What's expected of us

http://www.nature.com/nature/journal/v436/n7047/full/436150a.html

by luu • 13 years ago

129 56 13 years ago

pattisapu2• 13 years ago

Ted Chiang's Predictor would be quite the expansion pack for Hermann Hesse's Glass Bead Game.

svermeulen• 13 years ago

Anyone that liked this should definitely check this one as well by the same author: http://infinityplus.co.uk/stories/under.htm

friendcomputer2• 13 years ago

I highly recommend Ted Chiang. I really enjoyed "Story of your Life" available in this short story collection of the same name: http://www.amazon.com/Stories-Your-Life-Others-Chiang/dp/193...

Some of the stuff is a but uneven but he's written some of the most thought-provoking sci-fi I've read in some time.

pattisapu2• 13 years ago

Looks like "Story of Your Life" is going to be made into a movie. http://www.tor.com/blogs/2013/10/ted-chiang-story-of-your-li...

dvanduzer• 13 years ago

I last heard his name when a good friend recommended "Hell is the Absence of God" to me.

I later tried to recommend it to my mother, but found myself incapable of doing so.

lmm2• 13 years ago

Can you explain Story of your Life? To me it just doesn't make sense, our consciousnesses don't work like that - which undermines any point it might be making.

eloff• 13 years ago

The predictor already exists, there's a common FMRI experiment in neuroscience where with about 80% accuracy the scientists can predict you're about to push the button about one second before you decide to do it. Someone no doubt will post a link, I don't have time to Google it now. I'm guessing it's the inspiration for this story.

I'm more curious about what you guys think. I've long thought the arguments against free will are much stronger than those for it, but it's an unsettling idea to live with.

dangrossman• 13 years ago

It'd be a "predictor" if an FMRI could predict that the neurons are about to light up a second before that happens, rather than just measure the delay between neuron activity and conscious perception or conscious motor control.

Regardless, if anyone finds the thought that an FMRI can detect you making a decision a significant amount of time before you're aware of making that decision particularly interesting, this book dives into that research and much more about the nature of consciousness from a neuroscience and information theory POV:

http://www.amazon.com/The-User-Illusion-Cutting-Consciousnes...

diydsp• 13 years ago

Yes, loved the User Illusion. Fascinating Stuff.

Also, lately I've been hearing a lot about http://en.wikipedia.org/wiki/The_Happiness_Hypothesis

in which the brain is likened to a person riding and elephant. We can train the elephant, but mostly the elephant repeats its program.

It's billed as about "happiness," but imo that's only one slice of it. It applies to all habits, including eating, exercise, improving sw development skills...

dullcrisp2• 13 years ago

Interesting. If the observers were to play a distracting tone every time they predicted that you were about to decide to press the button, would you never experience deciding to press the button?

return0• 13 years ago

Not 1, but 7 seconds before

http://www.plosone.org/article/info%3Adoi%2F10.1371%2Fjourna...

moondowner• 13 years ago

Wow, thanks for sharing this, will definitely read it.

brazzy• 13 years ago

All it realls proves is that "deciding to do something" is not the simple, conscious event that it seems to us.

sarreph• 13 years ago

I love pieces of writing like this.

Determinism isn't something I've explored quite satisfactorily enough just yet, and reading articles such as this one sends me straight back into the depths of the literature.

rzimmerman• 13 years ago

This reminds me of the Isaac Asimov spoof article "The Endochronic Properties of Resublimated Thiotimoline" which describes "experiments" on a substance that reacts so aggressively with water that it reacts before the water is added. Can't find the text unfortunately, but Wikipedia has a summary: http://en.wikipedia.org/wiki/Thiotimoline

troymc• 13 years ago

I just read about that in Asimov's autobiography. He wrote it when he was a grad student at Columbia and was worried that it (a satire of chemistry papers) might harm his chances of passing his final (oral) PhD examination. His examiners never mentioned it until the very end... when one asked him what he knew about resublimated thiotimoline, at which time Asimov burst out in a fit of nervous laughter.

He passed.

_sabe_• 13 years ago

Even if we were free thinking individuals, able to make rational decisions on our own, we base those decisions on past events and the current situation. So even if you get to choose, you'd probably make the same rational choices as long as your past experience and the environment is the same.

This would be true one second ago, and a second ago this would be true about the second before that.

mnemonicsloth• 13 years ago

Here's a similar appearance by Vernor Vinge: http://www.nature.com/nature/journal/v407/n6805/full/407679a...

throwaway7808• 13 years ago

That reminds me of Feynman's Thesis and his (failed) attempt to build his Feynman's Radio...

jaw• 13 years ago

Sadly we're not informed what happens if you program a robot to press the button at fixed intervals unless it observes a flash.

dullcrisp2• 13 years ago

Or, for that matter, a human.

ksmiley• 13 years ago

Some simple resolutions would be: a freak lightning strike disables the robot's button-pressing hardware. Or a cosmic ray flips a bit in its machine code, causing it to shut down. Or the robot's designer suffers a heart attack before turning it on.

These may seem like contrived examples, but when you eliminate all outcomes that aren't self-consistent, it may be that _all_ the remaining possibilities are contrived. Probability gets weird when time travel is involved.

jrlocke• 13 years ago

Does such a detector imply that there is no freewill? Would determinism imply it?

Imagine the predictor works, what does that mean? One second prior to the action, the course you would take is set. This is actually probably a point in favor of the possibility of freewill; humans are course-taking machines, we have a whole life (and our evolution before that) to acquire dispositions that will one day help us survive. What kind of freewill do we want? Do we want to be "free to dodge a brick when thrown at us" or "free, one second before the brick arrives, to duck or not duck". Which of these is the important one? The kind of freewill we really want isn't the kind where we are free floating actors, it is the kind where our history, personal and evolutionary, dictates our present. The point is, real freedom is not about being without limits, it is about having sensible responses to ones environment; creatures evolve the freedom to avoid being eaten, the freedom to anticipate others' actions, the freedom to manipulate other agents with our clever words. Dan Dennett makes a better case for this than I, please look into his wonderful books on the topic.

Daniel Dennett: --Elbow Room --Freedom Evolves

gbrandt• 13 years ago

Relevant: http://en.wikipedia.org/wiki/Timequake

marmaduke• 13 years ago

This is known in the dynamical systems literature as strong anticipation:

http://digital.csic.es/bitstream/10261/7493/2/Dynamical_mech...

http://www.haskins.yale.edu/Reprints/HL1521.pdf

tomp2• 13 years ago

Oh my... another meditation on the issue of free will, and the existential angst we would experience if we learned that our choices are deterministic. Personally, however, I think that the whole question of free-will vs. determinism is a false dilemma. For me, another issue is more important:

Determinism vs. randomness.

If my choices are not deterministic, they must be random. I would rather be deterministic than random. In any case, if they are random, it means that I don't have free will either.

I like to think that all my choices are a result of the past; my genetics, my childhood upbringing (that I don't even remember, but which formed the neural pathways in my brain), my early memories, my environment, education, experiences, friends, knowledge that I have absorbed from the world... In each and every moment, I make a choice, which is the best choice I can make given my brain power/structure, my motivations and the external constraints given (is it raining? can I fly? when do I need to pay the rent). Even my motivations are largely determined by my genetics - avoid pain, strive for pleasure. I definitely hope that my choices are not random.

How does morality come into play, if our choices are deterministic? It doesn't - my morality is my internal concept that I use to make choices more quickly/easily. I don't impose my morality on other people and I don't really care how they make their choices, but I support different forms of punishment that modify the incentives of other people so that the society can function.

Finally, I don't think that the future is predictable, even though it is deterministic. Like you don't know what 1048936701349 * 13046871435 is before you calculate it, like even the computer cannot predict the result before calculating it (i.e. the fastest calculation algorithm is also the fastest prediction algorithm), the same way we cannot predict the future before it happens, i.e. before the universe "calculates" it.

jey• 13 years ago

You seem to be assuming that "random" must mean "uniformly at random". A weighted die that 99.99999% of the time comes up "3" is still random, but not uniformly random. The outcome of the die roll is nearly certain.

hmsimha• 13 years ago

I think you misunderstood tomp's message. Certainly they didn't mean to imply there was the chance our actions are uniformly random. Rather, the question seems to be whether there is a truly random component to our decisions. The dichotomy lies in the contrast between determinism and the possibility of a truly random quantam process.

manmal2• 13 years ago

I think that the brain is influenced by what is often called true randomness (e.g. radioactive decay). If neurons are influenced by true randomness in our environment, then the way they fire is not purely deterministic. But I'm not so sure anymore that radioactive decay and the like are purely random. Are they (because of quantum mechanical effects)?

vortmeester• 13 years ago

As best anyone knows, quantum mechanical effects (like radioactive decay) really are purely and objectively random -- i.e., it is not the case that they appear random because of our ignorance of some deterministic processes. The not-actually-random scenario is called Local Hidden Variable Theory [1], and is ruled out by a series of remarkable experiments based on Bell's Theorem [2].

You can, if you are stubborn, create a theory of QM with non-local hidden variables, but it turns out to be just as weird as orthodox QM; but instead of accepting objective randomness, you have to accept faster-than-light interactions. One example is David Bohm's "pilot wave" theory. [3]

[1] http://en.wikipedia.org/wiki/Local_hidden_variable_theory

[2] http://en.wikipedia.org/wiki/Bell%27s_theorem

[3] http://en.wikipedia.org/wiki/De_Broglie-Bohm_theory

nooneelse• 13 years ago

Just to bolster your point that neurons can be influenced by radioactive decay (in case a reader were wanting to dispute any direct influence exists): a person with a Geiger counter can make a decision of any kind based on the parity of a count produced by it. "Even I turn left and go to work, odd I turn right and go to a park."

benpbenp3• 13 years ago

If my choices are not deterministic, they must be random.

But this is not what is classically understood by free-will, and I suggest that it is your dilemma which is false. Quoting Aquinas, "Free-will is the cause of its own movement, because by his free-will man moves himself to act."

I anticipate that free-will thus understood is likely ruled out a priori by your philosophical disposition. Nonetheless, I think you can perhaps come around to see that is logically opposed neither to determinism nor randomness in the human brain as observed from an objective standpoint.

aninhumer3• 13 years ago

As I see it, if you consider the universe to be a series of states[1], either the next state is perfectly determined by the previous one, or there is a (finite) set of possibilities for the next state. The latter therefore exhibits randomness.

I'm happy to be challenged on my philosophical disposition though, because I'm aware that there are whole schools of thought that seem like vague informal nonsense to me, and I'm not so arrogant as to assume I'm simply smarter than them.

[1] I don't understand relativity well enough to adapt my argument to it, so I'm happy to accept a counterargument there.

lmm2• 13 years ago

I've just never heard a notion of "free will" that makes enough sense to even ask whether it's true or not. Unless you're talking about some sort of Cartesian duality where there's a nonphysical soul that determines your body's actions - which is at least a coherent idea, but not supported by the evidence.

spenuke• 13 years ago

I'm interested to see you show that determinism vs. randomness is a false dilemma. Not that it isn't - it's just that you haven't explained HOW it's a false dilemma. You just skate right onto the issue "more important for you".

Along the same lines, how is determinism vs. randomness any less of a false dilemma? You say "if my choices are not deterministic, they must be random." But the concept of free-will is another (well established) alternative to determinism.

We have already entered into an old metaphysical forest here, so why not go ahead and think about what exactly randomness is? From the way you describe the "more important" dilemma, randomness is a sort of determinism - when viewed at a systemic level. That is, we often speak of randomness in the terms of probability, as though the mathematical probabilities of a phenomenon are the a priori governing laws that determine that phenomenon's taking place.

An alternative way of seeing random probability is that the laws governing a system are simply unknown to us - though they are determined. We can calculate a-posteriori probabilities of a phenomenon, but it is not the probability itself that governs the system. This is the crucial point.

For example, assume that humans have free will. My favorite sandwich shop serves Reubens on average once every 18 days. But let's further say that the shop owner has some idea to sell Reubens 20 times a year.

Whether the universe is truly governed by random probabilities, or whether the real will (shop owner's) governing the system is just a black box to me, the calculated probabilities look the same.

So I suggest perhaps that yours is also a false dilemma. There seems to me no necessary implication of non-determinism and randomness, because I view randomness from an epistemologically humble position, i.e., that it is an a posteriori description of our lack of understanding.

Maybe in so much as randomness solves the false dilemma of free-will vs. determinism, free-will solves the false dilemma of randomness vs. determinism. A trilemma, perhaps?

poorelise2• 13 years ago

Reminds me of Comed-Tea in "Harry Potter And The Methods Of Rationality":

"If you drink it, something surprising is bound to happen which makes you spill it on yourself or someone else. But it's charmed to vanish just a few seconds later"

eli_gottlieb• 13 years ago

Except of course, that was done via a cheap trick with minor prescience: the tea transmits an urge to drink into your mind when it detects that something funny is about to happen.

poorelise2• 13 years ago

You might want to add a spoiler warning to that, as the actual mechanism behind it is not immediately revealed in the story :-)

Still, I don't see why the Predictor couldn't work in the same way - how would you determine that it doesn't?

pyalot2• 13 years ago

Cool idea for creative writing class, would've helped in plausibility if they'd thrown a few nano and quantum randomly into the essay instead of "negative time delay circuit".

warcode• 13 years ago

I'd say this touches more on the mechanics of time travel than free will.

So if this "predictor" is nothing more than a circuit that sends information to itself in the past, there must be a "me" that pressed the button without seeing the light, before the data is transmitted to the past to the "me" that then sees the light but at that exact point splits off into a separate timeline. And then I would have free will again.

Of course I am assuming time travel where travelling forward in your own timeline is impossible, something the article seem to not really go into.

wodow• 13 years ago

My predictor says that some of you will want to read The Hundred-Light-Year Diary by Greg Egan (in Axiomatic [1]) ... if you like this kind of thing.

[1] http://en.wikipedia.org/wiki/Axiomatic_%28story_collection%2...

afthonos• 13 years ago

What I really hate about these sorts of stories and attending discussions is that we all pretend we know what "free will" is, when no one has ever come up with satisfying definition (just look at the Wikipedia page). My humble opinion: all hand-wringing about lack of free will should be put on hold until we can answer the following question:

In what way will an entity that has free will behave differently than both a fully deterministic entity and an entity that is deterministic excepting occasional random events?

I have yet to see either an answer to that question, or a definition of free will that leads to one. Until then, I consider all speculation and meditation on the "revelation" that we have none to be a tale full of sound and fury, signifying nothing.

Kronopath• 13 years ago

The question of "Do we have free will?" always becomes a lot less confusing if you question the question itself, by asking what you mean by "free will" in the first place. There are some nice posts on Less Wrong that deal with this, they're worth a read if you're curious:

http://wiki.lesswrong.com/wiki/Free_will_(solution)

super_mario• 13 years ago

Just because you can predict what simple decision I will make before I am aware of it consciously, does not imply I don't have free will. All it says is that my conscious awareness of some other part of my brain making the decision is delayed by x milliseconds i.e. we are not making decisions in the upper layers of consciousness, but below it.

When you think about it, it is what I would expect. Your awareness is the product of nerves firing, so they have to fire before it can be reflected in your consciousness. If you decide to recall some information, the neurons storing that information have to be activated (they have to fire). If you can create a device (and we have) to detect those nerves as they are firing you will know what person is trying to recall before (by some miniscule lead time) they are aware of it.

nichtich• 13 years ago

Why is greater than one second is a problem? You can chain many predictor together using robots that when seeing a flash from one predictor presses another predictor. That way you can move the message back in time many seconds.

What's expected of us

Related Stories