# How can machine possibly possess consciousness if computation is a human idea?

Suppose a computer simulates a human brain so that it passes the Turing test. Suppose that this brain is given sensory stimuli that correspond to an experience of being in an Earth-like world. From my understanding, proponents of computationalism suppose that this simulated brain would have qualia.

The neural network, its state, and the program code are stored on a computer, and can only be observed if someone knows how to use this computer. How a neural network is represented in memory and on disk, depends on data structures, computer architecture, endianness, et cetera.

Suppose then that the last human on Earth dies, but the simulation keeps running on the computer. If there is no one in the Universe who understands how this particular computer works, then no one would be able to observe this simulated brain. In other words, this brain would be impossible to “decode”, and yet it would supposedly still experience qualia.

It follows that if this is the case, then nothing prevents other “invisible” brains from existing.

We know that a computer is a computational device, and it is running a complex neural network that corresponds to a human-like mind, but the Universe doesn't “know” it. It is our interpretation of what is happening. I can't stray from the conclusion that if the premise is true, then there may be different processes that isomorphically correspond to computations, some of which may isomorphically correspond to minds, but we won't know it because we don't see this isomorphism.

Why couldn't processes inside the Sun correspond to some computation (albeit in a non-straightforward way) which corresponds to a mind? By that logic, there could be infinitely many minds, all experiencing some sorts of qualia, “encoded” in different ways in all processes in the world.

Conversely, what if we replace a computer with a man (with a large lifespan) meticulously following the algorithm on paper. Would his computations somehow give rise to consciousness within a brain simulated on paper?

This just doesn't make sense to me. Do computationalists believe this, or do they have a counter-argument?

• This question might be of interest for a more nuanced definition of consciousness. – Artem Kaznatcheev Jan 4 '14 at 4:15
• "...Would his computations somehow give rise to consciousness within a brain simulated on paper?" this idea is explored in Searle's Chinese Room. Searle says it's not intelligence, others say it is. – obelia Jan 7 '14 at 2:54

# Introduction

Your thoughts seem to straddle panpsychism and computationalism. It is also possible you are just raising a question about physicalism: "if mental thoughts are a result of physical interactions, then why would consciousness be limited to things with brains?". Well, the short answer is that it's fundamentally not, but neither is a fusion reaction and you don't see them popping up all over the place. They only appear (or "emerge", as it were) in specific places where the conditions are just right. So, the physicalist presumes, does the mind. I will also note that physicalism (a form of monism) is basically the mainstream scientific stance.

# Computationalism

Would his computations somehow give rise to consciousness within a brain simulated on paper?

would be an extreme interpretation of computationalism. At it's core, computationalism just says that mental states are grounded in mental representations (semantics) operating on context (syntax): they are part of the cognitive process. Individual computationalists can interpret, to varying degree, how stringent the requirements are on the underlying hardware.

Chomskian linguistics and David Marr's theory of vision were two empirical studies that really influenced the perception of cognitive scientists in favor of computationalism in the 70's. More recent neuroscience literature has favored the connectionist view (note the two are not necessarily mutually exclusive).

A more recent literature review of consciousness, Towards a cognitive neuroscience of consciousness: basic evidence and a workspace framework, summarizes its empirical findings:

Three fundamental empirical findings on consciousness:

1. Cognitive processing is possible without consciousness
2. Attention is a prerequisite of consciousness
3. Consciousness is required for specific mental operations

3. would seem to be consistent with computationalism (the union of computation and consciousness). More support for the computationalist stance:

The free-energy principle: a unified brain theory?

Agents can suppress free energy by changing the two things it depends on: they can change sensory input by acting on the world or they can change their recognition density by changing their internal states. This distinction maps nicely onto action and perception

# Connectionism:

Yet another empirical perspective comes from another physicalist view called connectionism. Connectionism has gained popularity among neuroscientists as we develop technology to better probe brains and model neural mechanisms computationally.

Consciousness as Integrated Information: a Provisional Manifesto

(i) the quantity of consciousness corresponds to the amount of integrated information generated by a complex of elements; (ii) the quality of experience is specified by the set of informational relationships generated within that complex.

Some literature that supports Tononi's idea:

Breakdown of cortical effective connectivity during sleep.

A Theoretically Based Index of Consciousness Independent of Sensory Processing and Behavior

More support for connectionism:

The brainweb: Phase synchronization and large-scale integration

The emergence of a unified cognitive moment relies on the coordination of scattered mosaics of functionally specialized brain regions.

• Your answer does not really contradict mine. I wrote: "From the perspective of experimental science, there is no universally accepted operationalisation of 'consciousness'." (a) Conceiving of the brain as a computer, does not mean that you can measure consciousness. (b) And I don't see how "integrated information" relates in any way to what has been meant by consciousness over the past millennia: awareness and critical consideration of your own thoughts and motivations. That experiment (Casali et al., 2013) [contd.] – user3116 Jan 4 '14 at 10:26
• [contd.] just arrives at probabilities (so there is a large margin of erroneously misidentifying [lack of] consciousness), and the comparative method used only applies to subjects that we know (or believe) can have consciousness: the researchers compared coma patients to people that report (or that we know from our own experience) to be conscious. If we wanted to test a computer or animal with this method, we would have to trust that they knew and spoke the truth when they claimed consciousness. Which is not the same as directly measuring cousciousness. Which confirms my answer. – user3116 Jan 4 '14 at 10:27
• The only conflict is your tone of dismissal. We could go to the physics stack exchange and question their views on gravity using Solipsism, but that's not very constructive. We use empiricism in science, which is why we had Newtonian physics for a long time before quantum physics. And that's fine, we needed Newtonian physics as a foundation to build off of. – Keegan Keplinger Jan 4 '14 at 10:59
• Also, for clarity, with regards to "integrated information", you are cherry picking a follow-up study that demonstrates the clinical significance of Tononi's work. Tononi's original paper on integrated information does indeed mention qualia space (right in the abstract). Its extension to clinical practices (Casali et al 2013) is not a limitation, but a clinical measurement of the theoretical definition. Furthermore, many of your points are acknowledged in the introduction to Tononi's paper (I encourage you to read it, it is free). – Keegan Keplinger Jan 4 '14 at 13:42

Short answer: This is formally known as the hard problem of consciousness - if you can figure it out, then you'll probably win a Nobel Prize:

The hard problem of consciousness is the problem of explaining how and why we have qualia or phenomenal experiences—how sensations acquire characteristics, such as colors and tastes.

Note that this problem assumes that "qualia" are valid constructs, which is not universally accepted (see another question for more on that). All variants of functionalism (including behaviorism, physicalism, and computationalism) have this problem if they accept qualia as valid.

what if we replace a computer with a man (with a large lifespan) meticulously
following the algorithm on paper. Would his computations somehow give
rise to consciousness within a brain simulated on paper?


Ironically, the term "computer" used to refer to a man. There was a time when "computers" were people who ... computed. Alan Turing referred to such "computers" processing symbols on "tape" (paper spools) when he described computing machines in his highly prescient papers on artificial intelligence and machine learning in the first half of the 20th century.

These "paper machines" led philosopher Ned Block to formulate the Chinese Nation thought experiment (1978):

Suppose that the whole nation of China was reordered to simulate the workings of a single brain (that is, to act as a mind according to functionalism). Each Chinese person acts as (say) a neuron, and communicates by special two-way radio in the corresponding way to the other people. The current mental state of the China brain is displayed on satellites that may be seen from anywhere in China. The China brain would then be connected via radio to a body, one that provides the sensory inputs and behavioral outputs of the China brain.

Note that this differs from John Searle's Chinese Room argument, that focuses on whether the system "understands", while the Chinese Brain experiment focuses on whether the machine gives rise to phenomenology (qualia).

We know that a computer is a computational device, and it is running a complex
neural network that corresponds to a human-like mind, but the Universe doesn't
“know” it. It is our interpretation of what is happening.


Speaking of John Searle:

... informational processes are observer-relative: observers pick out certain patterns in the world and consider them information processes, but information processes are not things-in-the-world themselves. Since they do not exist at a physical level, Searle argues, they cannot have causal efficacy and thus cannot cause consciousness.

Note that this argument is by no means universally accepted (example, example).

Why couldn't processes inside the Sun correspond to some computation (albeit in
a non-straightforward way) which corresponds to a mind? By that logic, there
could be infinitely many minds, all experiencing some sorts of qualia,
“encoded” in different ways in all processes in the world.


This has been termed the "triviality argument". There are 4 well-known examples in the literature: Hinck's pail, Searle's (WordStar) wall, Putnam's rock, and Chalmers clock and dial. These philosophers argue that - far from needing a Sun - everyday objects such as walls, rocks, and pails of water have sufficient complexity to encode computationally conscious states.

A similar concept is a "Boltzmann brain":

... a self-aware entity that arises due to extremely rare random fluctuations out of a state of thermodynamic equilibrium. ... in many existing theories about the Universe, human brains in the current Universe appear to be vastly outnumbered by Boltzmann brains in the future Universe who, by chance, have the exact same perceptions that we do ...

Philosophers often use such thought experiments to argue by reductio ad absurdum. However, just because something seems absurd or doesn't make sense, doesn't necessarily mean that it's false (take for example, the Schrödinger's cat thought experiment), so this type of argument has limitations when applied to a mystery such as consciousness. The reality is that we don't know how or if machines can come to possess consciousness (see another question for more on that).

This being cogsci.SE, not philosophy.SE, we cannot simply accept a philosophical definition of consciousness. From the perspective of experimental science, there is no universally accepted operationalisation of "consciousness".

For example, it is impossible to measure consciousness directly. A usual approach is to let human subjects report their thoughts. But of course we do not know if they truthfully report (all of) their conscious thoughts, or if they are maybe even mistaken about what goes on in their minds. Also, we are thus unable to research the consciousness of beings that we cannot (reliably) communicate with, such as animals, children, and disabled persons. What we do is deduce from behavior that appears to us to reflect consciousness that consciousness must be present. But experiments have shown that even human subjects can show such (e.g. purposeful) behaviour while they are themselves completely unaware (i.e. not conscious) of this. The only consciousness that is directly available to a researcher is his or her own, but introspection has been largely abandoned as reliable scientific method more than a hundred years ago, because, of course, from the perspective of other people our reports are not trustworthy for the reasons outlined above.

What follows is that while the cognitive sciences attempt to study consciousness, they have not yet devised of a means to measure it. We, the cognitive scientists, are therefore unable to conclude if a computer, software or the sun have consciousness. Which makes discussions about the status of that consciousness after the disappearance of human beings rather futile for us.

But I am sure philosphers will be able to write whole books on the subject ;-) You might ask them.

• There's actually a wealth of neuroscience articles defining consciousness and making various empirical conclusions. I will post links later. I've been meaning to get to this thread. – Keegan Keplinger Jan 3 '14 at 11:05
• I should also note that computational approach is not exculsively philosophy; it is a prominent cognitive psychology approach (as discussed in the wiki link posted by the OP). – Keegan Keplinger Jan 3 '14 at 14:53
• @What - The cognitive sciences have many approaches and the philosophical approach might not be the most popular one right now, but if undertaken seriously it is of interest. As for computational definitions of consciousness in neuroscience, we already have a good question about that. – Artem Kaznatcheev Jan 4 '14 at 4:17