Noam Chomsky’s admirers are often surprised to hear that he spent the first decades of his MIT career working for the Pentagon. Many will be even more surprised to hear that he continued to work with a Pentagon contractor even after retiring from MIT.
In 2021, the military-funded research company Oceanit hosted an online lecture by Noam Chomsky. The lecture was introduced by Oceanit’s CEO, Patrick Sullivan, followed by its Director of Artificial Intelligence, Dr. Jeffrey Watumull.
Introducing the eminent linguist, Dr. Watumull explained:
[Chomsky] has worked with with us at Oceanit on AI programs to push precise ideas to absurd but successful conclusions and applications in this process that [Patrick Sullivan] describes as Intellectual Anarchy.
Chomsky repeated this striking phrase in his lecture, telling his audience:
What's needed is Patrick's Intellectual Anarchy, [a] willingness to challenge established doctrine with better ideas.
Intellectual Anarchy is also the title of Sullivan’s 2020 book – a book written to describe the ‘Intellectual Anarchy methodology followed at Oceanit’.
More recently, in October 2023, Sullivan gave a talk to a US Army roundtable on ‘Intellectual Anarchy’ and the need for ‘disruptive innovation’. Oceanit’s report of this talk points out that:
With ongoing conflicts and upheavals in places like Ukraine and Israel, gaining a competitive edge through disruptive innovations within the Army could act as a deterrent to other nations contemplating military aggression.
This whole concept of Intellectual Anarchy is reminiscent of Jerome Wiesner’s enthusiasm for the ‘anarchy of science’, ‘scientific anarchy’ and ‘planning for anarchy’ – all of which he saw as crucial in order to encourage the ‘free’ environment necessary for scientific creativity. As readers may know, Wiesner was the influential military scientist and lab director who in 1955 first recruited Chomsky to MIT.
In 2021, Patrick Sullivan contributed to another book, AI at War: How Big Data, Artificial Intelligence, and Machine Learning Are Changing Naval Warfare. In this book, Sullivan and his Oceanit co-authors (which presumably include Watumull) made several references to Chomsky’s work and closed their contribution with these words:
… the beginning of anthronoetic AI could be the end of war, and that surely is the ultimate objective of the US military.
Although this anthronoetic, or human-like, AI is inspired by Chomsky’s linguistics, it would be hard to think of a sentence that contrasts more with Chomsky’s passionately anti-militarist politics. Yet this didn’t stop Chomsky working with Oceanit. As the company’s website explains:
Oceanit has been working with Professor Chomsky for around five years, developing a linguistics-based Artificial Intelligence that we affectionately called NoME. NoME stands for Noetic Mathematical Engine, but is pronounced in the same way as its namesake.
Other Oceanit webpages reveal that this NoME system is intended for use in ‘war games’ and ‘cyberwarfare’. We also learn that the company has worked on ‘hundreds’ of US Government contracts. It appears that at least 90 of these were military contracts. Indeed, Sullivan himself says that Oceanit’s AI systems are ‘mostly in industrial and military applications’.
The actual Pentagon contracts that use this ‘anthronoetic AI’ are still more revealing. One 2020 contract is entitled: ‘Navy Technology Acceleration - Machine Learning (ML) and Artificial Intelligence (AI) to develop capabilities and impact mission success’. The abstract for this contract is as follows:
Consistent with the Navy objective, Oceanit is working from foundations in cognitive science, philosophy, and the mathematics of computability theory to create anthronoetic AI (a.k.a. strong AI): human-level/human-style AI capacitated with the linguistic competence to generate causal explanatory models via the Popperian philosophy of critical rationalism: given a set of big or small data, Oceanit’s AI – named NoME (the noetic mathematical engine) – constructs a Chomskyan ‘grammar’ (i.e., explicit theories/programs) to model causal relations, thereby transcending (but obviously including) descriptions that answer what is being observed, transforming data into evidence for/against conjectured explanations that answer why and how the data – or, to be precise, the phenomena underlying the data – exist and behave.
According to the contract, Jeffrey Watumull is the ‘senior scientist’ on the project.
Dr. Watumull was, apparently, Chomsky’s last PhD student. By 2017, he had become the ‘senior scientist’ on a million-dollar US Army project to develop a ‘virtual staff element … to provide a commander both organizational and operational recommendations to meet mission objectives.’
Today, Watumull is a 'principal investigator on programs for DARPA', the Department of Defense’s Advanced Research Projects Agency.
Although it is possible that Chomsky simply overlooked Watumull’s involvement with the US military, this is unlikely considering Chomsky has worked with Watumull on a co-authored book, Life, Information, Language, Intelligence, as well as on several linguistics articles.
Indeed, the author description for one of these articles explains that Watumull is presently a ‘principal investigator’ on both a DARPA seedling and an Office for Naval Research research proposal.
Another Watamull/Chomsky article is referenced in a 2022 US Navy proposal to ‘develop a human-level/human-style artificial intelligence (AI) that can perceive and explain signals ... to achieve long-range detection, tracking, and classification of maritime surface and subsurface contacts.’
This proposal – which explicitly invokes ‘Chomskyan grammar’ – concludes with the hope that it will ‘complete final testing and perform necessary integration and transition for use in antisubmarine and countermine warfare, counter surveillance … and future combat systems.’
So what’s going on here? Why has Chomsky continued to work with someone who is so clearly involved with the US military — despite his own passionate and heartfelt opposition to that same military? Could it be that in his eighties and nineties, Chomsky has been manipulated by a younger and smarter colleague?
I doubt it! Anyone who has watched Chomsky’s numerous YouTube lectures in recent years will know that he retained his wits and intelligence right up until the summer of 2023. Although he seemed to be letting Watumull do much of the writing, Chomsky was still quite capable of authorising texts written in his name.
So, could it be that Chomsky just doesn’t care if his linguistic theories are being used to benefit the US military? Perhaps he believes, as does Dr. Watumull, that it would be preferable for the military to work with ‘responsible defense contractors’ – i.e. those who are working on intelligent Chomskyan AI, as opposed to unintelligent AI. Again, I doubt it.
Rather, it seems, having become accustomed to Pentagon funding for so long – and having defended this funding for so long – Chomsky just got used to the risk that his work might one day prove useful to the military. The fact that his linguistics has always proved completely unworkable for the Pentagon, or indeed anyone else, has only reinforced Chomsky’s willingness to take such a risk.