A year ago, I stopped eating meat and became a lacto-ovo vegetarian. This year, I’ve stopped consuming eggs, and for the past few days I’ve observed a few neurons in my head engage in an increasingly decisive debate on veganism. When the argument on vegetarianism is resolved in favour of non-human animals, it is hard to defend battery farming or the condition of dairy cows. That said, it is not easy to convince oneself why one should care about an agglomeration of “living” cells. In this post I intend to raise as many issues as I can about the question of the animal, after much motivation from Peter Singer’s views here and here.

Motivation

I was brought up in an Indian vegetarian family, where meat was strictly banned on cultural and religious grounds. On a fine hot summer day, when I was around 9 years old, it suddenly struck me how weird it was that I was standing inside a temple, my hands in the Namaste position, praying for good health, prosperity and warding off demons by chanting Sanskrit mantras in front of a decorated stone idol. For one reason or another, I distinctly remember that day. I didn’t dare confront my parents about the uneasy feeling of “what the hell are we doing”, but the whistle was blown – I never again genuinely wished for anything from God even though I was made to frequent temples. Five years later, at a much too impressionable age, I read Ayn Rand, became a vocal anti-theist picking arguments constantly to my mom’s chagrin, and took the ideas of selfishness and hedonism rather seriously. I started eating meat in secrecy, outside home, and when I left for college I’d brandish a seemingly purposeful Darwinian quip against timid vegetarians at the dinner table : “I didn’t climb my way to the top of the food chain to be a vegetarian”.

To summarize, the point I’m trying to make is that the origin of your morality is supremely important, and if it has anything to do with faith in god or spirit or soul or other collective delusions, or even deference to a culture, I’d urge you to read this or this or this or watch this or this or this, because otherwise I wouldn’t know how to argue with you. Call it prerequisites.

Darwinism

After On the Origin of Species was published, humanity, for the first time, had an opportunity to reflect on the causes of its own existence – by natural selection, and hence was enabled to attempt to understand the game it had been playing. Plants, being autotrophic, don’t care about animals fundamentally, although interesting co-dependence between species have evolved. The first animals that evolved can be viewed upon as defectors in the game plants were playing – finding efficient ways to synthesize food using non-living things. They figured they might eat nutritious plants instead and thus avoid complicated photosynthesizing machinery in cells of their body. Other than a small minority, plants, being far more numerous in number and operating at time scales that are far greater, haven’t retaliated. Inevitably, some animals defected on other animals as well, became omnivorous or carnivorous and diversified. Some became predators and some became prey, all existing in various equilibria. And thus we happened to be. One remarkable fact is that not all animals that exist in equilibrium now have evolved to survive by eating other animals. Herbivores (blue whales, elephants, deer, etc.) are still common. I can easily imagine another world where I, a member of the dominant species, am incapable of surviving on primary producers of food – plants. In that case, the “vegetarianism” argument would be vastly different and presumably more convoluted. It really depends on how many levels of irreversible defecting have taken place for us to evolve.

To summarize, Darwinism tells us that survival of the fittest de facto involves individual members of animals killing others, and that this process eventually lead to our existence. One might be tempted to conclude from this, as Benjamin Franklin did, that since other animals kill each other for survival, and we are essentially animals, we shouldn’t hesitate to do the same. A counter argument, which I first heard from Singer, is that modern biology shows that we share an overwhelming amount of genetic, chemical and linguistic machinery with other extant species to varying degrees and this discovery makes a previously unknown connection between humans and animals, dethroning us from hitherto elevated status we accorded ourselves, enabling us to extend our socially evolved strategies, like empathy, to a wider circle. In any case, when one is using a Darwinian argument one ought to keep in mind the is-ought problem.

Strategies

In our world, there are two major categories of life – plants and animals. For now, let us regard Fungi as plants, although eating mushrooms is not an obvious vegan choice. Until we’ve figured out a way to synthesize all the chemicals we need to survive, in the lab, we don’t have an option of not eating plants. The question really is, how much do you identify your “self” with animals? Who all would you be willing to kill or harm for survival? A possible range of answers include –

Cannibalistic options :

1. Yourself (degenerate case), your clones, your siblings, your parents, your relatives and others closely related genetically – the first ones you’d protect from harm.
2. Your teachers, professors, mentors, guides and others closely related memetically – going forward, as more people identify their selves with their intellect, genes will begin to matter less.
3. People of the same race or culture – Comparisons have been drawn between animal rights and the holocaust. (wiki)
4. Women and children, other human beings – In any sexually reproducing species males are cheap and expendable. Two fold cost of sexWomen and children.
5. Humans killed in car accidents – Peter Singer says it should be acceptable to eat human road kills as long as one has permission of relatives and also if doesn’t incentivize killing in accidents.
6. Anyone who wanted to be eaten after dying – PETA president Ingrid Newkirk wants to be barbecued when she dies. Her will.

Non-Cannibalistic options :

1. Your pets, dogs, cats, chimps and other apes, dolphins etc. – Why was The Cove so sensational? Taboo foods vary considerably across cultures.
2. Cows, Swine, Poultry, etc. – For a detailed exposition of factory farming, watch Earthlings here.
3. Oysters and other creatures that do not have a central nervous system arguably don’t feel pain.
4. Eggs, Milk, Honey – Animal products that don’t necessarily result in killing. Separating vegetarianism from veganism.

The reason I list them out here is first, to remove any “shock” value from cannibalism – it is really not that different from eating meat. The second is to highlight the wide range of mutually inconsistent ethical and moral choices different cultures have made and how irresponsible one would be to make a choice purely on whimsical grounds (I’ve often heard people say “Aww, that’s cute. That shouldn’t be eaten.”). The third is to arrange the options in a somewhat decreasing order of cruelty in some sense, to point out that there is a continuum. From a utilitarian point of view, eating a human road kill is comparable to eating a slaughtered cow, if not better. Drawing rigid lines are purely for psychological efficiency – to avoid having to re-evaluate your stance every time you have a choice.

Suffering

Before we discuss why caring about any of this matters, it would be useful to explore the things that could matter. In the most broadest sense, pain and suffering can be viewed as a physiological response to a change in the internal or external environment that could be a potential threat to existence.  In humans, we can accept that some things are universally painful, ex. chopping off someone’s hand. For the more sensitive readers, the slight feeling of revulsion (“ugh”) when your neurons processed and mentally played out chopping of a hand is a form of pain too. A second order pain, if you will.

It is convenient to think of a brain as a buffer between input and output; one that processes and predicts the consequences of various combination of events, and eventually results in a reaction. It fast forwards evolution – instead of producing individuals and letting natural selection slowly affect changes, brains expend neuronal pathways to predict the future of the several probable evolutionary directions, and pick the one most likely to result in survival. If an individual is forced to choose a path that is not in its best interests, it will feel pain. The thesis is that for any consistent definition of pain one would come up with, the amount of pain felt by an individual increases with the predictive power of its brain. It is in this sense that I agree with Singer – terminating a severely disabled infant causes lesser pain than slaughtering an adult cow. To explore this further, let us bisect.

Atoms feel no pain. If you want to challenge that, you’d need an extremely broad definition of pain to claim for ex. that radiation is nuclear pain; that all radioactive nuclei tending to the more stable iron nucleus is an expression of pain. Similarly, stones or amino acids synthesized in the lab can be considered to feel no pain. It is not that I’d like to not be abstract as I can be, but theories become intractable and useless at the highest generality. Moving on, do viruses, bacteria, protista and other microorganisms feel pain? They are more a part of our environment than products of it, and other than ceasing to exist there is very little we can do about it, and even then we barely matter.

What about plants? This is an interesting case. Plants compete with each other for sunlight, they form codependent relationships with insects, self immolate to create forest fires to get rid of competitors ( I saw this in Private life of Plants ), and do many other interesting things to survive. When you chop a tree, is there pain involved? They don’t have a nervous system, don’t have a strong reaction when you run an electric saw through the bark ( although some do repair small damages ), and above all do not immediately retaliate. If there is a pain, it is a different kind. It is the pain of “knowing” that you will die without reproducing; for e.g an infertile human couple who really desired a baby, or a race horse unable to run fast enough to be selected for by humans, and so on. The pain is not physical, and perhaps need not even be mental ( Steven Pinker’s brain wishes his genes to go jump in the lake ). It is a genetic pain, not just felt by an individual but shared through the ancestry. China’s one child policy is an example with several criticisms.

Humans have only recently become developed enough to afford the luxury of avoiding constant physical pain and there’s still a long way to go. So, accounting for genetic pain will have to wait. The same reasoning applies to fungi – they’re as related to us as plants are. On the other hand, in the animal kingdom where we belong, as you move up the complexity, pain experienced by individual animals is very similar to our experience – it is felt in the body. We know this by looking at neural correlates, physical responses, and by an even more fundamental reason that we share our evolutionary history with them. We retract our hands instinctively when placed in a fire, and so do animals when their paws are introduced in a fire – it would be most surprising if the sensation that we feel is of a completely different nature than what they feel.

To summarize, I’ll just say that it is important for us to realize the difference between the kinds of pain experienced by animals and plants, and recognize the similarities between pains all animals can feel. Why should we care if some other individual feels pain? In the next section…

Who are you?

Having lived in Canada for more than a year now and interacting with local people raised in the Western culture has brought glaringly to my attention a deeply rooted and highly cherished notion of the self that each individual here possesses. I’ll go out on a limb here and make an observation, perhaps much too general, that sometimes I find individuality invoked for its own sake and justified with socially accepted norms on matters such as fashion apparel, party or wedding invitations and many other seemingly superfluous (or refined depending on your outlook) tastes in food, music and wine. I’m highly critical of the conformist attitude of the Indian culture as well, and the process minimizing my cognitive dissonance in understanding both cultures made me come up with a definition of self that I find useful. Think of yourself as a weighted sum of various levels that make you up, starting at the atomic level ( C,H,O), to the genetic level (selfish genes), to the mammalian level (social/family structure matters), to the memetic level (your mental picture of the world), and to the cultural level (where you realize you’re just an instance, playing a role in a larger game ), and all the levels in between where you think selection is taking place.

$\vec{self} = a_1*atoms + \dots + a_n*genes + \dots + a_m*memes + \dots$

This is clearly motivated by Quantum Mechanics’ wavefunction, but I’m not suggesting in any way, that atoms, genes, memes etc are linearly independent of each other; in fact the “higher” vectors wouldn’t exist without the “lower” ones. At each level where Doug Hofstadter’s strange loop occurs, or Daniel Dennett’s leveling up in design space happens, perhaps one would like to introduce a new basis vector increasing the “Hilbert” space. Before I proceed, I should make it clear that these analogies are in no way meant to be an actual model of life, and any pretence of mathematical rigour is purely for amusement and is not to be taken seriously. That said, I will take the liberty of lifting the veil of caution and say that I’ve found this definition rather useful.

When my sexual urge acts up, one part of my brain realizes “Ah, $a_n$ is quite big indeed”, and for people like Pinker who don’t care about procreating, $a_n$ must have attenuated in their evaluation of self. Now, how does one value oneself? If $\vec{s_1}, \vec{s_2}$ be two self vectors, then $\vec{s_1} - \vec{s_2}$ tells you how different they are. So, a person feels unique by diffing his self vector with one that comes closest to him/her, and his/her objective is to estimate the probability of survival of each of the basis vectors and maximize it’s share. So, how do you go about increasing your total fitness or prevalence of your self? By increasing each component – working hard to find mates who’s genetic component has higher probability of survival increases $a_n$; working hard to contribute to science increases $a_m$ and so on. Note that your self vector still has a magnitude of 1 in it’s own space, but its contribution to humanity’s self vector, $\vec{s_h}$, increases. For example, I think religion will eventually die out – it doesn’t make enough sense to continue existing even though most of the world is still religious. So, the “direction” in which a self $\vec{s_1}(t)$ proceeds is to maximize $\int_0^T \! \vec{s_1(t)}\cdot\vec{s_h(t)} \, \mathrm{d} t.$ to whatever extent $T$ a self can predict the future. The more you “align” with the total vector summed up over time, the more you “exist”.

To summarize, it is convenient to think of a self, its motivation, perceived purpose of existence and so on, as composed of various forces that operate at various time scales. With this definition, we can understand and express various phenomenon including kinship, family bonds, religion and speciesism. To understand why “you” should care about animals, we should keep in mind what “you” is and why you care about anything.

You and others

Much has been said about the Golden rule : “One should treat others as one would like others to treat oneself”. One common argument in favour of ethical treatment of animals is that we have the ability to put ourselves in the position of animals and imagine their suffering. A favourite line for vegans in defense of killing plants : “I can imagine what it must feel like for a cow to get hurt, but I can’t imagine being a tomato”. But why should one think of someone else in the first place?

Let us look at the simplest example of any introductory course in game theory : Prisoner’s Dilemma (Go through the game if you’re unfamiliar with it! ). In the usual setting of payoffs, (CC)>(CD)=(DC)>(DD) where C = cooperate, D = defect, Nash Equilibrium occurs at DD, meaning a rational player attempting to maximize his/her benefits will Defect – a strategy which is not only not Pareto efficient, but also potentially the worst possible utilitarian outcome (depending on the actual numbers). In real life, people behave differently ( and markedly so in iterated prisoner’s dilemma ). Naive Game theory doesn’t predict or dictate people’s behaviour. I’d recommend reading game theorist Ken Binmore’s easy-to-understand essay on The origin of fair play. Consider what would happen if both players observed the Golden Rule axiomatically. Then, since CD and DC aren’t valid options, the optimal strategy for each player would be to cooperate, and CC is a Pareto optimal Nash equilibrium. Again, we should be careful about deducing or concluding anything about human behaviour from this game, but we can nevertheless make a few observations.

First, people don’t always get stuck in selfish local optimums and neither are they superrationalist observants of the golden rule all the time. They are somewhere in between. One way to model this without restricting the game to superrational players would be to include in our definition of self terms that reflect one’s predilection to care for something bigger than oneself. So, every prisoner’s utility function would have a part that contributes to maximizing his “own” utility ( the individuality ), and a part that cares for both prisoners ( the commonality ). More generally, given a game of n players, any one player’s definition of self will not be independent of others, but rather include those components of $\vec{self}$ that are markedly different from those of others. Computing this difference can, of course, be done in many ways.

• “Best at everything” : For each component, “you” are the difference between your contribution to the component and the one that comes closest to you. $\forall \vec{a_i}, \vec{self_{new}}\cdot\vec{a_i} = \min((\vec{self_{old}} - \vec{self_{other}})\cdot\vec{a_i})$, where each $\vec{a_i}$ is a component and $\vec{self_{other}}$ is your evaluation of “other”‘s worth. You can then value yourself by computing the magnitude of $\vec{self_{new}}$ using some norm (e.g 2-norm). e.g : polymaths are revered because they’re good at some many things.
• “Better than everyone” : Instead of diffing component-wise, you diff yourself with a person that comes closest to you ( e.g an intellectual or your sibling ). You choose to update yourself such that $|\vec{self_{new}}| = \min( |\vec{self_{old}}-\vec{self_{other}} |)$. A situation where a valuation like this makes sense is, for example, a game where you are sure to die but by dying, you can save the life of someone you choose. Since you cannot save individual components, you will have to select a self that comes closest to you. e.g : Lily and James Potter dying to save Harry. Today, caring for ones children over others’ is considered normal, and that is historically based on genetic propagation. Time will come when people will be labeled “childist” in a most Huxlian fashion.
• “Better than average” : As a first approximation, the values of humanity can be considered to be the average of the values of each human. Instead of diffing ( component wise, or as a whole ) with the best, or the one that comes closest to you, you could opt to diff with the average, justification being that your novelty is with respect to humanity as a whole, and not just one person. For example, just because you are $\epsilon > 0$ worse than some other person ( say, in athletics ), doesn’t mean your self worth for that component has to be that small.

The reason I’m elaborately discussing $\vec{self}$ is two fold. The first is to point out that our sense of self, the singular feeling of “I” is just that – a feeling. There probably is an evolutionary/computational complexity explanation for why my feeling of sense of self is extremely localized to the body I inhabit and not “smeared” across many layers like an ant colony. Why do we find it strange that House elves like to be slaves? Or more much interestingly, why we find this bit (watch it!) in Restaurant at the End of the Universe ( read it! ) so damn amusing and oddly unsettling. Every time I have a need to use “I” in a serious conversation, I cannot help but break it down into components and therein find explanations. For example, questions like “Do I have free will?”, “What’s the purpose of existence?” and almost all philosophical conundrums cease to be mysterious eo ipso, when expressed this way.

The second point to note is that $\vec{self}$ is not calculated once and for all times, but keeps getting updated in a rather Bayesian way. The coefficients, or relative importance of components that make up your self, change when new observations are made; artificially reducing your operational Hilbert space to keep the magnitude of your $\vec{self_{new}}$ high, by showing off your superfluous tastes, only serves to flaunt your proverbial peacock’s feathers. However, lets not get into that – sex is a topic for another day.

In short, if there are two pairs of prisoners in a dilemma, the pair of prisoners that redefine their selves to include the welfare of the group over and above their previously lower dimensional individual selves has a winning advantage. I posit that this is one way to look at the increasing complexity of life. In fact, it is now a widely accepted scientific theory that eukaryotic evolution from prokaryotes was via symbiosis – instead of one prokaryote eating the other, it chose to incorporate it into its own cell to form the nucleus. The evolution of life and the corresponding increase in the dimensions of the $\vec{self}$ vector has happened concomitantly with the increase in accuracy of our map of the universe. As we live our lives, $\vec{self}$ gets updated in various obvious ways and some non-trivial ways. Although decomposition of $\vec{self}$ into components does not reveal anything new, it lends clarity to our sense of self.

You and animals

Now that we understand our $\vec{self}$ better, all that remains to be done is to diff our selves with our estimate of an animal’s $\vec{self}$. Lets get some broad perspective first before we compare.

Atoms are 13.7 billion years old (100%), while the earth is 4.5 billions years (32%) old. Soon after the earth was formed, cells arose. Miller Urey experiments demonstrate the ease with which organic molecules form. It is worth looking at the timeline of evolution : prokaryotes at 3.5 billion years (25%), eukaryotes at 2 billion years(15%), fish at 500 million years (4%), mammals at 200 million years (1.5%) and humans at 200,000 years ( 0.0015%). If the age of the universe is not apt for comparison, then consider this : Fish has lived 2500 times longer than humans. We are genetically similar to our closest living apes, chimps, by 98%. Our brains are remarkably similar to those of animals – they evolved incrementally. All big animals, including pigs, cows, chicken etc. feel pain. Their physiological responses, neural correlates are all the same; after all we evolved from them quite recently.

The only thing that separates humans from other animals is our large brain. Copying one another (watch episode 9) – nut cracking, obtaining honey, opening clam shells – started with monkeys. It was these phenotypes that had managed to communicate critical information to one another without relying on the long drawn process of embedding it in DNA. In essence, the word “copying” encapsulates all that we’ve managed to do so far. We build machines to do things over and over again. We’ve figured out that by not defecting in every prisoner’s dilemma presented to us, we can outsmart the other prisoners of earth – we’ve learned to cooperate by copying each other. Other than this fairly singular development in humans above other animals, there is nothing that we don’t share with them.

The question now is – how important is this? In terms of power – very much. We have the dictatorial capability to exterminate most major animal species ( maintaining the essential plants and other things we need ). In terms of evolutionary time – not much. If humans manage to destroy all humans, but spare a few forests with some monkeys or chimps, it is fair to expect intelligent species to evolve in a short time – even a generous 500,000 years is a blink in natural time scales. Humans as a species have become very smart, but the question of how each one of us individually compares is different and is important when one reasons for oneself.

My self assessed braininess has so far not resulted in contributing anything to pushing the periphery of human knowledge. I am, and from the looks of it, many are, just like monkeys introduced into an environment of plenty. A plenty that is a result of learning to cooperate instead of defect. Just based on this skill that our society has together figured out, I find it very hard to account for the differences in the way we treat humans from animals. I am a set of atoms in the universe that stands for an inaccurate map of the universe. And this map realizes that there are many other instances doing very similar things. Now, I don’t know why the universe is “folding” in on itself – but the differences in its solutions (i.e humans, non-human animals) are so minute that judging them differently on an existential level is like running Occam’s razor through your own neck. When I mentally diff myself with an average stranger, and that with a cow or a pig, the result is slightly in favour of the human only because he or she is a member of the human species.

Machines

As we noted above, the defining characteristic of a machine is copying. In computational parlance, we refer to it as iteration or recursion, in biological systems we call it reproduction. It started with a few carbon based molecules, to RNA, to DNA, and so on up the ladder to us, in principle, based on electromagnetic and gravitational forces. If you think of yourself as a big ass computer program ( as I sometimes do ), then the act of eating plants or animals, is essentially offloading some subroutines to other smaller computers. You are entrusting the responsibility of processing sunlight and nutrients from soil to plants, and that of converting plants ( animal feed ) to muscle and fat to animals, and thereby both profiting from them ( as their coefficients in your $\vec{self}$ are very small ), and increasing your dependence on them ( your $\vec{self}$ will not exist without them ). I’m concerned about the latter – existential risks.

If we wish to continue to exist, explore the universe and continue to grow, as we have done so far, dependence on incomprehensibly complex systems ( like a pig, or a chicken ) ought to reduce. I’m not suggesting that we’re anywhere near encoding our consciousness directly on atoms (you do want to know if the universe is isotropic, don’t you? ) or even diversifying our substrate to include silicon ( if you don’t want to end up like the Giant Panda, don’t keep eating bamboo! ). To exist longer, there is a direction ( bigger, more accurate maps ) in which we should aim to adapt our hardware ( chemical scum ) and software ( neuronal mess ); and offloading subroutines to mysterious machines is proceeding in the opposite direction.

In the long run, the above objection to increasing our dependence on animals might prove to be more relevant, but for now, it pales in comparison to another completely different kind of objection – The Matrix! We are afraid that machines will take over humanity one daylengthy proposals have been drafted to create “Friendly AI“. But guess what? For animals, that time has already come – gestation crates for pigs, battery farming of chickens; the pun on the word “battery”, unintentional as it may be, cannot slip ones mind. If you pause for a moment to think what is wrong with the situation, a simple question will do the job : What is the difference between me hunting down an animal and killing it with primitive stone tools or arrows and building a machine that automatically does the job for me? Objectively, the answer is simple – the effort required to hunt needs to be exerted over and over again for hunting to continue, where as in the case of machines, once an initial investment of thought process is made on the construction of machines, efforts scale beyond proportion. It’s the same reason why software companies rise to prominence rapidly.

Copying, and more specifically, asymmetry in the effort expended versus rewards reaped isn’t evil, per se – our brains make it incredibly cheap to communicate; our computers make it incredibly cheap to think and so on. The problem comes when non-linear effects in copying are neglected. Hunters respect their prey after hunting. Predators run for their dinner while prey run for their lives, but both are living on the edge. In Life of Mammals, there’s a wonderful clip on the last human persistence hunting. It is the ceremonial gestures and the internal understanding of the equilibrium that prevent genocides like the Holocaust from occurring in nature. When that equilibrium is disturbed, as we’ve done for cows, pigs, chickens and other animals exploited by us, there are no predator prey relationships; our predation has become so wide spread that we’re not predators any more – we’re the environment; we’re nature. And the problem is much worse than the Matrix. There is no simulation to keep the mind healthy, and there’s no sanitation to keep the body healthy. Above all, the creatures imprisoned are not very different from us at all – a difference of 50 million years over 4000 million years of evolution of life.

To sum up, I’ll just point out that when something scales to a very large extent quantitatively, its qualitative nature can change so dramatically that it cannot be expressed just in its elements. As a pathological example, wavefunction of a human makes little utilitarian sense. A more realistic example – the success of our species cannot be described solely as a set of humans, without describing their interactions. A lot of people hunting $n$ animals in the wild is very different from a few people factory farming $n$ animals. Without reciprocity of effort, one should be very careful about what one automates, more so when the automaton works on lives potentially causing prolonged pain and suffering, and even more so when the lives involved are very similar to us. The human brain easily saturates in its judgement of magnitude of extremes – the death of a pet dog is tragic while the death of million equally conscious pigs is a statistic.

Legality

There is much to say here, but I’ll attempt to keep this section short. Some facts : many countries have laws preventing cruelty to animals, some countries have banned dog meat, some have restrictions on using primates for research, dog fighting is illegal in many countries, and so is cock fighting. There are no laws on how to treat animals in factory farms; the size, cleanliness of shelters, or the density of packing animals in them is solely based on industrial practice. After going through several articles, it appears to me that laws on animal welfare are rather arbitrarily drawn out.

I would highly recommend this exchange of letters between Peter Singer and Richard Posner, a distinguished law scholar. The writing is enjoyable and the arguments are sound. At various places, the debate reduces to a question of either upholding moral instinct or applying ethical reasoning. Racism serves as a running example. By comparing the mental abilities of animals to retarded people and children, and 101 chimps to a human, some notion of utility is touched upon. For the majority of the debate, I found myself agreeing to most of what Singer said, but towards the end, I agree with Posner – introducing laws to restrict animal killing would be unwelcome. I’m a libertarian – imposing any rules on a society at large, and especially ones that lead to heated arguments, is almost axiomatically wrong. The cost of a dictatorship is efficiency. The Chinese are mostly an atheistic society, but I’d much rather prefer a democratic, secular country even though its people are mostly religious, despite the ridiculousness of religion. There is a certain long term benefit reaped from having every individual work out the logic for herself – the logic sticks. It comes from within. Meanwhile, each person who feels strongly should be allowed to put his/her arguments together, get the facts right, and open up the discussion for comments and criticisms.

I don’t think we should force others to stop eating meat, or in any way restrict their freedom to interact with animals as they please, but equally important is the fact that every vote counts. Just because 99% are responsible for eating meat, and that you will make an insignificant difference by stopping does not mean at all that you should renege. The epsilons add up, and as unsatisfying as the result of your abstinence might be, that is the only way to make a difference. Any alternative legal recourse is a dictatorship time bomb.

Nutrition

I’m much less worried about not obtaining the requisite chemicals from a vegan diet than most people. This is because, I know for a fact that my ancestry, being adherent brahmins, going back to several generations have led a healthy life being lacto-vegetarians. Further, I grew up on a vegetarian diet and there were no issues. And this brings me to a point on taste – Indian cuisine offers plenty of vegetarian options. Even for a western palate, when enough people choose to go vegan, chefs will come up with recipes that aren’t awful meat substitutes or a mash up of “greens”.

In fact, in US and other developed countries, excessive indulgence in meat is resulting in significant health problems for its citizens. Food that kills, provides many numbers tackling issues including coronary diseases, obesity, and whether human physiology evolved to survive on meat or leaves. The jury is, however, out on that one. David Attenborough is of the opinion that human incisors constitute enough evidence to make the case. It is interesting to note that chimps and other great apes are not exclusively vegetarians but derive 1-3% of their diet from other animals. Members of genus Homo are opportunists, and there is no reason to believe that meat is essential for humans.

Diary is an essential part of the Hindu diet, and presumably, that’s why Hindus respect cows and abstain from eating them. In any case, we are now aware of most of the vitamins and minerals needed for growth and maintenance, and can scientifically evaluate our diet for completeness. I’ve switched to almond milk, and am able to avoid all diary products except yogurt. My meals include a generous serving of Indian pickle, which causes “internal body heat” – don’t know the technical word for it – and yogurt helps cool it down. I’m still looking for a yogurt substitute, but other than that I’m a vegan.

(Update 29/3/2012) – Over the past two months, I started experiencing symptoms of dry mouth and reduced stamina. I couldn’t sleep for 8 hours without waking up 4 times to drink water. And I couldn’t run my usual 5k without stopping at least once to drink water apart from increased tiredness (anemic?). Got a blood test done which showed Vitamin B12 deficiency – popping supplements now. Just an FYI.

Conclusion

We started with Darwinism to point out that the observation of animals killing each other in the wild does not directly provide a justification for us doing the same. We then went on to note that it is not yet possible to cut ourselves off fully from organic beings – eating plants in unavoidable. Then we listed in a somewhat increasing order of preference, various places where we could draw the line. Knowing what we share with other animals is important, and we dedicated a section on understand pain and suffering. The analysis naturally proceeds to ask why you should care about suffering of others, but before it would help to obtain a deeper understanding of what we mean by a self. So, we devoted a section to breaking down our self into various components, against the unifying feeling of “I” we consciously possess, to better appreciate our internal motives. Using a simple game, we went on to explore the causes and benefits of caring about other players, and noted that from an evolutionary perspective, there is a feedback loop between broadening our definition of self and the increase in complexity of life.

The next three sections apply the abstraction to animals. First we observe that our success as a species is almost exclusively a result of cooperating as a society, and that taken individually, the difference between humans and animals is very small. This asymmetry of success as a result of interaction is a double edged sword – we can bring about a lot of constructive change by pooling our efforts in one direction, and also cause tremendous destruction by neglecting to account for scaling of minor unpleasant deeds. Machines have already spelled doom for many lives.

We then touch upon legality, to point out that our current system, formed out of a combination of evolutionary inertia and an intermittently growing sense of responsibility to other creatures, has drawn seemingly arbitrary lines. I am, however, of the opinion that forcing a change will potentially cause more trouble in the long run outweighing the result of adding up all short term benefits, and in this regard, I differ with Peter Singer. We finally conclude with a short section on nutrition to state that it isn’t much of an issue, even practically.

I write about this subject at length, because I think we have a responsibility to investigate the cost of keeping our bodies running. The analysis becomes even more important when its conclusions are shared by a minority. During my meat eating years, I knew of no one else who argued against vegetarianism more strongly than I, and now that, after reading, watching, and acquiring more evidence, my stance has changed, I find it imperative that I defend my position. Of course, I’m open to revision, and welcome critics who can point out holes in my argument, but until then I rest my case.

Advertisements

[Cross-posted from goodreads]

If Godel proved that no sufficiently complex system, i.e one that is capable of arithmetic, can prove its own consistency or if you assume the system is consistent there will always exist (infinitely many) true statements that cannot be deduced from its axioms, in what system did he prove it in? Is that system consistent? In what sense is the Godel statement true if not by proof? You’ll have hundreds of questions popping in your mind every few minutes, and this short book does a very good job of tackling most of them.

Godel numbering is a way to map all the expressions generated by the successive application of axioms back onto numbers, which are themselves instantiated as a “model” of the axioms. The hard part of it is to do this by avoiding the “circular hell”. Russell in Principia Mathematica tried hard to avoid the kind of paradoxes like “Set of all elements which do not belong to the set”. Godel’s proof tries hard to avoid more complicated paradoxes like this :

Let p = “Is a sum of two primes” be a property some numbers might possess. This property can be stated precisely using axioms, and symbols can be mapped to numbers. ( for e.g open a text file, write down the statement and look at its ASCII representation ). The let n(p) be the number corresponding to p. If n(p) satisfies p, then we say n(p) is Richardian, else not. Being Richardian itself is a meta-mathematical property r = “A number which satisfies the property described by its reverse ASCII representation”. Note that it is a proper statement represented by the symbols that make up your axioms. Now, you ask if n(r) is Richardian, and the usual problem emerges : n(r) is Richardian iff it is not Richardian. This apparent conundrum, as the authors say, is a hoax. We wanted to represent arithmetical statements as numbers, but switched over to representing meta-mathematical statements as numbers. Godel’s proof avoid cheating like this by carefully mirroring all meta-mathematical statements within the arithmetic, and not just conflating the two. Four parts to it.

1. Construct a meta-mathemtical formula G that represents “The formula G is not demonstratable”. ( Like Richardian )
2. G is demonstrable if and only if ~G is demonstrable ( Like Richardian)
3. Though G is not demonstrable, G is true in the sense that it asserts a certain arithmetical property which can be exactly defined. ( Unlike Richardian ).
4. Finally, Godel showed that the meta-mathematical statement “if ‘Arithmetic is consistent’ then G follows” is demonstrable. Then he showed that “Arithmetic is consistent” is not demonstrable.

It took me a while to pour over the details, back and forth between pages. I’m still not at the level where I can explain the proof to anyone clearly, but I intend to get there eventually. Iterating is the key.

When I first came across Godel’s theorem, I was horrified, dismayed, disillusioned and above all confounded – how can successively applying axioms over and over not fill up the space of all theorems? Now, I’m slowly recuperating. One non-mathematical, intuitive, consoling thought that keeps popping into my mind is : If the axioms to describe arithmetic ( or something of a higher, but finite complexity ) were consistent and complete, then why those axioms? Who ordained them? Why not something else? If it turned out that way, then the question of which is more fundamental : physics or logic would be resolved. I would be shocked if it were possible to decouple the two and rank them – one as more fundamental than the other. I’m very slowly beginning to understand why Godel’s discovery was a shock to me.

You see, I’m good at rolling with it while I’m working away, but deep down, I don’t believe in Mathematical platonism, or logicism, or formalism or any philosophical ideal that tries to universally quantify.

My primary reason for reading Origin is its historical significance. It bothers me that humans, until as recently as Darwin, did not seriously ask about the origin of species despite interacting with its members every day. That this revolutionary idea should have occurred to one person as opposed to many in an incremental fashion, is far from obvious – and extremely interesting retrospectively. Are there any other such revolutionary ideas that’ll later be trivial to everyone?

Darwin supported his argument meticulously by observations, made by him and others, and their sheer number is so mind boggling that I fell asleep more times reading this book than any other. It amazes me how he summoned up so much enthusiasm to study and compare the most boring habits of some of the dullest creatures.

There is no reason to read this book to understand natural selection – our knowledge now is a far more superior and complete. I had already read The Selfish Gene, had already experienced the profound “OMFG! That’s Brilliant!” reaction that rapidly morphed into a “Duh! Isn’t that obvious?” . Yet, Origin was an enthralling read for the most part, with many opportunities to pause, wonder, daydream, extrapolate, apply the theory to modern humans and computers, and so on.

This is a fairly long book about a seemingly tautological argument. I highlighted a large number of lines and took down a few notes on my kindle. A short summary of my observations, and a few quotes to give you the taste of what most of the book is like, follows.

1. Variation – Darwin appreciated that variability exists in nature,
but he did not seem to explore the causes or consequences – which is understandable as genetics was ahead of his time.

“It may seem fanciful, but I suspect that a similar parallelism
extends to an allied yet very different class of facts. It is an old
and almost universal belief, founded, I think, on a considerable body
of evidence, that slight changes in the conditions of life are
beneficial to all living things.”

“Again, both with plants and animals, there is abundant evidence, that
a cross between very distinct individuals of the same species, that is
between members of different strains or sub-breeds, gives vigour and
fertility to the offspring.”

“Dominant species belonging to the larger groups tend to give birth to
new and dominant forms; so that each large group tends to become still
larger, and at the same time more divergent in character.”

“Widely ranging species vary most, (…) and varieties are often at
first local,–both causes rendering the discovery of intermediate
links less likely. (…) And if there be any variability under nature,
it would be an unaccountable fact if natural selection had not come
into play.”

2. Cooperation – Darwin appreciated that environment in the form of
sunshine, water, temperature ( non living stuff ), wasn’t the primary
reason for success of life in its complexity, and that inter-species
interaction is a large cause. However, he talks about competition,
much like Dawkins, but not much about cooperation, like Lynn Margulis. I
think cooperation is much less apparent when you just look at the
phenotype. It seems to me that it is more widespread on the smaller
gene level with microbes and viruses exchanging chemicals frequently, and
hence completely understandable why Darwin was unaware of it.

“This long appeared to me a great difficulty: but it arises in chief
part from the deeply-seated error of considering the physical
conditions of a country as the most important for its inhabitants;
whereas it cannot, I think, be disputed that the nature of the other
inhabitants, with which each has to compete, is at least as important,
and generally a far more important element of success.”

“Bearing in mind that the mutual relations of organism to organism are
of the highest importance, we can see why two areas having nearly the
same physical conditions should often be inhabited by very different
forms of life; for according to the length of time which has elapsed
since new inhabitants entered one region; according to the nature of
the communication which allowed certain forms and not others to enter,
either in greater or lesser numbers; according or not, as those which
entered happened to come in more or less direct competition with each
other and with the aborigines; and according as the immigrants were
capable of varying more or less rapidly, there would ensue in
different regions, independently of their physical conditions,
infinitely diversified conditions of life,–there would be an almost
endless amount of organic action and reaction,–and we should find, as
we do find, some groups of beings greatly, and some only slightly
modified,–some developed in great force, some existing in scanty
numbers–in the different great geographical provinces of the world.”

3. Ontology of god – Darwin did not comment much on Man, but he did
make strong rational arguments against creation, and understood the philosophy of science.

“Why should all the parts and organs of many independent beings, each
supposed to have been separately created for its proper place in
nature, be so invariably linked together by graduated steps? Why
should not Nature have taken a leap from structure to structure? On
the theory of natural selection, we can clearly understand why she
should not; for natural selection can act only by taking advantage of
slight successive variations; she can never take a leap, but must
advance by the shortest and slowest steps.”

“But many naturalists think that something more is meant by the
Natural System; they believe that it reveals the plan of the Creator;
but unless it be specified whether order in time or space, or what
else is meant by the plan of the Creator, it seems to me that nothing
is thus added to our knowledge.”

4. Altruism – Darwin identified and seemed to have understood the
conundrum posed by “selfless” individuals of the ant of the bee
communities. Since his argument is at the species level, it was easier
for him to reconcile the selfless acts of a few “slave” members as
beneficial to the species. Dawkins’ must have had a tougher time as
his arguments were at the genetic level, but even that was resolved
when it was shown that slave members had a different gene composition
that made it advantageous for their genes to die in the service of the
queen’s genes.

“Finally, it may not be a logical deduction, but to my imagination it
is far more satisfactory to look at such instincts as the young cuckoo
ejecting its foster-brothers,–ants making slaves,–the larvae of
ichneumonidae feeding within the live bodies of caterpillars,–not as
specially endowed or created instincts, but as small consequences of
one general law, leading to the advancement of all organic beings,
namely, multiply, vary, let the strongest live and the weakest die.”

“for if on the whole the power of stinging be useful to the community,
it will fulfil all the requirements of natural selection, though it
may cause the death of some few members.”

5. Falsification – I think it is important to note that when a
groundbreaking theory is introduced, it should make some substantial claims
that are against common knowledge. Although Darwin did not tackle all
the implications of his theory in this book, he did make remarks
which, at that time, would perhaps have been considered quite bold.

“hence there seems to me to be no great difficulty in believing that
natural selection has actually converted a swimbladder into a lung, or
organ used exclusively for respiration.”

I failed to post a blog entry last week, and in order to make up for it, I should post two this week. Here’s the second one.

Lesswrong is a blogging group that is “dedicated to the art of refining human rationality”. There are a lot of posts about Bayesian probability, Cognitive biases, Artificial intelligence etc, and if you’re new to it, I’d recommend starting at one of the sequences. Apart from online discussions, lesswrong community has come to realize the importance of meeting up in person. This is particularly important considering the kind of audience and participants the blog attracts – slightly geeky individuals possessing high IQ who are presumably not the most social people. I can relate.

The Ottawa meetup group started sometime in the summer with an average 6 people attending every week. For various reasons, people have stopped coming. It seems, those who were aware of it when it started were the only ones to join the group – around 10 people. We did not make any efforts to spread the word around. In any case, the average group size is 3 ( Alex, Andrew and me), and occasionally a couple of other people show up. So, here’s a summary of what goes in a typical meetup. This one was a couple of weeks ago.

Summary of Monday’s meeting –
We were supposed to meet at Andrew’s at 7.30, but it got postponed to 8. Alex was 15 minutes late, and when he arrived at 8.30 he brought a pack of “Set” cards. The three of us played one round while doing some bistro combinatorics on the game. As usual, for every game there’s a mathematician who’s studied it. http://www.math.rutgers.edu/~maclagan/papers/set.pdf

We then moved over to the couch. I had printed 3 copies of chapters 2 and 3, and surprise surprise!! only 3 people showed up. Talking about making decisions under uncertainty, I asked Alex to ask me about any outward-looking issue lesswrong deals with ( not the self-help ones ), and I said I’d reduce it to an issue in complexity. He asked something along the lines of “making decisions under uncertainty”, and I quickly interpreted it as “optimal decisions” and hand-wavingly introduced bounded rationality – the process of optimization of ones perceived “decision function” is critically ,possibly non-linearly, affected by ones intelligence. Alex then asked for some examples of problems in P and some in NP. After an embarrassing amount of thought to come up with a simple example, i gave a couple of terrible examples. But since I’m writing stuff down, here are some better ones –

NP Complete – http://en.wikipedia.org/wiki/List_of_NP-complete_problems#Games_and_puzzles
P – I mentioned matrix multiplication.

TODO for next monday – Clarify the meaning of decision problems and then the classes P, and NP. And basically why P=NP is the most important problem in math.

Other usual favourite topics that surfaced included the genetic influence of homosexuality, cryogenics etc. Andrew asked the following question : What if there was a machine into which you stepped, and out came an exact copy of you but your original was destroyed. I obviously pointed out the no-cloning theorem while Alex remarked that perhaps no significant things that make us us have Quantum origins. Then we proceeded into debating about how we  compute “I” – this was a long conversation. I pointed out that until very recently in human evolution any definition of “I” had a significant genetic/family/culture factor, where as now. in the context of “uploading our brains”, we’re defining “I” in a purely intellectual manner. Basically arguing about whether it’s ok for genes to ‘go jump in the lake”. This went on for a while, and Alex remarked that my thoughts race ahead of my words to the extent that the conversation seemed like a random walk. On homosexuality, I quickly mentioned that it wasn’t an ESS – homosexuals will not procreate and will die out. Alex mentioned that a recessive gay gene could exist in a person, while his homosexual brother who has two recessive gay genes can help him survive and procreate by providing resources. Andrew pointed that incidence of homosexuality is observed to be higher in the younger siblings. I wasn’t totally convinced that Andrew’s gay-producing theory and Alex’s gay-maintaining theory added up to their prevalence. We also wondered about homosexuality proportions in animals.

That’s about as much as my aching head could remember. We did not cover much of the paper, but we will continue doing it. Next week, I propose we play some card game ( Sets was fun ), sit down and go through some concepts and try to cover chapters 2 and hopefully some part of 3 in reasonable detail. “Fart sessions” as we used to, in undergrad,  refer to any undirected, dissipative discussion sessions ( http://archiv.tu-chemnitz.de/pub/2006/0020/data/MAthesis_EvelynRichter.pdf ) will inextricably follow.

In Ottawa lesswrong meetups, we’ve started to discuss Scott Aaronson’s Why Philosophers Should Care About Computational Complexity. We read and discuss three chapters every week, and I figured it’d be nice to summarize the material. The original pdf is by itself isn’t very long at all, and contains summaries – I’d highly recommend you read that instead of this. In any case, I’ve decided to create a summary and add context to make things clearer for me. I’ll keep adding chapters as we cover them.

Chapter 1 – Introduction [ Added background. Pedagogic ]

Analogous to how cars and other mechanical inventions were developed to cater to our laziness, computers are devices onto which we offload our most boring and repetitive thought cycles. Computability theory is the study of the extent to which it is possible to identify and demarcate those boring human thoughts, “algorithms”, and have them work on other devices, “computers”. In the 1930s, people came up with different models of computation, but most surprisingly they all turned out to be equivalent! Any boring thought you might have will run the same on any computer you make. Turing machines became a popular model for analysis, but you’re free to choose your own – ants (Langton’s), game of life (Conway’s ) or lambda calculus (Church’s).

You can see why philosophers got excited – they wanted to know if all human thoughts are boring. In other words, can we come up with a thought that is not boring but well-defined? One of the first questions posed about this was Hilbert’s Entscheidungsproblem. He was a mathematician and he knew that statements about numbers are really boring. But is it possible that there are some statements about numbers that are not boring? Can we come up with such a statement? How? – simple – Figure out a way to ask the same question using numbers! – “Can you come up with a boring thought that’ll classify statements about numbers as true or false?”. Kurt Godel figured out a way to ask that question precisely by mapping it to numbers. Viola! All hell broke loose. The answer was a loud screaming “NO! Not all true statements about numbers can be proven to be true” – Godel’s incompleteness theorem. But wait. What do we mean by true? Can we even define precisely what we mean by true. This time Alfred Tarski yelled “NO! Arithmetical truth cannot be defined in arithmetic.” – Tarski’s undefinability theorem. The hardest blow came when Turing asked a similar question “Can you build a Turing machine that can tell if a given Turing machine halts”, and elegantly explained “No, you cannot” – the Halting problem. This was worse because it wasn’t about numbers anymore, or about the nature of truth – it was about actual real physical devices you can build. All these powerful negative results led to the Foundational Crisis of Mathematics. For decades to come, mathematicians and logicians provided enough fodder for generating non-boring thoughts in the minds of philosophers. They devoured it.

But that was a long time ago. Since the 1970s, when we started building real computers, we became increasingly concerned with the complexity and efficiency of algorithms – how much space and time are required for efficient computations? Computational complexity theory was born. It has now developed into a rich theory but it appears that philosophers are still dining on computability theory. We have a new restaurant at this end of the universe not yet frequented by philosophers. Scott Aaronson is our cook and I’m your waiter.

Chapter 2 – Complexity 101 [ Added background ]

If we know something is computable does it matter if it can be computed in 10^5 seconds (1 day) or 10^100 seconds? Yes. Our universe is less than 10^20 seconds old. Clearly, the problem is in the exponent. So, if you want to solve a problem of some size, n, and the most efficient algorithm to do it takes time 10^n, then you’re screwed. If the size is in the exponent, we’re not efficient. But if it takes something like n^10 or even n^100, we’re happy. If the size occurs in the base, our algorithm is efficient. The next obvious thing we’d like to know is which problems admit efficient algorithms and which don’t. Just like in computability theory where we figured out that many models of computations were equivalent, we’ve now learned that many hard problems are equivalent. They’re called NP(Non-deterministic, Polynomial time) -complete problems. It doesn’t matter which problem you figure out an efficient algorithm for – if you find for one, you’ve found for all. The efficient ones, where size occurs in the base and not in the exponent, are called P (Polynomial time) solvable problems.

I will take a moment here to explain what NP-completeness means, because it is crucial. Imagine you’re asked to solve a problem, for e.g sort a pack of cards. You may be interested to know how many cards you’d have to compare in the process – so you ask “What is the minimum number of comparisons needed?”. This is an optimization problem and the answer is a number. Other problems may have answers in various other forms – list of all cities in the shortest route etc. In order to be able to compare these problems and their algorithms, it’d be nicer to have answers that are comparable, and the simplest such form is a boolean value. So, you instead ask “Is the minimum number of comparisons needed to sort greater than 100”, to which the computer only answers True or False – these are called decision problems. Another feature of any algorithm, you might have noticed, is that it must make choices and depending on the outcome of these choices has to proceed further. Think of it as a binary tree – you start at the root, make your choice and proceed leaf-ward. Leafs contain answers – True or False. When a Deterministic Turing machine (DTM) makes a choice, it has to go down its chosen path, and if it turns out that it was a bad decision, it has to come back and go down the right path. A Non-Deterministic Turing machine (NTM), on the other hand magically always make the right choice. All problems efficiently solvable using a DTM is defined as the set P, and all problems efficiently solvable by the NTM is defined as the set NP. Good. Now to NP-Completeness. Stephen Cook in 1970s discovered that many many problems ( but not all ) in NP were equally hard and were the hardest problems in NP – these are called the NP-complete problems. What exactly is the power of this hypothetical magic in an NTM ? We don’t know – that is the P =? NP question.

After that long digression which has hopefully brought most readers on a common footing, let me return to what Scott Aaronson has to say in this chapter. First, why are we so comfortable in dividing up efficiency ranges into polynomial and exponential, and not something in between like n^logn ? A second objection is that in practice algorithms that take an exponential 1.00001^n time might run faster than n^100000, because in practice inputs are bounded. These questions are tackled in later chapters. A third point noted is that while computability theory concerns itself with logic about what can and cannot be computed, complexity theory deals with what can be done best given some resources, and in this sense it is more closely related to physical sciences. So, one can ask, if we could travel backwards in time, could be compute faster, or is there a limit physics poses on the amount of information a limited space can store.

Chapter 3 – The relevance of polynomial time

3.1 The Entscheidungsproblem Revisited

Church and Turing showed that even if you take an incomplete formal system, you cannot classify all statements into “provable in F”, “disprovable in F”, and “undecidable in F”. Godel in a famous “Lost letter” wrote to von Neumann asking if there would be a way to decide if a given statement has a proof in F of length n bits or less efficiently. i.e if a proof of a reasonable length exists, can it be found in a reasonable amount of time? Godel had anticipated the P vs NP problem.

3.2 Evolvability

Is 4 billions years time enough to construct a human body ( which presumably optimized many of its inner workings w.r.t its environment ) from a much more homogeneous environment that formed the earth? Godel thought not, but modern biology says otherwise. Basically, complexity theory is nowhere near even hinting at us an answer to the possibility of Evolution.

3.3 Known Integers

This is an interesting question – when you say you know something, what do you mean? Let us, for a moment, keep aside knowledge of the kind most people spend their lives trying to acquire – mating habits or courtship rituals of humans – and ask more specific questions, for instance, ones that mathematicians claim they know – proving something. When a mathematician claims he knows how to prove a statement to you in a system, what he’s saying is that he can, in an amount of time polynomial in the size of the statement, reduce it to something you both agree on. To make this more concrete, when you say you “know” a prime number p what you’re saying is that you have an algorithm that proves in polynomial(p) time that p is a prime. The same thing cannot be said for the prime := “next prime larger than p” even though it is well defined. Although this seems like a very interesting way to think about things, Gil Kalai opines “The opinion that once you show that a certain human activity represents an effort in P means that it is “trivialized” is a bit too far since it is quite plausible that all human activities represent efforts in P”.

Today, I’d like to talk about the idea of permanence from an ontological standpoint. The issue i’m concerned with is the ease with which we talk about the existence of something independent of time. I will make observations in common subjects and move on to less general ones.

Richard Dawkins’  book The Selfish Gene theorizes the gene-centric view of evolution where its phenotypes,  you and I, are carriers of genetic material on which selection happens. After reading the book, I was left with a feeling of inferiority in the manner of a helpless subordinate. The feeling might have well been due to a strong sense of associating my self with my mind (“I am the most abstract thought I can think”) as opposed to computing my identity by weighting several well regarded societal good-to-haves. In any case, it is true that certain patterns of A, T, C, G have been very successful over 4 billion years over other patterns of nucleotides . It is true that some biological structures like warm blood or vertebrae or neurons have been more successful than other structures. On the other hand, it is also true that some elements are most abundant in the universe than others – hydrogen and helium for obvious reasons, iron for a more interesting reason ( watch BBC Atom ). Hydrogen atoms are almost as old as the universe, younger by about 300,000 years. Atoms, neucleotides, genes, proteins, organs, individual organisms, species and many distinct layers in between all have existed for different time periods with each comparable layer exhibiting a non-uniform pattern in its space of all possible patterns. All right. With these observations in mind, let us ask some questions :

Q. What is the purpose of life?

A. 101010 due to some random effects in Douglas Adams’ neural connects.

Q. No, seriously. If I think of myself as a computer, as I most often do, what function f(x) do I optimize?

A. When I read Dawkins, one answer immediately seemed to be a perfect match – “Procreate as much as possible to optimize the survival of all your children”, or “Don’t disappoint the genes that formed you”. This explanation of what we’ve unconsciously been doing all along applied very well to the people around me. People and non-human animals spend a majority of time directly or indirectly in search of a mate, or finding resources for the upbringing of a young one. For them, f(x) is necessarily in terms of their gene patterns; “phenotypic” residues ( for ex. intellectual contributions ) have very short life spans. I suspect their strategy of having as many kids as possible is a way to produce future computational instances that can continue crunching on their f(x), just as they had been of their ancestors’. All of this is fine, but I think there was a very very crucial event in the 1800s, that changed the whole game. In On the origin of species, Darwin clearly explained this in great detail. We now know roughly the characteristics of f(x) that was being optimized, and in my opinion this knowledge has significantly altered that f(x). It is the measurement problem. World human population is expected to decline. Populations in developed countries, presumably with “wiser’ people, are already on the decline. Genes that evolved to code for a human phenotype are probably not going to live the longest. Why then should I consider the spreading of my genes as my primary purpose?

Q. So, what has this f(x) become?

A. It can be whatever you want it to be. I realize that doesn’t mean much as it quickly reduces to the problem of free will vs determinism, which is really non-problem. What I mean by that, is f(x) doesn’t have to be continuous. Every once in a while a “level-crossing” feedback loop is completed – f(x) becomes complicated enough to model the previous version of f(x). At that point what appears to be a linear growth of complexity encounters a spike. A black swan. In practical terms, what I’m trying to say is that we as humans seem to be beginning to learn that “permanence” or “longevity” of something is a worthy candidate for optimizing. Genetic configuration is not primary anymore. Welcome, gays, lesbians and smart people who don’t want to have kids. Your bodies ( including minds ) have figured out a way to experience the pleasure of sex, possibly created and selected as a major driving force of survival, without procreation. Find your new thing that has a larger value of permanence – contribute to the scientific body! That is the higher being!

I wanted to talk a lot more about permanence in Mathematics. About universal quantifier “for all”, about Continuum hypothesis, about Mathematical platonism etc. I wanted to mention our quest for permanence in Physics – of virtual particles and an obsession with finding theory of everything,  Maybe if time permits some other time. Good Night folks!

Halloween weekend. Woke up late, read a slashdot article  “When Having the US Debt Paid Off Was a Problem” which claimed that US had a budget surplus in 2000 and could pay off its entire debt by 2012 and much more surprisingly how that would be harmful to the US economy. I don’t even pretend to understand the finance and my usual strategy is to dismiss it as a shitty unproductive zero-sum game greedy people play. But today, I had more important things to study – i.e more important things that could be put off. So, I ended up watching 15 Khan Academy videos about “Paulson Bailout“.

Here’s what I learnt…for future reference.

In order to calculate what you’re worth, you list all that you own (Assets)  and all that you owe (Debt) in some currency and the difference is the answer (Equity). The whole problem is to determine what constitutes your Asset, Debt, and figure out how to map them to, say, USD. We should learn some terminology first.

Mortgage Backed Security – People need to buy houses. They don’t have all the money. They get loans from commercial banks or wherever, and pay mortgages for 30 years covering interest and principal. Investment banks with liquid cash (C) set up a company “Special Purpose Entity” with some liquid cash and buy these mortgages. So now, homeowners pay their monthly mortgage to these SPEs. SPEs bundle up these mortgages into bundles called “Mortgage Backed Securities” and trade them on stock exchange (perhaps to make up their initial cost C). Why would people want to invest in such stocks? Because stocks pay dividend and if it is more than what your Savings account interest is, then why not? Wait – what is going on? Investing in a regular stock means that you are placing your confidence in the productivity of that company, i.e you think d = company_productivity(your money +  company’s money) – company_productivity(company’s money) > 0. Moreover, you think that d > your money, which means that investing in this company makes the society more productive. But in the case of SPEs, or MBS the how-does-my-investment-make-society-more-productive-loop is much much bigger – you invest in an Investment bank, which lends more money to people to buy houses, which makes them work harder to pay their mortgages improving society’s value. If everyone paid their mortgages, it wouldn’t be a problem. But there exists a discontinuity in the system – insolvency. Now, the game is about who’s going to hit the bound statistically speaking. So, SPEs package their MBSs into classes “tranches” – high risk (unsecured), medium risk(mezzanine), and low risk(super secured), with correspondingly decreasing dividend rates, and these are called Collateral Debt Obligations (CDO), where the risks are evaluated by rating agencies. When someone defaults, money vanishes from the unsecured CDOs first and then maybe mezzanine if needed.

US Treasury Bonds – When you lend someone some money, they give you a receipt which contains details, i.e who borrowed from whom at how much and for how long at what interest etc etc. This receipt is called a bond certificate. When you buy US Treasury bonds, it means that you’re lending the US govt. some money.

Good. So, now what does a bank’s balance sheet look like? Its assets include some liquid cash, perhaps some good corporate bonds (money lent to Microsoft perhaps :P), maybe some US treasury bonds, and maybe some “high risk” CDOs. For some reason banks have liabilities – they take short term loans, ( for ex. 3 months ), pay interest on them, and pay the principal off by taking out another loan. I don’t exactly understand why, but it seems these loans form a large part of their liabilities. For ex. today wikipedia says Goldman Sachs has about $900B of assets, but only$77B of equity, which means over $800B in loans. Lehmann Brothers had over$600B of Assets when it declared bankruptcy! Equity = (Assets – Liabilities) can be hard to estimate when there are risky CDOs. Banks bought these CDOs at a high price during the housing bubble when everyone thought that housing prices keep rising eternally which was reflected on their “book” value on balance sheets. Some people begin to default on their mortgages. This affects the “high risk” category of CDOs which are now worth far less than what the banks bought them for, i.e their “market” value on balance sheets is far less. Banks need to write them off to retain trust. This brings down their Equity. If Equity goes below zero, the bank declares insolvency ( like Lehmann Brothers did). This means all its creditors lose money. But its creditors are other banks, which had these loans listed in their assets. Now people look at these surviving banks’ balance sheets with a suspicious eye wondering if their assets represent their real market value. Banks don’t want to default, so they hold on to their liquid cash to pay off their loans when they’re due. Market slows down, and the “real” value of anything is not easy to determine. A few more banks default.

What can the government do? Well, nothing. If a society is capitalistic, it should be so when banks are profitable and when they make losses. That did not happen. Paulson bailout injected \$700 billion into these banks by buying the risky CDOs at a much higher price than market value. Reverse auction – where the feds setup an auction saying they want to buy come CDOs with some cash. The company that sells it to them for the cheapest price wins it. This is a terrible idea – a company desperate enough to sell its CDOs for some quick liquid cash isn’t going to garner more trust. Who know what else it has overstated or fudged in its balance sheet? Moreover, what will a bank that receives this bailout fund do? It will hold on to its cash so it can repay its loans and not default. All these bailout funds go stagnant. Some banks die, some survive. Government debt increases, and a few weeks ago that hit the ceiling as well.

Making money off money stinks.

Advertisements