But great, that you found interesting too. Keep having dialectic discourse and reflecting about the topics until the truths emerge out of the pure reason just like Socrates and his interlocutors had done, seems still one of the best ways doing philosophy. I must thank you for that. cheers. — Corvus

Maaan! This "guy" has a way! Pure reason (something I despiced before) are a shining beackon now. Dialectic discourse made clear in practice! He truly deserves a medal. And then to think Popper was assigned the title "Sir"... The real sir resides here. A tear leaves my eye. NO irony involved!

What is an interlocutor has become clear now too! Gee, you should have been a philosophy teacher! But enough but licking now. I'll give a kiss only.

What is relevant to OP with the analogue devices would be their in/outputs being continuous voltage rather than digital 0/1 bits, that is same with the human brain. — Corvus

You have answered my question! — Prishon

Not really. Corvus is making a similar mistake that Hermeticus made in the first page of this topic; he's conflating analog and digital signals with analog and digital computation.

As a generic example, I offer the internet. In particular, I'm connected to the internet via a cable modem. The cable modem I use communicates using QPSK (quadrature phase shift keying); in this scheme, four symbols are communicated at once over a carrier wave (sine wave) by shifting the phase of the wave. QPSK signaling is digital; however, the carrier uses continuous values. The point being, you can't just assume that since you're measuring continuous values, you've got analog computation on your hands.

Neurons do indeed have continuous valued voltages, but they aren't wires or electric circuits... they're living cells. Neurons communicate using an all-or-nothing principle; they basically either fire or don't fire. A neuron that signals another neuron is firing, and in the act of firing, the neuron releases neurotransmitters along its axis to the next neuron. A neuron that isn't firing just isn't doing that, and in not doing that it does not send signals to the next neuron. These signals involve for the most part sodium and potasium ions, and it is the distribution of said ions that generate the electric charges you measure with probes.

At this level, computation requires transmission of signals, not just having them. My QPSK cable modem uses its continuous signals to communicate 4 symbols, which are decoded/encoded into the 2-symbol values as they start to travel along my network (initially over ethernet). The continuous-valued-nature of the carrier wave is irrelevant, because what actually makes it across is the 4-state symbols. For an example the other way, Hermeticus showed a nice stereotypical square wave, which was presumed to carry the bits 101010 (I'll cut it off there; annoying doing six pairs)... but if I were to use a generic scheme that looked like that to carry digital signals, the wave might be carrying 110011001100; or maybe even just 111, depending on the exact times I'm supposed to be getting the signal out. Neural signaling is even messier, because it doesn't seem to follow nice clean clocks all of the time; how long between neural firings does it take for the neuron to represent three 0's in a row? What if it fires a fraction of a way into the fourth? How does it signal two 1's in a row?

Despite the discrete signaling of neurons, the fact that they fire at some frequency allows for the possibility that there is indeed something analog going on; only, not with the "voltage" as Corvus is oversimplifying it, but with the frequency. The general problem of how neurons communicate signals is referred to as neural coding.

We can sketch how things work by looking at particular areas. For example, the cones in your eyes generally transmit signals based on opsins absorbing photons. There's a particular probability that an opsin will absorb a photon and kick off the cascade, based on the opsin and the frequency of the photon. What this generally means is that if more light is present, more of those opsins will absorb it, and therefore there will be more cascades involving the cone signals. But again, this doesn't trigger continuous signals coming out; rather, it modulates the frequency at which the cones fire. Those signals come out of the cones and travel down through the optic nerve; so it must be true that at least at some points along the brain, the intensity of the signals the cone measured is encoded more or less in the firing rate of those signals. But note that even if this is an apt description of how color perception works at a signaling level, it's inadequate to establish how the brain works overall.

Finally a second answer actually adressing the brain question! But I have to say that Im on the side of Corvus still. Grainy as the currents in the neural network might be there is stil an analogue process flowing on the network. Massive parallel non-externaly-driven flows representing external processes. Or externally driven by the senses. The huge variety of possible flowpaths creates the opportunity to represent virtually all processes in the universe. The number of possible paths is about a 1 followed by 10exp30 zeros!

But I have to say that Im on the side of Corvus still. — Prishon

Not sure what you mean by Corvus's side.

"What is relevant to OP with the analogue devices would be their in/outputs being continuous voltage rather than digital 0/1 bits, that is same with the human"

And he even had the decency not to reduce people to the brain: human. I dont like choosing sides though.

What is relevant to OP with the analogue devices would be their in/outputs being continuous voltage rather than digital 0/1 bits, that is same with the human — Prishon

So, am I typing on a digital computer? My QPSK cable modem uses a continuous signal.

So, am I typing on a digital computer? My QPSK cable modem uses a continuous signal — InPitzotl

No, you are not typing on a digital computer. You are typing on a dial board.

Let's try it this way. Imagine a machine performing this toy calculation:

...to compute (A+B)×C.

Let's consider two scenarios. In each scenario, there will be two cases. In case 1, we compute (2+3)×4, using A=2, B=3, and C=4. In case 2, we compute (3+3)×4, using A=3, B=3, C=4.

In scenario one, we will use a toy analog computer that works with continuous voltage. In case 1, we pump 2v into A, 3v into B, and 4v into C. Via this value encoding, and by definition of an adder, our computer's adder must then produce 5v at (A+B); and by definition of the multiplier must produce 20v at (A+B)×C. In case 2, we change the input at A to 3v. The result of doing this must then produce 6v at (A+B), and consequentially 24v at (A+B)×C.

In scenario two, we will use a toy binary computer. Binary computers only have 2-value symbols, so we'll have the inputs be strings of 5 symbols, and let's label the two values 0 and 1. So in case 1, we pump in the string 00010 into A, the string 00011 into B, and the string 00100 into C. Via this value encoding, and by definition of an adder, our computer's adder must then produce the string 00101 at A+B, and produce the string 10100 at (A+B)×C. In case 2, we change the input at A to the string 00011. The result of doing this must then produce the string 00110 at A+B, and the string 11000 at (A+B)×C.

For the machine to work in scenario one, it's insufficient that the input signals on A, B, and C are continuous values; and that the output signals on (A+B) and (A+B)×C are also continuous values. The components absolutely must be capable of distinguishing the input values to affect the output values as required by each computation.

If it turns out that the 2v versus 3v inputs into A don't affect the adder's output at all, then you can't possibly compute addition, or multiplication, or any function that depends on A (aka produces different results for 2 and 3). Likewise for scenario 2, it's insufficient that you are able to encode strings and send them to the adder; the adder absolutely must be able to distinguish all 2^5 values of the strings.

So back to the neuron case, it's firing at different rates, let's say. Okay. But the resulting signals go from neuron to neuron. For it to use 200Hz and 202Hz as distinct values in computations, it is absolutely necessary that there is something in the network that can distinguish those two values to produce different outputs; otherwise, the fact that they are continuous inputs is completely irrelevant.

Now, as I've sketched out before, there probably is indeed a significance here, as in the color perception case. But you cannot derive this simply from the fact that you found some analog inputs going into the neurons, and analog outputs coming out of them. If the neurons don't distinguish the values when reacting, the neurons can't compute using them. The processing of values is critical in a computer.

Damned! I contemplated your answer. Like this the brain looks more like a digital one... Or maybe a quantized analogue one.Thanks for the lecture!

Like this the brain looks more like a digital one... Or maybe a quantized analogue one. — Prishon

Analogue signals can always be converted to digital signals using ADC within the system. I don't see much significance in making and emphasising difference in their nature here.

What can be measured from human brain via monitoring instruments is analogue voltage signals, not the digital signal. Maybe the workings of the brain can be explained in digital signal forms.

Whatever the case, it seems human mind cannot be reduced to the workings of the signals. But the signals could be converted to replicate / emulate some of the functions of human brain.

Whatever the case, it seems human mind cannot be reduced to the workings of the signals. But the signals could be converted to replicate some of the functions of human brain. — Corvus

It took a while to get a grip (I have read this sentence over and over again, with different emphasis, even loudup, th annoyance of my wife...Women...). But now I have that grip its crystal clear, another example of how concourse is done). If you have not eaten yet: buon appetito! ☺

It took a while to get a grip (I have read this sentence over and over again, with different emphasis, even loudup, th annoyance of my wife...Women...). But now I have that grip its crystal clear, another example of how concourse is done). If you have not eaten yet: buon appetito! ☺ — Prishon

:100: :up:

Thanks for the great OP - very interesting, and I learnt a lot via the thinking and dialectic process.
Just returned from lunch. Thanks. You too.

There are high level aspects of brains and computer that are analogus. For instance they might be the only general information processing machines in the universe. But to pretend that the brain is running Java is just nonsense.

There are high level aspects of brains and computer that are analogus. For instance they might be the only general information processing machines in the universe. But to pretend that the brain is running Java is just nonsense. — hypericin

:100: