Print

Please login or register.
1 Hour 1 Day 1 Week 1 Month Forever
Login with username, password and session length

[SirQuiamus]

Oh, sure. I could certainly be wrong, I just haven't seen any actual refutations that prove it. That's all I mean to say.

Man, I'm having religion thread flashbacks. Not you specifically Bauglir, but this whole argument.

Your flashbacks are right on the money. The internet has ushered in a New Dark Age, and transhumanism is 21-st century scholasticism.

Heh, you're not wrong. This isn't that different. I try not to take these things too seriously anymore, but I actually really do want to see this refutation. I'm not too hung up on proving myself right, I'm just trying to be clear about what I believe and why so that there's a very clear target to demolish.

To summarize: I believe that there's a hardware/software analogy. I believe minds are a naturally-occurring analogue to software. Because hardware and software demonstrably exist in reality, to the extent that anything can be said to, appeals to natural law or ontology are unlikely to work well - a good refutation would most likely demonstrate that the analogy is impossible in the specific case of minds. That said, it's not my place to dictate the terms of somebody else's argument. That's more my expectation than an instruction.

This really, really got me thinking... so much so that my thinky bits are getting inflamed and my writey bits no longer produce coherent sentences. I might try again tomorrow, or save it for a future blogpost five years from now, or just forget about it because it is ultimately pointless academic wankery with no practical relevance whatsoever. :p

[Bohandas]

Food for thought:

There exists a kind of programmable microchip, called an FPGA.  It is little more than a whole bunch of logic gates on a silicon substrate, that can be flipped on and off, to create a simulation of nearly any other kind of chip, on the fly. The most frequent use case is for testing CPU designs. Thus, the very same FPGA can go from being a DSP, to being a GPU, to being a CPU, just with a little extra "software."  The resulting simulation is not emulation; the logic gates in the FPGA are given a configuration that is identical to the gate layout that would be inside those chips, and then it "Simply does" those things. 

Why is this relavent?

Recent research suggests that memories are not stored in the synapes between neurons, but rather inside the neurons themselves.

I'm surprised,. When I clicked on that link I was sure it was going to be that old story about the scientist who had supposedly shown that a worm which ate another worm's brain could gain the other worm's memories (an experiment which IIRC was found to not have been particularly rigorous and which at any rate had alternate possible explanations for it's results regardless.

[MetalSlimeHunt]

Yes, the alternate explanation was that the worms left chemical signals that could be followed.

[TheBiggerFish]

PTW.

[Bohandas]

Yes, the alternate explanation was that the worms left chemical signals that could be followed.

One article I read on it also hypothesized that the training process could have made the first worm produce neural chemicals important to the formation of memories, thus allowing the second worm which ingested these to learn the task faster

[wierd]

There's additional research to suggest that biological neurons do substantially more than just summate the incoming signal data, and that there is active processing hardware inside the denritic and axonal portions of the neuron's cell body. (The synapse is the space between the terminal buton at the end of the dendrite--which branches off the end of the axon--, and the receiving neuron cell body. These structures [dendrite and axon] are inside the myleinated tail of the neuron itself.)

http://www.ncbi.nlm.nih.gov/pmc/articles/PMC1559659/

[Bauglir]

Neat! That suggests neural networks that we want to emulate brains are gonna need another layer of abstraction. Should be interesting to see the field develop. Can't have a neurology discovery without somebody trying to figure out a way to implement something similar in software, or at least that's the impression I get from a very narrow, brief foray.

[TheBiggerFish]

Whurgh.  More AI complexifications, check!

[SirQuiamus]

Oh, sure. I could certainly be wrong, I just haven't seen any actual refutations that prove it. That's all I mean to say.

Man, I'm having religion thread flashbacks. Not you specifically Bauglir, but this whole argument.

Heh, you're not wrong. This isn't that different. I try not to take these things too seriously anymore, but I actually really do want to see this refutation. I'm not too hung up on proving myself right, I'm just trying to be clear about what I believe and why so that there's a very clear target to demolish.

To summarize: I believe that there's a hardware/software analogy. I believe minds are a naturally-occurring analogue to software. Because hardware and software demonstrably exist in reality, to the extent that anything can be said to, appeals to natural law or ontology are unlikely to work well - a good refutation would most likely demonstrate that the analogy is impossible in the specific case of minds. That said, it's not my place to dictate the terms of somebody else's argument. That's more my expectation than an instruction.

Spoiler (click to show/hide)

Oookay... you assert that the mind is analogous to computer software because it is, in your own words, "a system of ongoing processes taking place in the medium of brains," and more specifically "information ... stored in the arrangement and function of its component matter." There are a number of hairy philosophical problems embedded in that definition, but I think the most relevant ones in the context of computing are the concepts of "process" and "function." The gist of your argument seems to be that logical structures like processes are substrate-independent and exactly reproducible in another medium, and they therefore deserve an ontological status of their own – but surely you agree that all processes have to be stored on some physical substrate at all times, be it 1s and 0s or marks on paper? (It would be rather silly to claim that a process rests in the Platonic heaven of ideas whenever it isn’t running.)

You asked whether I think the process of my web browser is physically located somewhere on my computer, and I have to say yes, in the same way as the "concept" (whatever that means) of web browser is "physically located" somewhere in my brain. When the process called Firefox isn’t running, it is logically a bunch of metaphorical "files" sitting in a metaphorical "folder" called /usr/lib/firefox, and those files are linked to certain inodes pointing at certain physical locations on the hard disk. Computing makes a natural distinction between logical and physical structures because binary data on a computer system has to be symbolically represented in order to be human-readable, but this does not change the fact that the magnetic grains on the hard disk platter are real physical things – just like synapses and neurotransmitters – and replicating the logical structure of the filesystem sans data will not reproduce a working web browser, just as mapping the connectome will not reproduce a functional mind. The situation is even more complicated when we are talking about running processes: One could argue that the idea of reconstructing the mind qua process would be like dumping the RAM contents of a computer running Firefox and trying to reverse-engineer the program’s source code – this effort would run into severe underdetermination problems even in the computing medium, and the neurological medium would be much, much worse, I suspect.

You see, the problem is that whenever you talk about this or that real-world process as a "real" object, you are in fact talking about the symbolic abstraction, via language, of a complex network of causal relationships within a physical medium. There are (at least) two reasons for this: First, you do not have all the relevant data about all the parts in that process, and secondly, you have absolutely no direct empirical data of the causal relationships involved. Causality cannot be directly observed in action – it has to be reconstructed after the fact by weaving discrete observational statements into a web of cause and effect, and there is always an enormous amount of leeway between the observations and your causal theory.

I already pointed out how difficult it would be to reconstruct the causality of a running program, but the crucial difference between brains and computers is that in the latter case we don’t have to reconstruct anything. Computers are purposefully designed to accept simple instructions that make all causal connections completely explicit, and the problem of underdetermination is non-existent when you are reading scrupulously commented source code – something that neuroscientists will never (? ? ?) have access to. That is to say that computer software is a poor analogy for the "processes" of the mind because the so-called brain software has to be painstakingly abstracted by neuroscientists, with a nigh-infinite number of equally valid theories to choose from, whereas computer software is already a purposeful abstraction by design, and it obviously has to have a documented and perfectly valid way to construe the processes that are running on it.

Another problem is that even the causality between software and hardware seems to be reversed in the computing analogy, in comparison with the brain: Computer software is unquestionably a series of explicit instructions that cause the hardware to perform the requested operations, whereas the mind qua process could just as well be construed as an epiphenomenon of the neurological brain hardware. It’s often said that consciousness is epiphenomenal and I very much agree with that idea, and I see absolutely no reason why these abstract mental processes could not be regarded as epiphenomenal by-products of physical processes. Because they are abstractions, after all, and they are never observed in nature except in conjunction with physical brain activity.

Computer software is naturally substrate-independent because it is already an unambiguous symbolic representation, and symbolic systems like languages and math are ontologically weird because they exist as social conventions that are seemingly not tied to physical substrates. It would not be too far-fetched to assume that words and programming languages are stored physically in individual brains, but nevertheless, the distributed nature of this massive social system called language gives one a very strong impression of disembodied symbolic "things" floating around in the aether – which is, to be honest, the first explanation that comes to mind when someone talks of substrate-independent logical objects.

I find this topic deeply fascinating and there is much more to say if one really gets into it, but for now, I ask anyone who accepts the software metaphor at face value to ask themselves: "Am I being misled by language here?"

EDIT:
Ian Bogost has written a lot of stupid shit, but I think this article is a rather good introduction into the ways in which technologically motivated metaphors muddle the distinction between abstract and concrete things.

Spoiler (click to show/hide)

It turns out that we have a long history of explaining the present via the output of industry. These rationalizations are always grounded in familiarity, and thus they feel convincing. But mostly they are metaphors. Here’s Nicholas Carr’s take on metaphorizing progress in terms of contemporary technology, from the 2008 Atlantic cover story that he expanded into his bestselling book The Shallows:

    'The process of adapting to new intellectual technologies is reflected in the changing metaphors we use to explain ourselves to ourselves. When the mechanical clock arrived, people began thinking of their brains as operating “like clockwork.” Today, in the age of software, we have come to think of them as operating “like computers.”'

Carr’s point is that there’s a gap between the world and the metaphors people use to describe that world. We can see how erroneous or incomplete or just plain metaphorical these metaphors are when we look at them in retrospect.
...
If algorithms aren’t gods, what are they instead? Like metaphors, algorithms are simplifications, or distortions. They are caricatures. They take a complex system from the world and abstract it into processes that capture some of that system’s logic and discard others. And they couple to other processes, machines, and materials that carry out the extra-computational part of their work.

I think there just might be a commonplace, common-sense lesson to be learned here: "Don't mistake the map for the terrain." (Also known as the fallacy of reification.)

EDITEDIT: Spoilered that wall of text.
EDITEDITEDIT: Added italics because fuck yeah, italics.
EDIT⁴: Still fixing typos and bad grammar.

[dwarobaki]

The book: "The Old Axolotl: Hardware Dreams" by J. Dukaj was the first thing I thought of when I saw the thread. Although discussion here is very "advanced" at the moment, the book is quite good reading about "technological immortality".

[i2amroy]

-snip huge response-

I'm pretty sure I've got what you are getting at, and I've got 2 pieces to respond with:
1) You mention that reconstructing the process without the data would be like "dumping the RAM contents of a computer running Firefox and trying to reverse-engineer the program’s source code". However if this was the case it would imply that there is some part of the brain that can story memories that is not actually in the brain itself, a "soul" more or less. Recreating a single brain would be less like dumping the RAM, and more like dumping the RAM of the computer, the hard drive of said computer, and the RAM and HD of the web server, i.e. every bit of memory and hardware related to the particular program you were running. This would mean you would certainly have much more data (and it might not fit on your puny 4GB flash drive anymore), but at that point because you have all of the data, reverse engineering becomes possible.

2) I can see what you are getting at in the second part; but I think you are slightly confusing something that is extremely difficult to separate hardware from software with something that is impossible to separate (which really don't exist in this universe, at least). Even if the mind is just an ephemeral byproduct, that doesn't mean we can't simulate it, it just makes it a lot harder to find the base logic that the system runs on. As long as there is no "unique true random number generator" included determining what the "mind" does, then we can simulate it. In cases like ephemeral processes this becomes a problem very similar to attempting to reverse engineer a deterministic random number generator and it's accompanying seed from a single set of consecutive generated numbers sampled at a random location; it's a very difficult task (and in some cases is beyond our current technological prowess), but that doesn't make it totally impossible.

To claim it a slightly different way, I could claim that "fire" is an ephemeral byproduct of the physical chemical processes taking place at that location. But that doesn't mean that we couldn't model the chemical processes, use the data from that to model the thermodynamic processes, then use the data from that to model the fluid mechanical processes of the air, and then from that use the data to model the way the fire moves and acts. This was a data process that took us huge amounts of work, on multiple levels, to slowly work our way backwards and down, combining our knowledge as we go, from "ooh, fire, glowy" all of the way down to "the oxygen and the CO₂ combine to produce a net thermal output of X amount, which is then released with this distribution, therefore creating these thermal flows in the air, thus causing the fire to move and look like this", but it's not an impossible process, just a very difficult one to reverse engineer (and one that we are still working on some parts of our simulations today).

[wierd]

This is just opinion, but--

I view the problem this way:  Currently, we have a very very complicated black box, that we want to simulate. The only way we know to go about this, is to break bits of the black box off, and examine them individually.  Naturally, this destroys the whole of the black box in the process, and we have no guarantee that this will allow us to map out the whole of the black box; some of the data is lost via the process of breaking the box apart to examine it.

However, it is possible that we dont need the whole box. A good portion of that data is going to be strictly related to the physical materials the box is composed of; we want to make a new box out of different materials that does the same high level process. This is where the concept of "high level emulation" comes in-- You dont need to slavishly emulate the low level processes, if you can get a grasp of what the emergent property of those processes is, and what its high level requirements are.

Initially, artificial brains will be low level simulations, likely created using surrogate hardware that slavishly simulates all of the components-- because we dont currently understand the high level process we are trying to isolate. However, given sufficient time, resources, and investigation, it may become possible to create high level simulations that require less effort to produce, and still have all the bang, for less of the buck. That's when you have software-only minds.


edit:

But since we are talking about philosophers, and dealing with simulations and simulacra, might as well throw in some baudrillard.  That's at least  somewhat fresher and newer than plato's cave and descartism.

The goal here is simulation, per baudrillard's definition, at least in the short term.  The high level emulation would be a simulacrum-- a construct that now bears no resemblance to any original.

[LordBucket]

surely you agree that all processes have to be stored on some physical substrate at all times, be it 1s and 0s or marks on paper?

When you listen to a television or radio broadcast, what physical substrate do those broadcasts exist on between time of braodcast and time of reception? When the path of light is altered by gravity, on what physical substrate does that path alteration occur?

It’s often said that consciousness is epiphenomenal and I very much agree with that idea, and I see absolutely no reason why these abstract mental processes could not be regarded as epiphenomenal by-products of physical processes. Because they are abstractions, after all, and they are never observed in nature except in conjunction with physical brain activity.

1) No, consciousness has never been (forgive the imprecise phrasing) "objectively or scientificly observed." We don't have any means of measuring consciousness apart from looking at somebody and asking "hey, dude you conscious?" and observing their response. If two rocks bumping into to each other are conscious of that event, we would have no way of knowing that, because rocks don't talk. If a human were a not-conscious zombie, we would have no way of knowing that either, because humans do talk. Give me five minutes and I can write a computer program that will claim to be conscious. Are you going to believe it just because it says it is? I don't think so. Are animals conscious? That's a point of debate. We can only observe behavior. What about animals that don't have brains? Are they conscious? Many of them are animate. But are they conscious? We don't know.

Yes, you can measure brain activity, but if you use that as a measure, then you're assuming your conclusion: that brains correlate with consciousness.

Here's a video of single celled creatures eating each other under an electron microscope. Look at the way the paramecium "panic" and try to escape once they've been engulfed. Does that not appear to be the behavior of an entity that it aware of its environment?

Yet they lack brains in the conventional sense. How do you reconcile this with your position?


2) If "asking somebody" is sufficient evidence, there's no shortage of anecdotes about life after death experiences, including a few actual studies concluding that some people do continue to have awareness after brain activity ceases. And if you hunt around, a few of them involve cases where the clinically dead person was able to recall sounds in the room after their brain stopped. Sounds that the doctors in the room agree did occur after this person's brain shut down. Tough to dismiss that as false/hallucinated memory.

Yes, you can suggest that maybe our instruments are not as good as we think they are, and that maybe brain activity does still continue after the EEG claims its stopped, but again: are you examining evidence or are you assuming your conclusion?

[Graknorke]

When you listen to a television or radio broadcast, what physical substrate do those broadcasts exist on between time of braodcast and time of reception?

The electromagnetic field. Or radio photons.

[LordBucket]

The gist of your argument seems to be that logical structures like processes are substrate-independent and exactly reproducible in another medium, and they therefore deserve an ontological status of their own – but surely you agree that all processes have to be stored on some physical substrate at all times, be it 1s and 0s or marks on paper? (It would be rather silly to claim that a process rests in the Platonic heaven of ideas whenever it isn’t running.

When you listen to a television or radio broadcast, what physical substrate do those broadcasts exist on between time of braodcast and time of reception? When the path of light is altered by gravity, on what physical substrate does that path alteration occur?

The electromagnetic field. Or radio photons.

I think you've missed the context of the question.

An electromagnetic field is "physical" in the sense of "relating to physics." It is not "physical" in the sense of being a "physical, material object."

If consciousness can occur within the medium of electromagnetic phenomenon, then I think SirQuiamus's premise becomes invalid.