The Paradox of Mary and Mark

What is “consciousness”? (part 2)

This is a variation on Frank Jackson’s thought experiment “Mary’s Room” (1982).

This much of the argument should convince you that you believe subjective consciousness is real. There is nothing confusing or mysterious about the concept. You know exactly what I mean when I say “there is a difference between knowing which things are red and knowing what red looks like.”

This is the first corollary: you can’t detect subjective consciousness from the outside. Philosophers call this “the problem of other minds“.

This is supposed to convince you that mechanism* can’t explain phenomenal consciousness.

This last is a form of the hard problem of consciousness. There appears to be a huge gulf between “experiencing something for yourself” and “finding out about something using a machine”. But we are machines, so how can there be a difference?

---

Solving the Hard Problem

Answers to the question posed in this thought experiment:

Substance Dualism: We aren’t just machines. There’s also another substance in us, a “spirit” or “metaphysical consciousness” that actually has the subjectivity.

This is no help, for two reasons. First, if there was some kind of “invisible stuff”, physics and neuroscience would probably have found evidence of it by now. Second, you’ve solved the hard problem but created a harder problem — now you have to explain why this “invisible stuff” feels something. Remember, it’s invisible, so … good luck with that. 

Emergence: Mary and her equipment form a system that “knows what red looks like”, even if Mary herself doesn’t know what red looks like. The system has phenomenal consciousness of red, even if Mary does not.

Panpsychism: Machines actually do have experiences. It’s undetectable from the outside, but every machine, down to a single atom, actually has some subjective consciousness. Mary’s equipment knows what red looks like, the exact same way Mark’s eyes know what red looks like. All things know what their experiences feel like.

Eliminative materialism: There isn’t really a difference between “knowing what red looks like” and “knowing which things are red”. These are two ways of saying the same thing. The fact that it appears to be different to Mark is just a useful illusion— an artifact of his equipment, a distortion in his knowledge.

Biological Naturalism: Yes, they are both machines, yes they both implement the same function, but they implement the function using different machinery. Mark’s machine does different things to get the same result. The technical details of how the machine is built matters, even if the information it processes and function it implements is the same.

Mysterianism: You can’t explain it. The paradox above proves there is no way to solve the problem using logic, language, or human minds.

None of these solutions are particularly satisfying, at least not to me. I feel like they all create even harder problems. Either A: non-human “stuff” or “systems of stuff” or “invisible stuff” have subjective experiences (which doesn’t really tell us what it actually is or how it works, and actually makes it harder to explain, not easier). Or B: subjective experience isn’t what it appears to be, so even our own experience doesn’t count as evidence for anything. Or C: Don’t ask. (Not helping.)

---

Subjectivity in General

Subjectivity is more than just identifying colors, of course. It includes all other kinds of sensory experience; the experience of our emotions; what it feels like to have thoughts, memories, imagination, “inner speech”, “train of thought” and other mental experience; the way it feels to do things on purpose (that is, “free will”).

(This is just a loose list — I’m sure I’ve missed some things and I’m sure there are better ways to organize this.)

Objective (machines or programs)	Subjective (people)
Know which things are red	Know what red looks like
Know what you should avoid	Be afraid
Know what’s damaged	Feel pain, suffer (have “sentience”)
Take an action based on a calculation	Doing something on purpose (have “free will”)
Cause an accident, due to malfunction or design flaw or misuse.	Do deliberate damage
Be dangerous	Be evil
Have a goal, a utility function, or an objective function.	Feel desire
Have a program trace of your computations.	Experience a “train of thought”
Communication between objects in your program.	Experience “inner speech”
Direct most of your computational resources at one problem.	Pay attention to something.
Model the features of your process inside your data (as a debugger or operating system can)	Experience what is happening inside your mind and know that you are experiencing it. (Be “self-aware”)

It’s not hard to devise similar thought experiments for Mary and Mark that shows how the paradox works for these other things as well. Just suppose that, for some reason, Mary doesn’t have one of these things and uses machinery that serves the same function — machinery that is functionally equivalent.

---

The Point

The main thing I hope you take away from this is that:

1. You know what subjective consciousness is and you believe in it, or at least, you accept and act on its implications in your day to to day life (regardless of what you might say when you’re caught up in a philosophical discussion of some kind).
2. You can see that it leads to a paradox and that this paradox is hard to solve. No simple answer works for everyone all the time; it is profoundly arguable.

---

* I say “machines” and “mechanism” here, but the argument works exactly the same way if we want to focus on the system of information or the function that their brains (or equipment) are implementing. It doesn’t matter if we’re talking about the hardware of software; you will find the same paradox. So this thought experiment works equally well for the philosophical theories of functionalism or computationalism as it does for mechanism.

---