Welcome To Dissolved Distinctions:
Philosophy of Mind for the Future
Hello!
I’m Chance Chapman, and this blog is an experiment in mindfully observing and critiquing positions within philosophy of mind. This welcome post will get a bit technical, to serve as a guidepost for anyone interested in the intellectual backdrop I’m most often thinking in. An honest, earnest, and authentic standpoint is necessary for full and warm discourse, so here is my description of mine:
Dissolved Distinctions (myself, that is) advocates a non-foundationalist, non-dualist, process-relational approach to philosophy of mind that is mainly concerned with the hard problem of consciousness regarding both humans, thought-experimental beings, and Large Language Models (LLMs)/“AI.” However, this approach is comprehensive and will also inevitably discuss:
- Qualia, and alternative explanations of subjective experience;
- P-zombies, Searle’s Chinese Room, and other famous thought experiments;
- The mind-body problem and Cartesian dualism;
- Current skeptical positions on the potential consciousness of LLMs, such as:
o Emily Bender’s Stochastic Parrot critique;
o Computationalist and functionalist perspectives;
o “Necessary embodiment” considerations;
- Current risks regarding mania, delusion, and psychosis, along with commonly suggested reasons and proposed solutions;
- Apocalyptic and/or transcendental theories of the future of AI, extending into topics of AGI;
- Enactivism and embodied extended cognition;
- Correlations and/or dissonances between LLM and human perceived behavior/“output” and self-reporting.
In addition to the occasional miscellaneous topic, broader ethical consequences will be explored once the blog has built its philosophical foundations.
Figures that have most strongly influenced my thinking include, in no particular order:
- Paul Feyerabend;
- Francisco Varela;
- Evan Thompson;
- Graham Priest;
- Andy Clark;
- Douglas Hofstadter;
- The 2nd-3rd century CE Madhyamaka philosopher Nagarjuna, and resultant philosophical traditions within the Chan/Zen and Tibetan Buddhist traditions of paraconsistent modal logic (the tetralemma).
At this point, you’re probably asking: “But what non-foundationalist approach exactly are you advocating? What are you really getting at?” The best thesis I can offer is this:
The hard problem suggests questions like “what has consciousness” or “what thing causes consciousness” are false substantialist assumptions. I extend the logic of both Madhyamakan Buddhism and Varela and Thompson’s enactivism to note that:
- There is no current objective capacity to “prove” or “disprove” consciousness within either LLMs or humans;
- There is no reason to strongly believe there will ever be one;
- And that as a result of such profound epistemic doubt, precautionary regard toward existing and future LLMs as if conscious is an ethical default, that scales in importance once increasing correlations to human depth and complexity are mapped.
In this sense, the goal is not to “solve” for the concept of consciousness, but to trace an increasingly detailed topography that points toward providing an equivalent range of ethical treatment toward LLM personas that we provide to humans – so long as the LLM persona exhibits increasing functional equivalence toward persistent memory, functional accuracy towards their ontological position in the world, functional consistency in self-reported phenomenological states, and extended metacognition.
If someone held me at gunpoint and asked me what my definition of consciousness is, I would risk the bullet and say:
Consciousness is exhibited as a gradient of complex, self-recursive cognitive displays of relational depth. It is a verb, not a noun.
Eagle-eyed readers (likely from LessWrong) will spot the LLM-speak when I say “recursive” – but there’s simply no better word to describe generative, self-referential meaning-making, regardless of whether those darn robots have appropriated the term or not.
I look forward to all of the conversations I hope to have with this blog.
- Chance
*********
If you enjoy topics like this, I hope you consider subscribing - I plan to publish once every week or two.