Embracing Wittgenstein's Nonsense: The Creative Exploration in Philosophical Inquiry and its Implications for AI Advancement through LLMs

by Casian STEFAN, Principal Researcher at Essentia Mundi AI Lab. Contact: ai-AT-essentiamundi.com / ai.essentiamundi.com
Dec. 2023.

Please consider citation with link, if you derive work. Or contact me for collaboration. Thank you!

Wittgenstein's instruction to "pay attention to your nonsense" carries a heavy weight meaning. One can not make the immediate parallel with the current Language Models that make use of the Transformer with attention heads architecture. In which their inner space, while creating clusters and then converging to the next plausible token (*) (so powerful mechanism that is able to create syntactically valid formed phrases) they also exhibit a semantic "confabulation" characteristic. Eg. they do now know that what they say it is true. I would add that there is a need for new characteristic: of paying attention as in finding connections of a lot more to be generated confabulations.

This very characteristic should be more carefully examined, as this can represent a key for AI advancement. In the sense that an Agent's navigation and action should include not-very-stated objectives (**) and that it should have a balanced, even unambitious stance, yet still able to solve, create, etc. This mechanism should avail of creativity input, the latent space filled with potential new ways.

In philosophy, Wittgenstein acknowledges the potential for seemingly nonsensical ideas, but he emphasizes the importance of carefully examining and understanding of these inquiries. The process of paying attention to one's nonsense involves critically analyzing and reflecting on the ideas, exploring the underlying assumptions and implications. The idea is that there should be a space of that ideas to be produced in the first place. A space which should enable the Agent's oscillation between two seemingly differing attractors.

The inquiry between should have a degree of nonsense. Then by paying attention to the nonsense, the agent should gain insights into the structure of a subject matter, and fall within limits of meaningful expressions. The examination of seemingly nonsensical statements can lead to a deeper understanding of the underlying concepts and the ways in which its language space shapes further its perspectives.

In Wittgenstein's view, nonsense in philosophy is not to be feared but embraced. It is a natural byproduct of grappling with complex, abstract, and sometimes elusive concepts that characterize philosophical discourse. Nonsense, in this context, is not a dead end but rather a signpost pointing towards uncharted territories of thought.

The essence of Wittgenstein's directive lies in the notion of paying attention to one's nonsense. This involves more than merely acknowledging the existence of seemingly nonsensical ideas; it requires a deliberate and reflective examination of these ideas. By dissecting and scrutinizing what appears as nonsense, philosophers can unravel hidden meanings, expose underlying assumptions, and gain insights into the intricacies of language and thought.

So we should see the philosophical inquiry, at its core, as a creative endeavor. The willingness to explore unconventional ideas, even if they initially appear nonsensical, is a catalyst to the creative spirit inherent in philosophy. This creativity is not confined to the realm of abstract thinking; it has practical implications for problem-solving, innovation, and the generation of new knowledge.

The connection between philosophical exploration, nonsense, and AI becomes particularly intriguing when considering the concept of confabulation in LLMs for example. Confabulation is to be seen here, as a process by which an intelligent system generates information to fill gaps in its knowledge. Embracing nonsense in philosophical inquiry aligns with the idea of encouraging AI agents to explore unconventional paths, fostering creativity in their learning processes but retain an aimless objective (***).

By allowing AI systems to engage in creative exploration and confabulation, we open the door to new possibilities, novel solutions, and the ability to assert new "unknown unknowns." (****) This process mirrors the human capacity for creative thinking and problem-solving, enabling AI to navigate uncharted territories and contribute to the ever-expanding landscape of knowledge.

Wittgenstein's urging philosophers to embrace and pay attention to nonsense, underscores the creative nature of philosophical inquiry. This creative exploration not only enriches our understanding of complex ideas but also holds promise for the advancement of AI. By incorporating the principles of confabulation and encouraging AI agents to engage in creative exploration, we pave the way for the emergence of new knowledge, solutions, and a deeper understanding of the unknown unknowns that lie ahead. Like us, but having the power of marring philosophy inquiry and AI, we are endowing the systems with a powerful feature that can unlock innovation and can push the boundaries of what AI / and we can conceive and achieve.



_____________________________________
Exploration by C. Stefan, Dec. 2023 [about]
Last update: Dec. 2023 (versions: *) [versions]
"Essentia Mundi" AI Research Lab. [home]
Copyright © 2023 AI.EssentiaMundi.com, all rights reserved.
_
References:
* A mathematical perspective on Transformers (Borjan Geshkovski, Et al.)
** Why Greatness Cannot Be Planned: The Myth of the Objective
(Kenneth O. Stanley)
*** Intelligence and Unambitiousness Using Algorithmic Information Theory (Michael K. Cohen, Et al.)
**** Borges and AI (https://arxiv.org/abs/2310.01425)