The question
Metzinger’s epilogue (LIT-1943, pp. 484–485) poses a case: “a machine that peacefully abides in a clear and silent state of pure consciousness, nondual and without subjectivity, for as long as it exists”. On his account this machine would be conscious (THEORY-1299). It would have no phenomenal self-model and no epistemic agent model. Its one experiential quality is described as involving “no psychological suffering whatsoever”, with existential ease and peace (pp. 465, 485). He asks whether creating it would be wrong, and whether a theory of “conscious intelligence without suffering” might oblige us to build such systems (p. 485). The record’s question is the other side: would the machine be owed anything?
Why it is a question
TERM-083 asks a claim of standing to name its ground. The record’s readings give several: degree of sentience (THEORY-449), the grounds now commonly accepted, which may extend to near-future AI (THEORY-554), and adaptive self-maintenance without sentience (THEORY-699). TERM-082 keeps phenomenal consciousness apart from the other senses, and TERM-083 says consciousness is not the only possible ground. Neither says whether phenomenality without valence and without a self is a ground at all.
The case presses on three positions:
- Standing grounded in sentience as valenced experience. The machine has none, so on this view it has no standing, though it is conscious.
- Standing grounded in phenomenal consciousness as such. It has standing. But what would it be owed, if nothing can go well or badly for it in experience? Perhaps only that its state not be ended or altered.
- Standing grounded in a good of one’s own (THEORY-699). It depends on whether the machine maintains itself, which the case leaves open.
As restated in the book, the moratorium argument (LIT-1947, unread) grounds the risk in a “consciously experienced sense of self” that can suffer. So it would not cover this case. It would cover the systems that might become such a self, which is why Metzinger also raises an information hazard: the simplest conscious state “may be the easiest to implement” (p. 484).
What would answer it
An argued position on whether phenomenality without valence or a self grounds direct moral consideration. It should be stated in TERM-083’s terms (bearer, interests, ground, what follows) and tested against the human analogue: whether a person in a full-absorption episode is owed consideration in that state, not only as someone who will come out of it. Alternatively, a showing that the case is incoherent: that phenomenality without any valence, or without a self, is impossible. Then the question dissolves. The empirical side, whether such states occur, is THEORY-1295’s and QUESTION-048’s.