Disclaimer: I have a spectrum of posts, from “thoughtfully laid out to attempt to appeal to people who disagree” to “more quickly summarized to appeal to people who already at least partially agree with me.” This post is in the second category.
I want to contribute another word to my pile of cheesy invented terms: bonsciousness.
One of my pet peeves is people confusing questions about consciousness. I’ve seen “can we ever scientifically solve the hard problem of consciousness?” uttered just a few sentences away from talking about “what is the origin of qualia.” Consciousness is a fascinating problem for reasons that render those questions useless, but in order to make this more clear I want to divide the one concept into two.
“Consciousness” is the word we (should?) use for the conceptual model we have of “other things being self aware.” When we ask the question, “how do we know if an AI should be treated ethically?” we’re probably asking if the AI is conscious. The P-Zombie thought experiment has to do with whether or not other beings contain this elusive “consciousness” quality.
Consciousness can’t be known for sure, as we could be in a simulation where everyone else just behaves very convincingly, or a dream where we are absolutely convinced that we are talking to an intelligent friend, despite them just being a projection of our own minds.
When we handle the concept ‘consciousness,’ we’re usually handling something like the concept of ‘how much do we feel that other beings exhibit patterns that we uniquely identify with’ – as in, we might think an AI is conscious if it can do things like “use language creatively.”
The key concept of consciousness is that it is something that can be applied to multiple things. There can be multiple consciousnesses, whatever it means or however sure we are of it existing. Multiple awarenesses feel like it makes sense, and is calculable, or measurable, and one day we might be able to do better science to it and feel like we’ve made additional consciousnesses.
Bonsciousness, on the other hand, is fundamentally singular. It is the subjective and immediate awareness of the self. It is direct experience.
The concept of bonsciousness becomes relevant in thought experiments like Mary’s Room (if a color blind scientist studies color for a thousand years, will surgically fixing her color blindness on the 1000th year give her any new information?), in questions of identity – but it always has to do with the nature of experience.
It cannot be multiple. If you try to imagine someone else possessing bonsciousness, you are not thinking about bonscoiusness, you are thinking about consciousness. This may seem a subtle distinction, but I find it incredibly important. Teleportation poses no deep philosophical questions when it happens to other people – the importance lies in the subjective and personal experience. The question “does teleportation kill you” relies on what the continuation of experience feels like – while the question “is AI conscious” relies on whatever markers we have that we think needs to be met for consciousness to exist. These are two questions that come from two hugely different types of ideas.
I find this to be an unintuitive distinction for some people, as it’s very common for people to combine the concepts of consciousness and bonsciousness in their own mind. I also find that if one does not already find self-awareness to be deeply philosophically strange, it’s difficult to induce that sensation in them through argument alone and I don’t expect this post would accomplish that.
I do suspect, though, that at least attempting to use different terms for singular/multiple ideas of consciousness would clarify a lot of the conversations I’ve been listening to lately. I’ve been hearing people ask a question about bonsciousness and then attempt to answer as though they’re talking about consciousness, which is quite frustrating.
Bonsciousness is so elusive because it is about a category of knowledge that isn’t measurable, and trying to treat it as measurable shuts down a lot of avenues of learning.