The Cabinet of Wisdom IX Emerging from the Emergent Property Argument
The Cabinet of Wisdom
PART IX
Emerging from the Emergent Property Argument
A frequent counter argument takes the form of a word fetish called ’emergent properties.’
The argument in some form, as best I understand it, runs as follows: the shape of the water molecule does not, for a single molecule, display the hexagonal crystalwork of a snowflake.
Nonetheless, the shape of the snowflake is a physical byproduct of the water molecule shape and the Van Der Waals forces adhering molecule to molecule. The snowflake geometry is an emergent property of the molecule shape.
Likewise, no one braincell, in isolation, is capable for thought, nor is a single letter (with the exception, in English, of I and O and A) able to form a word or sentence. But human thought is clearly the emergent property of many braincells acting in concert in the nervous system, and all sentences written in English, including this one, emerge from the combination of letter meaningless in themselves.
Therefore the mere fact that cogwheels or electronic circuits, in isolation, being inanimate and incapable of thought, cannot, by itself, justify the conclusion that electronic brains and artificial intelligence networks, with all parts acting in concert, do not, could not, and will never think as well as a human brain, possessing self-awareness and free will as much as any biological human.
Now, this argument, as far as it goes, is perfectly sound.
The mere fact that the components of a clockwork man with a clockwork brain, like Tik-Tok of Oz, are themselves made of unliving copper components, does not necessarily mean those components when acting in concert do not think.
But the argument being given here is not that argument. No one here is arguing that a thinking being made of components, because he is made of components, cannot and does not think.
A living brain and a dead brain have the same number of components, arranged in the same way, even as the Book of Kells and the Voynich Codex are both made of paper inscribed with ink with lavish illustrations. The difference is that whatever last thought, if any, linger in the motionless braincells of the dead brain cannot be read, and neither can the unknown script of the Voynich Codex.
The argument posits a formal logical error. While it is true that merely being made of components does not necessarily prove that electronic brains do not think, or never could, it is also true that merely being made of components does not necessarily prove that electronic brains do think, or ever could. It is irrelevant to the argument.
An emergent physical property can indeed emerge from a simpler physical property. But a nonphysical property cannot emerge from a physical property at all, simple or complex.
Because the argument given here may be unclear, let us remind ourselves.
The argument given here is that a simple process can mimic the form of human thought. For example, to examine all possible variations of a Hexapawn game, eliminating losing moves, is a mechanical process that mimics a human thought process by mimicking its form. The Hexapawn engine can present a result that will be meaningful to a human observer able to interpret the end form back into the thought associated with it.
The chessplaying cabinet can do the same for chess, even though, unlike Hexapawn, the speed and number of the mechanical operations needed to mimic the thought process is more than a human observer can do in a reasonable amount of time, if ever. The chessplaying cabinet nonetheless is displaying a formal result that a human observer, if he is able to interpret the symbols used for chessmen and knows the rules of chess, can interpret as a move in a game.
The argument given here is that self-awareness is a quality known from the point of view of the self-aware being. We make no comment on how self-awareness arises from non-self-awareness, if it does, but point out that merely adding gears to a chessplaying cabinet or beads and pictures to a stack of Hexapawn matchboxes is not that process, without more. Self-awareness may be a simple thing, or may be complex, but it is not an intentional by produce of complexity in a tool. No matter how many gears or beads are added to cabinets or matchboxes, they do not change their fundamental nature, which is, mechanical parts being pushed by other mechanical parts being pushed by a third, and so on.
Indeed, since it is established, even by the most skeptical, that rocks to not think, therefore beads do not think, nor is a thinking ability granted to them merely when a human observer arbitrarily assigns a symbolic meaning to them, in this case, having the bead color represent a game move.
The act of representation is always arbitrary, that is, always dependent ultimately on human will. Contrariwise, the mechanical motions of a machine, or of the inanimate objects found in nature, are never arbitrary, nor ever purposeful, since the mechanical effects are entirely determined, defined, and controlled, by the combination of mechanical causes operating on them.
An inanimate object, such as a the color of a bead, or the position of the hands of a clock, need not have any physical change made to it in any physical way, in order to have a symbolic meaning associated arbitrarily with it.
In the West, for example, when the longer and shorter hand of a clock points straight up, this is noon. Unless it is midnight. This is true even if the clockface has numbers written in Latin numerals or Arabic numerals or no numerals at all. The medieval monks who decided to put noon at the apex of the clock were mimicking the form of the sun circling the earth, according to the Aristotelian model of the solar system, so the symbol of having the hands be highest when the sun at noon was highest no doubt seemed fitting. But, with equal ease, the monks could have modeled a different scheme, making dawn the first hour (as it is in the Gospel) and placing it at the top of the clockface, so that, at noon, the hand would be horizontal.
One might argue that assignation of the physical properties of assigning symbolic values to inanimate objects does indeed involved a physical change in the braincells of anyone to whom the symbols have been explained, but this has no relevance to the current question, which is, namely, whether or not assigning symbolic values to the motions of animate clockworks assigns or awards them the ability to think. It is safe to say that if rock do not think, assigning a symbolic value to the rock based on color does not grant the rock the power to think, nor if a lump of iron is melted and cast into the shape of a wheel or spring or clock-hand.
Again, and for the same reason, painting or chalking or cutting shapes into rock or paper or any substance those shapes, such as ideograms, hieroglyphs, cuneiforms, Greek or Cyrillic or Latin alphabetical letters, or Roman numerals or Arabic, does not grant the power to think. Indeed, the shapes remain meaningless to anyone illiterate to that language.
Again, in this case, for the Latin letter shaped like a circle, its meaning does not depend on the physical properties of the shape. In other contexts, the same shape is used for the Arabic numeral representing the concept of zero, or the Greek letter omicron, or the astrological sign for the full moon, or the halo above a figure in an icon to represent holiness.
If written in written in chalk on slate, ink on paper or velum or papyrus, carved into a monument of marble for a millennium, or written by a skywriter in white smoke in the wind for a minute, the symbol is the same. No information is more or less carried by it. Zero is still zero. Omicron is still omicron. O is O.
This is because the symbol’s meaning in this case is based on its form. If the form changes, such as if it grows a tail, the O becomes Q, and has a different interpretation.
Symbol can mimic mechanical operations if the thinker so wills for so long as he wills. This is particularly true in matters of mathematics, formal logic, and geometry.
We can arbitrarily assign the Greek letter Zeta to represent the abstract number two and the astrological sign of Jupiter to represent the number four, and the Japanese number two to represent an equality, while using the cross of Saint Andrew for multiplication or the Greek cross for addition. So doing, we can write 2+2=4, or 2×2=4, but arbitrarily could write 2+4=2 and be wrong, or could write 2+2=x and be neither right nor wrong. Because the cross of Saint Andrew can also be used to represent an unknown algebraic value.
Contariwise, an adding machine, when the user depresses the keys representing for two added to two, and pulls the lever to engage the gears, makes no choice and has no choice but to turn the numbers in the display window and display the answer the machinery is designed to display. If designed correctly, the image carried to the display window on the cogwheel will be some sign, in this example, the astrological sign of Jupiter, which anyone who has been told the meaning of this shape knows represents the number four.
And the observer will be delighted that the adding machine did a mental operation of adding up two numbers! The machine thinks! The adding machine has passed the Turing Test.
Indeed, all machines from clocks to calculators pass the Turing Test. That is what is so wrongheaded about the Turing Test. If you ask me the time of day, and I look at the sun and tell you, I have done no more and no less than if a sundial or grandfather clock tells you the same thing. The difference is that the sundial can only show the wrong time if its carefully positioned before a mirror, and the grandfather clock if it is running slow, but I can tell you the wrong time because I am trying to trick you into arriving where the surprise birthday is planned with just enough delay to allow a dancing girl to hide in a hollow cake.
This is because neither sundial nor clockwork are telling you anything, that is, they are unliving physical objects with physical properties only, and you, the human observer, attribute or ascribe symbolic values to certain of those properties, particularly to forms that have unchanging formal existence, but only have temporary manifestation in temporal existence.
The abstract concept of duality in magnitude, more commonly called the number two, is the same for all minds, human or angelic, elfin, or extraterrestrial, no matter where situate, and no matter what symbol, if any, is used to represent that concept. No matter what mind where situate otherwise thinks, if it thinks truthfully, the number two is an even prime; it is both half of four and two less than four; its square root is irrational. These are not physical properties, because, if they were, then some physical change of some physical object could change these properties.
But because truthful symbol follow nature, a symbol that shows twice two is four can also be assigned to a mechanical system, such as beads on an abacus. Lo and behold, if I put two beads to the left and two more, then there are four beads to the left.
A Hottentot, or so I have been told, has no number for any numbers above three, and cannot speak aloud a symbolic sentence to convey that twice two is four. He has no word for four. But he can push two pairs of abacus beads to the left, and see the result.
In his case, he is seeing a mechanical operation represent the form of a thought for which he himself has no apt symbols. Nonetheless, the result is correct. In my case, if I go to a chess-playing cabinet, the turning gears in the machine, by purely mechanical operation, can and will defeat my amateur level of skill. For the arithmetic problem of the Hottentot, as with the chess problem of the amateur, the level of calculation is beyond our mental reach.
This is because, like the Hottentot who cannot count above three, I cannot calculate out every legal chessmove and assign weighted values based on anticipated responses to end each game in a stalemate. But I can do that when I face a tic-tac-toe playing cabinet.
It is, nonetheless, a calculation. Calculations follow the logical forms of mathematical operations. Hence, unliving objects moved according to logical patterns can follow the selfsame forms, and, being interpreted by a living mind, will reach the same result a thinking being could have done, were it able in the time given.
It is beyond question that a simple version of this calculation process, as when using rocks or beads to track simple arithmetical operations, involves no thought on the part of the rocks or beads. The process of elimination of a simple game like Hexapawn can mimic an idiot version of learning, and for this reason is called machine learning. Even the most ardent of eliminative materialists will admit that rocks do not think. Presumably, cogwheels and clockfaces likewise do not think.
Unless the act of a human observer arbitrarily ascribing a symbolic meaning to some physical property of a rock or bead or wheel, such as their color, or position, or marks inscribed on them grants the power of self-awareness hence of thought, we can likewise say that merely by being used as a tool does not grant the power of voluntary thought or decision to animate objects.
If true for simple clockworks, then also true for complex clockworks. If true for formal mathematical processes simple enough for a human onlooker to grasp, such as Hexapawn machine learning, then also true for those too complex and rapid for the human onlooker to grasp, such a chessplaying cabinet machine learning.
By Occam’s razor, if a complex and rapid mechanical process can be described in term of pure unintentional mechanical motion, without assuming intentionality or an entity capable of intentionality, then we should not so describe it.
If, on the other hand, we cannot describe the human thought process without reference to intentionality, nor without reference to an entity capable of intentionality.
Hence the two things, human thought versus complex mechanical motions, cannot be described in the same way, do not fit the same model, or, in other words, are not and cannot be the same.