https://commencement.stanford.edu/traditions/heraldry
-A scholastic hermit.
Thursday, January 24, 2013
Sunday, January 20, 2013
Differences
I've been playing with a cellular automaton simulator. Read the wiki article.
The standard "Conway's Game of Life" uses the rule 3/23, i.e., cells are activated ("born") when they have 3 active neighbors, and active cells remain active if they have 2 or 3 active neighbors. This rule is at a temperate medium between too fecund (every input grows out of control, chaotically) and too sterile (every input dies). Certain engineered structures can replicate themselves (actually a slightly different rule, 36/23) or spit-out new patterns indefinitely. Most patterns, however, tend over many turns to a static system of oscillators and still-lifes.
The "replicator" rule is 1357/1357--born or survive with an odd number of active neighbors. Every pattern is a replicator. Here's an example. The input is a 5x5 square at the center of the map.
Here is the 2nd-to-last step.
And the last step. Nearly all of the material from the previous step has disappeared, which means that, excepting the 25x2 positions constituting the two squares, all of the positions had an even number of active neighbors (remember, the rule is 1357/1357).
The original 5x5 square has produced doppelgangers equidistant from its starting position. This last step becomes the first part of a 64-cycle loop that continually returns to this simple starting-point (I'll call it "step 1"). Here is the 63rd step of the cycle:
The two squares in Step 1 are linked together in interesting ways. Consider what happens when you modify one of the squares but not the other. I've hollowed-out the lower square, i.e., removed a 3x3 square from its center.
I've added a little crown to the difference, a 3x1 bar.
One of many intermediary stages. Notice how many combinations of figures and differences have emerged. Notationally, M is the filled 5x5, N is the hollow 5x5, and f is the modified difference. Below we see M, N, M-N, Mf, Nf and f -- six figures in all.
But by the last step, only three remain: [Mf, Mf-Nf, Nf] = [Mf, M-N, Nf].
Remember that all of this calculation follows from a simple rule governing cell activity. I think that only particular sorts of maps can support this pattern -- the one used here has 192x256 cells, both multiples of 64. The smallest map I've found that works is 6x8 cells. It has a 2-cycle but follows all the same properties as the big map above. Cool stuff!
P.S.
The rule could also be thought of as a boolean calculator. Let's consider a simple 12x16 map with only 1D figures. Each is like a boolean vector. So let's have two figures, A = {TTF} and B = {FTF}. The operation AB is defined as: A XOR B = {a1 XOR b1, a2 XOR b2, a3 XOR b3}, so AB = {TFF}.
XOR is defined for two inputs, either T or F. It outputs F if the inputs are the same, T if different.
So, in our automata map, the middle figure calculates A XOR B:
Indeed, {TTF} XOR {FTF} = {TFF}. Note that A and B have traded places in the process. This inversion is pretty odd, especially since it doesn't seem to happen during regular cycles (i.e., when step 1 is configured as top, middle, bottom: (A, AB, B), respectively. But in fact, the same inversion rule applies across the board--it's just invisible in some cases. This surface variety is a consequence of the properties of XOR.
Let's consider the transparent cases first.
If we take [M, MN, N] and give the bottom figure an additional figure x, then the new initial state is [M, MN, Nx]. The final state will be [Mx, MNx, N]. A technicality here: the "additional" figure x is defined according to the same XOR operation that defines any "XY" relation. So, e.g., if X = {TTF} and Z = {TFT}, and Xy = Z, then y = {FTT}, since {TTF} XOR {FTT} = {TFT}. Or intuitively, the additional element "y" adds to X where X has nothing but subtracts from X where there is overlap. Back to [Mx, MNx, N], note that the additional element x has been transferred from the bottom to the top figure. In effect, the original state [M, MN, Nx] (1) --> [MNNx, MNx, MMN] (2) --> [Mx, MNx, N]. What happened there? I'll get to step (1) below. Step (2) follows the general properties of XOR. It is both commutative and associative, so we can group like terms together. For any boolean vector A, A XOR A = 0, the false vector (since XOR returns F for all like input pairs). T XOR F = T and F XOR F = F, so XOR 0 preserves its first input. Thus, MNNx = Mx, and MMN = N. The first step follows the proposed generalized transformation: [a, b, c] --> [bc, ac, ab]. In other words, the automata game calculates all possible XOR combinations on the map. Let's try it on the other cases.
The input [M, 0, N] will return [N, MN, M]. Explanation? The generalized transformation gives [N0, MN, M0]. Since 0 preserves inputs, the final state is [N, MN, M].
The input [M, MN, N] will return itself. The generalized transformation is [MNN, MN, MMN]. Likes cancel, so the output is [M, MN, N].
One last case. The input [M, MNx, N] will return [Mx, MN, Nx]. The generalized transformation is [MNNx, MNxx, MNMx]. Apply XOR rules and the output is [Mx, MN, Nx].
Here are some examples:
M = {TFFFT], N = {FTFTF}, MN = {TTFTT}. Suppose {M, MNx, N}, where x = {FFTFF}. Note that the vector x does not appear graphically per se, but through its XOR addition to the middle figure. In the first image below, it thus appears as the filled-in slot in the middle of the middle figure (since {TTFTT} XOR {FFTFF} = {TTTTT}).
The whole process should yield {Mx, MN, Nx} = {TFTFT, TTFTT, FTTTF}. Graphically:
A more complicated example:
[M, MN, N] = [TTTT, FFFT, TTTF]
[Mx, MNy, Nz]: x = {TFFF}, y={FTFF}, z = {FFTF}; [FTTT, FTFT, TTFF]
[MNNyz, MNxz, MMNxy] = [Myz, MNxz, Nxy] = [TFFT, TFTT, FFTF]
--A Hermit
The standard "Conway's Game of Life" uses the rule 3/23, i.e., cells are activated ("born") when they have 3 active neighbors, and active cells remain active if they have 2 or 3 active neighbors. This rule is at a temperate medium between too fecund (every input grows out of control, chaotically) and too sterile (every input dies). Certain engineered structures can replicate themselves (actually a slightly different rule, 36/23) or spit-out new patterns indefinitely. Most patterns, however, tend over many turns to a static system of oscillators and still-lifes.
The "replicator" rule is 1357/1357--born or survive with an odd number of active neighbors. Every pattern is a replicator. Here's an example. The input is a 5x5 square at the center of the map.
I've omitted a lot of steps. Here's the 8th, where the original pattern has produced eight copies.
In the course of their replication, the patterns become pretty baroque, but they return in embedded cycles to the simplicity of the original pattern (again, many steps omitted):And the last step. Nearly all of the material from the previous step has disappeared, which means that, excepting the 25x2 positions constituting the two squares, all of the positions had an even number of active neighbors (remember, the rule is 1357/1357).
The original 5x5 square has produced doppelgangers equidistant from its starting position. This last step becomes the first part of a 64-cycle loop that continually returns to this simple starting-point (I'll call it "step 1"). Here is the 63rd step of the cycle:
The two squares in Step 1 are linked together in interesting ways. Consider what happens when you modify one of the squares but not the other. I've hollowed-out the lower square, i.e., removed a 3x3 square from its center.
The 2nd-to-last step of the new cycle.
The final step:
The change from step 1 has been inverted--now the top square bears it. More interestingly, the difference between the doppelgangers has been rendered equidistant between them. That is the 3x3 square in the middle of the map. The system is actually a calculator of geometric difference. It supports a sort of arithmetic. Let's call the initial unmodified 5x5 figure "M" and the modified figure "N." There are 3 positions on the map: top, difference, and bottom. So, respectively, the initial state is [M, Ø, N]. The final is [N, M-N, M]. Figure subtraction produces whatever sub-figure is not shared by the inputs. Modifying the difference imposes identical changes on the top and bottom:I've added a little crown to the difference, a 3x1 bar.
One of many intermediary stages. Notice how many combinations of figures and differences have emerged. Notationally, M is the filled 5x5, N is the hollow 5x5, and f is the modified difference. Below we see M, N, M-N, Mf, Nf and f -- six figures in all.
But by the last step, only three remain: [Mf, Mf-Nf, Nf] = [Mf, M-N, Nf].
Remember that all of this calculation follows from a simple rule governing cell activity. I think that only particular sorts of maps can support this pattern -- the one used here has 192x256 cells, both multiples of 64. The smallest map I've found that works is 6x8 cells. It has a 2-cycle but follows all the same properties as the big map above. Cool stuff!
P.S.
The rule could also be thought of as a boolean calculator. Let's consider a simple 12x16 map with only 1D figures. Each is like a boolean vector. So let's have two figures, A = {TTF} and B = {FTF}. The operation AB is defined as: A XOR B = {a1 XOR b1, a2 XOR b2, a3 XOR b3}, so AB = {TFF}.
XOR is defined for two inputs, either T or F. It outputs F if the inputs are the same, T if different.
So, in our automata map, the middle figure calculates A XOR B:
Indeed, {TTF} XOR {FTF} = {TFF}. Note that A and B have traded places in the process. This inversion is pretty odd, especially since it doesn't seem to happen during regular cycles (i.e., when step 1 is configured as top, middle, bottom: (A, AB, B), respectively. But in fact, the same inversion rule applies across the board--it's just invisible in some cases. This surface variety is a consequence of the properties of XOR.
Let's consider the transparent cases first.
If we take [M, MN, N] and give the bottom figure an additional figure x, then the new initial state is [M, MN, Nx]. The final state will be [Mx, MNx, N]. A technicality here: the "additional" figure x is defined according to the same XOR operation that defines any "XY" relation. So, e.g., if X = {TTF} and Z = {TFT}, and Xy = Z, then y = {FTT}, since {TTF} XOR {FTT} = {TFT}. Or intuitively, the additional element "y" adds to X where X has nothing but subtracts from X where there is overlap. Back to [Mx, MNx, N], note that the additional element x has been transferred from the bottom to the top figure. In effect, the original state [M, MN, Nx] (1) --> [MNNx, MNx, MMN] (2) --> [Mx, MNx, N]. What happened there? I'll get to step (1) below. Step (2) follows the general properties of XOR. It is both commutative and associative, so we can group like terms together. For any boolean vector A, A XOR A = 0, the false vector (since XOR returns F for all like input pairs). T XOR F = T and F XOR F = F, so XOR 0 preserves its first input. Thus, MNNx = Mx, and MMN = N. The first step follows the proposed generalized transformation: [a, b, c] --> [bc, ac, ab]. In other words, the automata game calculates all possible XOR combinations on the map. Let's try it on the other cases.
The input [M, 0, N] will return [N, MN, M]. Explanation? The generalized transformation gives [N0, MN, M0]. Since 0 preserves inputs, the final state is [N, MN, M].
The input [M, MN, N] will return itself. The generalized transformation is [MNN, MN, MMN]. Likes cancel, so the output is [M, MN, N].
One last case. The input [M, MNx, N] will return [Mx, MN, Nx]. The generalized transformation is [MNNx, MNxx, MNMx]. Apply XOR rules and the output is [Mx, MN, Nx].
Here are some examples:
M = {TFFFT], N = {FTFTF}, MN = {TTFTT}. Suppose {M, MNx, N}, where x = {FFTFF}. Note that the vector x does not appear graphically per se, but through its XOR addition to the middle figure. In the first image below, it thus appears as the filled-in slot in the middle of the middle figure (since {TTFTT} XOR {FFTFF} = {TTTTT}).
The whole process should yield {Mx, MN, Nx} = {TFTFT, TTFTT, FTTTF}. Graphically:
[M, MN, N]
[M, MNx, N]
[Mx, MN, Nx]
[M, MN, N] = [TTTT, FFFT, TTTF]
[Mx, MNy, Nz]: x = {TFFF}, y={FTFF}, z = {FFTF}; [FTTT, FTFT, TTFF]
[MNNyz, MNxz, MMNxy] = [Myz, MNxz, Nxy] = [TFFT, TFTT, FFTF]
--A Hermit
Monday, January 7, 2013
Thoughts Conceived during a Period of Deeeep Introspection, by Dermit and Lermit (recently rediscovered notes on Lermit's computer)
As a head beside me bobs up and down, besides itself in its ability to gleefully disable the terrible expiration of timeline. Words tumble in reasonable streams, conveying nothing. Only a clearly-articulated vibration, quelling the scream that is everywhere bursting out and everywhere hidden in two-week-old bedsheets and once-used napkins.
I understand the class' enunciations as I would define the individual words of a dadaist word-salad--each one comprehensible as a solyp, but no meaning is even attempted by the whole.
I imagine great-souledness, and feel that it is not just mere chimera: it is a chimera imagined by a man dreamed by the imaginary friend of a fictional hero of a novel within a novel written by a nom-de-plume in a theater piece by a fake alias of Chekhov, who critics may well say was trying to pass of an unknown Moskovite's work as his own. And my mind is the only thing, in fact, that is holding this entire chain in its fluid integrity, and the symbols on this page.
Chinese people are racist…have fundamental prejudices… Feudal midst still remains, tried to industrialize rapidly to match the material wealth of the west… But missed the crucial step of humanistic enlightenment. Which is why they can't understand hippies and bohemianism at ALL, because it is the black-sheep progeny of humanism (decadence presupposes the disintegration of humanism)
Comprehension of a heat sensation and its puissance over people--if you UNDERSTAND it, at the very least callous indecency will be eliminated in your conduct. Comprehending why Tarrou adopts as his ethical principle COMPREHENSION.
fuzzy vs. sludgy sound. (both applied to the black sabbath tone. but fuzzy and sludgy are two very different things…)
Thinking that someone you don't know is someone you know, switching back and forth between recognition and non-recognition… Two rapidly switching worlds of potential action and obligation.
a book could be written of every individual character in the plague --the prisons, the merchants, the prison guards… Books or works that completely depend on the reader having read something else. But aren't all books derivative? The reader may not have read fhe source material, so a new perspective is laid across it. (Wide Sargasso Sea)
a typographical (calligrammic) philosophical treatise
a philosophical treatise with novelistic or theatrical components (concepts as characters)
Kant. Each link obtains without fail, but the text as a unity is unfathomable or unpersuasive. Duty as the categorical. Is that repugnance to the whole a valid objection? Or is the person being repelled by the whole just completely out of jilt, and our commonplace ways of behaving just completely delusional? The truth is always counter-intuitive?
We don't KNOW anything that is beyond our horizons. We believe in what is beyond them--Reason come into being at the very moment at which we choose to believe in a certain configuration of the world beyond our horizons, at the very moment at which we choose a certain action based on the way we believe (imagine the external landforms, or are persuaded by a certain explanation of how the landforms are) the world to look.
(We only--empirically, certainly, observationally--KNOW things that are in our sphere of perception, our horizons. But we have to know our beliefs, we know what we believe insofar as we need those beliefs to determine the way we act in the world. We don't KNOW our beliefs, but we know our beliefs--this "knowledge" (not KNOWLEDGE) allows an imaginary expansion of our horizons (so we can act like we KNOW--see--what is actually outside them), when in fact we are just--imagining, creating--the landforms that are outside of our real horizons and believing (having faith) that they really ARE like that. That is why beliefs have the weight of real KNOWLEDGE.
God KNOWS all. So it seems like he should be impassive to everything that happens in the world, but humans (limited beings) have always wanted to attribute the mental state of a limited being onto God in saying that his "ultimate plan" is oriented towards some good (good, as in it will be good for myself as a member of the human race or something like that). But to KNOW everything means that every event that takes place is just oriented to an indifferent future state, a non-preferential state of being. Like someone who keeps ants in a glass case. He can see their every move, so the only legitimate position he can take up towards the ants is one of indifferent observation, perhaps curiosity but not emotional investment. But from the point of view of a single ant, that ant has no idea what is going on beyond what he can perceive, so he acts and feels emotions based on what he imagines the external world to look like.
Religions are beliefs, in this structure. That is why religious debates are so deeply entrenched.
Languages are beliefs. "Summer", "Sommer".
Walking around in a empty square, yes there are infinitely many paths that one can take, but none of them are meaningful. You cannot impose a narrative on them, you can't describe it other than "I'm wandering". Bookshelves are obstacles. But you can take a certain path involving many many choices "turn left then right then go down two rows etc.) and that trajectory becomes meaningful…you can describe each step, there is a narrative, a story that can be told about it. When you form relationships with another human being, that's like putting bookshelves in your empty space. Humans are bookshelves… You have to navigate around them but that makes your life more meaningful. (Peruse what titles they contain, too! That's a neat metaphor too.)
Someone who only eats half an apple because he doesn't want to get juice on his hands, but the juiciness of the apple is intrinsic to the enjoyment of it.
A one-night stand involves objectification. Partner as a desirable object rather than as desirable person, because when you find a person desirable that interest ought to last more than one night.
You can create a hypothetical situation in which someone who want to set up an anonymous one-night stand is perfectly legitimate (both parties are aware that this is just a sexual exchange for physical pleasure and possession of a beautiful object) but when you substitute real people… Hypothetical one-night-standers are perfectly fine and even model behavior., but…you know certain things about those people that either support or make hypocritical this decision. Also the nitty-gritty details of having to send a picture, objectification of the self… If one-night stand relationships were to become the norm the most important thing about yourself would be your value as object (physical attractiveness).
To see a crystal globe in any correct way, you have to close one eye. Double vision--two viewpoints… blurs the thing that is being perceived. Same goes for Kierkegaard's pseudonyms… They are points of view that are all valid when taken singularly. But taken as a whole, a man is never a single point of view because he is free and embodies/can embody a whole host of viewpoints.
(Ironic existence…)
4am. Hour of the Wolf. The plague.
Simplicity creates complexity. Complexity creates simplicity.
Complicated philosophical/scientific systems are devised to assert simple things. Conversely, Valéry's conception of poetry is the ultimate simplicity (diamond metaphor)
The opening line of a book like Grapes of Wrath is enough to set astir all of one's voluminous mess of expectations regarding the crystal genius of the author.
blue icons glowing on my screen
signals on a life-support machine
just glow
without a flicker
blue like glass eyeballs
steady, indifferent
while
fretful judders of impatience
refluxed acid eating into my innards
and bilious swollen fingers
that hover,
waiting,
for signs of life.
--Two Hermits Shaking their Heads at Themselves. (Sigh)
Subscribe to:
Comments (Atom)