Here is my attempt at a summary (disclaimer: going from memory; watched this and related lectures during my MA thesis last year):<p>- Early scientific inquiry in the enlightenment period (by those who referred to themselves as "philosophers", but who we would now call "scientists") was based on the "Mechanical Philosophy". At this time skilled artisans [who we might call "hackers" today] were creating sophisticated machines, e.g. clocks, talking machines and simulations of digestive systems.<p>-These new machines stimulated the minds of the scientist-philosophers in a similar fashion to the machine learning/AI demos today. A nascent philosophy of science emerged at this time: they wondered if the whole universe was one giant machine, viewing God as a kind of skilled artisan who carefully tuned the universe, which as a result operated on the same mechanical principles. In this philosophy of science - the "Mechanical Philosophy" - everything in the universe can be understood in mechanical terms [which are intuitive to human beings; e.g. babies assume a hidden point of contact when shown a cause and effect at a distance]. In the Mechanical Philosophy, the nature of the universe is intelligible to us by direct observation.<p>- Newton, to his own dismay (and ridicule by his contemporaries), showed that the Mechanical Philosophy was incoherent, because the "action at a distance" posited in his theory of gravity could not be explained in purely mechanical terms. Chomsky argues that post Newton, the goals of science ("philosophy" at the time) became more modest: the goal of understanding nature/the universe in direct terms [by applying intuitive explanations] was abandoned! Now we try to build _models_ of natural phenomena that are intelligible to us. The model is intelligible, even if the underlying phenomenon is not. All modern science works this way, e.g. quantum physics, chemistry, biology, linguistics.<p>- This is what Chomsky means when he says that Newton exorcised the machine, but left the ghost intact. We no longer seek mechanical explanations - the machine was exorcised from our theory of science - but the "ghost" is still there. After Newton, the problem was largely forgotten. But "exorcising the ghost" - i.e. forming theories about nature that are directly intelligible to the senses - is considered too hard to approach (you can't even "ask stupid questions" about it.). Nobody in chemistry is trying to form directly intelligible theories of chemical bonds for example. We just expect the model to be intelligible.<p>- Chomsky argues that "exorcising the ghost" - understanding nature in direct mechanical terms - might be beyond the capability of our species. It would not be surprising that our species has inherent biological limitations such as this, since we are ordinary animals, rather than angels. Every other animal has well defined scopes and limits in their various capacities. The scopes of our capacities as a species go hand in hand with the limits. The science-forming capacity is just one aspect of our biological endowment, which is limited in this way. Plausibly, aliens watching us might observe that we are unable to make progress in certain areas of science due to these inherent limitations. [This ties somewhat into Chomsky's view of language, as a biological system with inherent scope and limits].<p>Takeaway: we shouldn't expect the study of all of nature to be reducible to the study of physics, or some unifying theory, as is often assumed. Two different models might be reducible or integrate (e.g., attempts to unify physics and chemistry), but this is not necessarily the case, because the study of physics is not "closer" to reality or more fundamental in any meaningful sense (since the study of physics is really just another set of intelligible models). Likewise, the models we use to study biological and mental systems are not necessarily reducible to physical models [it is only worth pursuing these unifications if they improve the study of one or another field, they are not useful goals in their own sake].