# Minimum brain size for consciousness?

In the limit, how small can a brain get and still host an identifiably human-like consciousness?

Obviously, they can be smaller than an adult human's, since children (with smaller brain volumes) and patients who lost upwards of half their brain exhibit behavior that we would consider conscious and report experiencing an internal representation of the world through qualia. Furthermore, since human brains are large, noisy and lossy environments, this means that many brain structures have redundant features to get around this noise problem. Smaller organisms could and often do have smaller cells and don't need as much error-correcting duplication.

So are we talking chimp-brain-size? Dog-brain-size? Ant-ganglion-size? To simplify, assume that the biology in question retains opposable thumbs or at least vocal chords to ease communication.

Re: What do I mean by conscious: do other entities we recognize as conscious recognize it in turn as conscious? This might miss on some (or even most) forms of consciousness, but saves me endless definitional hassles.

• To give some context, I'm assuming at least some humans are adopting the Rats survival solution to the Humans as Pets question I posted earlier, with a host of earth-bound AGIs at the top of the metaphorical food chain. – Serban Tanasa Dec 28 '14 at 1:24
• As an aside, you can get really small multi-cellural animals if you are willing to make trade offs. Granted, this example is done by sacrificing neurons (it doesn't need that many) and the nucleus of many of the cells. See the Fairy Wasp. One question is how small are you measuring things? Are you willing to accept the cheela of Dragon's Egg or the humans of Flux? – user487 Dec 28 '14 at 3:09
• +1 for Flux and Dragons Egg. Don't forget Forward's inspiration, A Mission of Gravity by Hal Clement. – JDługosz Dec 28 '14 at 3:20
• The assumption "Obviously, they can be smaller than an adult human's, since children [...]" is wrong. That's because if an adult has a smaller brain, its children will have an even smaller brain. And a species with children with brains so small, might be incapable of producing consciousness. Please notice I'm not saying it's impossibile, I'm just saying that that assumption is wrong. – o0'. Dec 28 '14 at 21:04
• @Lohoris, your argument does not follow logically. Human children past a certain age are recognizably conscious. Week-old human embryos, for instance, are not. There's no reason why a biological entity can't attain consciousness despite not having it previously. In fact, that's what we all do. – Serban Tanasa Dec 28 '14 at 21:14

I don't think the terms "self-aware" and "conscious" are defined well enough to give you an answer. On one hand, we haven't given the title "self aware" to anything smaller than an Orangutang. However, depending on your definitions, you could make a hard sell that a standing wave such as a quantum waveform has self-aware traits.

My answer, if I had to peg definitions for "self-aware" and "conscious," would be that the answer depends on how hostile the environment is. The more hostile the environment is, the less value there is in understanding yourself (you spend all your energy battling the environment). The friendlier the environment is, the more an understanding of how "you" operate becomes an important trait.

At the smallest scales, Brownian motion is a dominant force. Its randomness is going to keep a small organism evolving to deal with its effects, not its effects on itself. However, if temperatures were chilled, it may evolve to be more aware of itsself on its surroundings.

Pack your local area close enough, however, and your neighbors all work like you. Suddenly its really useful to understand yourself so you can understand your neighbors.

Bacteria exhibit interesting behavior in biofilms. They effectively form a quorum, at which point they all agree that banding together is good for the group. They then react differently, "realizing" the power of their combined force is greater than the sum of their Selves.

What about parts of a body? Currently there is belief that the Anterior cingulate cortex may possibly be the "throne" of the brain's consciousness. Is it? We don't yet know. But this does suggest that there's a smaller part of our brain responsible for this funny thing we call consciousness.

Going all the way to the waveform argument, consider a standing wave in a flute. If any air molecule is perturbed, such as from Brownian motion, it falls back in sync, because it "knows" the relationship between its momentum and its position. Or does it? This quickly becomes a philosophy question. What does it mean "to be aware of one's self?"

Philosophy majors will take it all the way. There is a category of philosophies called idealism which all center around the idea that everything is conscious thought. Every grain that makes up the sandy beach is a conscious aware entity. It's a bit extreme, but it points out just how hard of a question you are asking. In the other direction, there is a physicalilst argument that consciousness is simply an illusion, and nothing is conscious, not even us.

Edit: in light of your edit, I recommend two Wikipedia articles:

• Animal consciousness is not a well defined topic. Some tests, such as the mirror test, admit some primates and pigs, but dogs often fail that test (which is counter intuitive for many dog owners). On the other hand, some tests are with regard to the handling of pain, which many animals, including dogs, pass.

• The actual issue you are going to have to grapple with is the Problem of Other Minds. It is an extremely difficult topic, which can go so far as to say "all other humans are merely zombies... I am the only conscious one." I recommend looking into this problem, because your edit suggests this direction: you are seeking to make them "similar" enough to humans to ease identification via the ability to vocalize and so forth.

Because of how hard the Problem of Other Minds is, and how much harder it is to convey this issue as an author, consider side-stepping the issue entirely. Use Sanderson's First Law of Magic to make your life simpler:

Sanderson’s First Law of Magics: An author’s ability to solve conflict with magic is DIRECTLY PROPORTIONAL to how well the reader understands said magic.

Decide to yourself "this creature size is conscious." It doesn't really matter what line you pick, pick a line that suits you and your intuition. Call it magic -- you defined something magical called consciousness. Cool! Now by Sanderson's first law, you are free to call them "conscious" as long as you please, up until the point where you try to resolve conflict with it. At that time, you need to have taught the reader enough about this "micro-consciousness" that the reader understands enough to let you solve the conflict.

Because the understanding of your particular definition of consciousness is narrow at the start of the work, concentrate on showing how it solves problems that the reader would relate to that sized animal. For instance, ants don't like water on their nest. It does really bad things when the nests flood, so they have to take action like move the young closer to the entrance. An ant would be interested in keeping the nest area dry. However, they would not invent robotics to cut down a giant banana leaf to cover the nest, but they would develop gorgeous Mandela of small leaves which interlock to make something watertight, but which can be moved by a single ant. Now we have demonstrated problem solving.

Now lets say there's a particular parasite of a larger creature that is considered "good food" by the ants (for any reason you please). Every now and then, an ant might come across one and get to take it back, but its a lot of work. However, the larger creatures are smart -- they can be trained. Consider an ant colony that will rearrange their Mandela to signal nearby creatures with parasites to come be groomed (or they may arrange their Mandela differently to signify that they have enough food, and would prefer to leave their nest less-trampled). This would be along the lines of the small cleaner wrasse who clean the mouths of sharks. Suddenly we have the ant colony communicating with a greater species, and communication is a strong "walks like a duck" moment in any consciousness story.

At some point you should be able to wrap the reader up into questioning "is the ant conscious, or is the ant colony as a whole conscious?" That is the point where you have them wrapped up in your particular definition of consciousness, and you can start saying "let me explain to you why this is the sound of a duck quacking," and no reader will disagree.

The key to all of this is that you never use consciousness to solve a conflict that the reader does not already understand: and you start from conflicts that could be reasonably resolved by a creature of similar size and move up.

• +1 for the first sentence. How do we define consciousness and self-awareness? I seem to remember some research on sense of self in lizards, but is that human like? – ckersch Dec 28 '14 at 4:33
• I'm going for the easy engineering way: quacks like a duck, walks like a duck... Do other entities we recognize as conscious recognize it in turn as conscious? This might miss on some (or even most) forms of consciousness, but saves me definitional hassles. – Serban Tanasa Dec 28 '14 at 13:04
• @Serban Tanasa: If you look into the study of consciousness, you'll find that the "quacks like a duck" technique is not as effective as we are used to it being when it comes to this topic. Consider Sentience Quotient, which argues that "the appearance of consciousness" (or more specifically: sentience) is related to processing density, not size. There is literally that much disagreement as to what a quacking duck sounds like (en.wikipedia.org/wiki/Sentience_quotient). – Cort Ammon - Reinstate Monica Dec 28 '14 at 17:37
• @Serban Tanasa: I added an edit to my answer. The section on Sanderson's First Law of Magic may help to sidestep you having to define consciousness entirely, while still exploring it. – Cort Ammon - Reinstate Monica Dec 28 '14 at 18:03

What if the said consciousnesses are digitised into a computer?

(Summary: Probably not possible with existing silicon-based computer technology, but very possible with future exotic computing techniques)

With the ever-decreasing size of transistors, it could theoretically be very possible that a future digitial mind will be capable of becoming vanishingly small. Of course, this requires that artificial intelligences be natively programmed, instead of being emulated (which would be significantly smaller and more efficient, but we cannot tell for sure how much smaller).

To do a back-of-envelope calculation

A supercomputer cluster (10 petaflops) achieved a speed ratio of 1:2400 while emulating a fraction (~2%) of the human brain's neurons. Using naive scaling, it would seem that 1.2 zettaflops $(10^{21}$) would be required to emulate the entire brain.

The said supercomputer uses 45nm node transistors made using single-layer silicon lithography. Assuming Moore's Law is broken at 1nm (atomic-sized) transistors, and that this "future CPU" design stacks 1000 layers of silicon, the proposed "future silicon processor" will be approximately 2 million times faster than the supercomputer, which therefore would be able to simulate the human brain with a 16x speedup, or simulate 16 simultaneous human brains. Since even simulating one brain would still require multiple server racks, this would not represent a shrink.

What about computronium? A Nature paper on the ultimate limit of computational miniaturisation seems to suggest that 1kg of computronium would at best be able to carry out $5 \times 10^{50}$ flops (by collapsing it into a black hole). Assuming 1.2 zettaflops are required to simulate a human brain, less than one atomic mass unit ($1.66 \times 10^{-27}$ kilograms) of computronium would be required to simulate the human brain. A native computational intelligence would be significantly smaller.

• Thanks again for the answer! Deserves more upvotes! What would be the cost of running this sim for a year? For something with K's efficiency, I'm getting something absurd like when scaled to 1:1 speed factor. Is Koomey's law a reasonable projection of power efficiency gain? – Serban Tanasa Dec 29 '14 at 22:22
• Why would a computation machine involve consciousness? What a computer does is from mechanical instructions. It would never need a consciousness to do everything it does. And how would a consciousness ever even affect what it does in any way? Do you think having it be very fast an complex will attract a consciousness to drive the program? Why would it have a consciousness, or more accurately, why would any consciousness attach itself to these mechanical operations, except the consciousnesses of humans studying its computations or outside conscious observers of a robot observing its actions? – Dronz Feb 27 '15 at 21:29
• @Dronz This answer isn't saying that a sufficiently fast computer would necessarily become conscious as a byproduct of being a l33t r1g, but that a sufficiently fast computer can emulate, simulate, and run complex systems. If consciousness is such a complex system, then theoretically a computer could run a consciousness as a piece of software, or (as suggested in the answer) could even be specifically constructed in order to run a consciousness natively. – user867 Oct 20 '15 at 0:52
• @user867 Why would a consciousness be a system such as a computer runs, at all (complex or not)? I can model a consciousness to the degree that I understand consciousness, in a computer simulation, but it won't make the program be a consciousness, even if it acts equivalent to my understanding of what one is, any more than programming a spaceship simulation makes a spaceship come into being in the real universe. – Dronz Oct 20 '15 at 4:08
• @Dronz See this Wikipedia article. It is debatable whether this is true, but if you reject it, you would have to accept philosophical mind-body dualism, which is problematic to say the least. – March Ho Oct 20 '15 at 6:15

As a writing exercise (I never figured out the plot details) I wrote exposition on an ET who has gone unnoticed by Earthlings because he is the size of a mosquito.

Looking upmthe complexity of a human brain and estimates of its processing power, I was able to imagine a plausable technology to just barely fit. This would be an uploaded mind or AI in an advanced engineered computer, not natural biology.

Another idea I played with was sentient life on the scale of our single cells. Lacking physical count of parts, they used quantum mecanical processes and lots of time to think: their perception of time was about a day to our year. Nanomachines are rheir natural scale and (e.g.) motors are grown on dometicated lifestock (flagellium motor bread to be more what they want and less natural). The explaination is handwaving, not real engineering, but OK for a story that really treats it as a given.