory papers you gang of philosophers might be interested in.

1. We now are pretty sure we know how the human brain (actually any living nervous system) computes. And thereby what it can't compute.

https://medium.com/liecatcher/how-your-brain-computes-41ebe7428ff9

2. Human Natural Language is sure enough a window on how the brain computes precisely because it evolved explicitly an an efficient way for one neocortex to communicate its own computations with another. NL IS the Rosetta Stone of Computational Cognitive Neuroscience.

https://medium.com/liecatcher/natural-language-and-your-brain-237185770b00

3. Many philosophers, I have noted, conflate natural language and intelligence. They forget that one preceded the other for hundreds of millions of years. Separate them, and you see the brain and human evolution much more clearly, as well as fundamental limitations in what we can know, as well as the unlimited nature of what we can know. Chomsky, a linguist and not a computation guy, saw this as clearly. (e.g., Grammar derives from how higher order nervous systems actually compute even in dogs and mice).

https://medium.com/liecatcher/the-natural-evolution-of-human-lies-655e983ee6c6

4. Sadly, we can now see the death of humankind in a lack of vital NL dialogue that we learned from the Age of Reason, had to involve enforced moderated dialogue (lacking almost completely on the Internet). This comes from what is now a four year interest of mine in human (NL) lying. This is important for philosophy as well as society because it scopes precisely, who we are, where we are, and what we should be doing. In this article I invoke Descartes, but in my book on the same subject, I don't even mention him.

https://medium.com/liecatcher/https-medium-com-rhtcmu-fiat-lies-are-genocide-on-the-human-race-a4d76b093530

5. Finally the answer much discussed in this paper and many of the replies on progress in any discipline which turns out to be scholarship-specific "enforced moderated dialogue" which leads to progress. I point out why "Physicists Are the Best Liars."

https://medium.com/liecatcher/physicists-are-the-best-liars-c8e9ccfcf6b4

It is the fear of 4 spreading from bad cults, and the hope of 5 spreading as good cults, that I have been implementing the "Internet Court of Lies". To see how this court creates an ungoverned (i.e., self-governed private) place to correct the inevitable damage possible, and now observed, of NL in modern media, sign up for

https://www.liecourt.com.

This is to get a Zoom link for a weekly public trial. Trials are designed to last less than ten minutes, with just as many minutes to bring a case to trial. Volunteer to be a juror to see what enforced moderated dialogue on truth and lies can do for individual thought in the spirit of Descartes.