Previous Entry Share Next Entry
Interesting Links for 06-03-2012
Illuminati
andrewducker

Original post on Dreamwidth - there are comment count unavailable comments there.

Re the Mortgage Indemnity scheme. I think your note pretty sums up the situation. Although I’ve not been able to find the actual policy document yet I think what they are proposing is a scheme whereby the government (you and I) stand guarantor for the first tranche of any capital losses.

So in the event of the mortgage holders not being able to make their repayments and the cash price of the home having fallen we, through the government, take the hit.

One the one hand, one might ask why the government is protecting private lenders from losses on their lending. You might also worry about taxpayers being left to pick up the cost of providing subsidies mortgages to relatively well off people i.e. we are guaranteeing the loans of people with jobs and not providing social housing directly to those not in employment.

On the other hand, this might be a cheap way of providing a Keynsian boost to the economy.



My usual back of a spreadsheet useful only as a contextual starting point number crunching follows.

Say average cost of a new build house is £100k. The scheme aims to help 6,000 borrowers buying new build housing. Be generous and assume that that housing would not have been built but for the scheme.

That’s £600m of house building over the next few years.

The direct cost of the scheme would be any guarantees the government has to pay out.

Assuming £600 mn of loans guaranteed for the first 25% of capital losses is £150m of guarantees. With a 1% delinquency rate the direct cost to the government is going to be about £1.5m spread over a couple of years.


The next point is slightly off because the Scottish government doesn’t collect income tax or pay social security but there would also be the net gain of switching some builders’ labourers from unemployment benefit to waged employment, probably paying a little bit of income tax.

Algernon's Law - can anyone spot the obvious flaw?

There are so MANY flaws it's hard to know which one to find first.

1) The "requires a leap". A slight increase in intelligence may be no evolutionary advantage, a massive increase in intelligence might be.

2) The "lessons of prehistory". The same argument could have been made at any point in pre-history "Clearly these neanderthals could not be more
intelligent or evolution would have made them that way hence increased intelligence would be an evolutionary disadvantage". Evolution in the traditional sense is very very slow (which is why many people think human changes are now governed by memetic not genetic changes).

3) The "we're going that way anyway". The Flynn effect is the name for the effect whereby IQ scores have systematically been increasing throughout the world (arguably slowing or reversing equally). There's moderate evidence that evolution (or societal influence... or something) is currently increasing our intelligence incredibly rapidly in evolutionary scales.

4) The "evolution is not that good at optimising". The red squirrel
turned out to be pretty good in its environment. The grey squirrel was better. There are countless examples of species being killed off by introduction of another set of species (most of South America's large animal life died off rapidly in an event known as the "Great American Interchange" when the North American animals proved simply better at competing in that niche). People tend to assume that evolution produces things maximally fit for their local environment, forgetting that the local environment includes the species extant... so if there's nothing in particular preying upon an animal, there's no need for it to develop a way to avoid that.

5) The "it can't easily evolve". A wheel is a tremendously good solution for moving things but it has (arguably) never evolved (there are some microscopic organisms which arguably use such a mechanism). You can read a lot of tracts on why "wheel evolution" is difficult. It may be that the mechanism to create three times intelligence is relatively easy to engineer but difficult to evolve -- that is we're far more likely to get there from gene manipulation than from evolution.

6) The "optimising the wrong thing". Evolution doesn't improve what most people think it improves. Evolution improves an animals ability to replicate its genes (either through having children or promoting survival of kin). A peacock is a tremendously silly design from most points of view apart from aesthetic... a male peacock is hugely compromised by its design being that thing which female peacocks fancy. So, for example, if the article had argued "any attempt to artificially create a land animal larger than the elephant would automatically fail because all these changes would have a cost and the elephant is already optimised" we'd know it was silly (because there have been such animals). But evolution is no more trying for "smarter" than it is for "larger".

7) The "no selection pressure for that". (Variant of above) Is there actually selection pressure on humans to be intelligent, indeed is there any selection pressure on humans at all? Humans aren't significantly predated, in the developed world, humans rarely die before the end of their reproductive age. While people often say they find intelligence attractive, do intelligent people have more children? Indeed, many people (often nasty people) argued the complete opposite, that unintelligent people have more children and any evolutionary effect is lowering not raising intelligence. While I'm not going to comment on that, it is in no way clear at all that there's any evolutionary pressure right now to raise intelligence. Indeed the most dominant selection pressure on humans in the developed world is probably that we evolve a strong desire to have children.

At least some of these are discussed in the article.

Indeed, for example the Flynn effect was mentioned and described as "slow"! (I kid you not.)

There's also a lot of handwaving about what the author assumes is an evolutionary advantage.

> 1) The "requires a leap". A slight increase in intelligence may be no evolutionary advantage, a massive increase in intelligence might be.

I don't know what this objection is supposed to be. Large mutations, as in multiple simultaneous mutations which all work together, are exponentially rare. This is why the creationists focused so much on the eye, and why the step-wise incremental multiple evolutions of the eye were so important: because the odds of multiple simultaneous mutations is so small it might as well be zero.

> The same argument could have been made at any point in pre-history "Clearly these neanderthals could not be more
intelligent or evolution would have made them that way hence increased intelligence would be an evolutionary disadvantage".

That doesn't mean the argument is wrong. For example, buying a lottery ticket is net expected value; yet some people still win.

We don't know that Neanderthals were not smarter. Brain size correlates with IQ in existing homo sapiens, and skull volume was both smaller than Neanderthals and may have shrunk since then to now. (This is discussed in footnotes.) And as already pointed out, humans pay heavy prices for their large heads as it is: maternal mortality was no joking matter.

> The Flynn effect is the name for the effect whereby IQ scores have systematically been increasing throughout the world (arguably slowing or reversing equally)

Yep. There you go: we've hit the genetic ceiling, now that environmental variations like iodine deficiency have been eliminated. Where are your cheap easy fitness-improving intelligence increases now?

> 4) The "evolution is not that good at optimising". The red squirrel turned out to be pretty good in its environment. The grey squirrel was better.

I really don't see how this is relevant.

> It may be that the mechanism to create three times intelligence is relatively easy to engineer but difficult to evolve -- that is we're far more likely to get there from gene manipulation than from evolution.

You're just agreeing with EOC loophole #3 here.

> 6) The "optimising the wrong thing". Evolution doesn't improve what most people think it improves. Evolution improves an animals ability to replicate its genes (either through having children or promoting survival of kin).

Agreeing with EOC loophole #2.

> 7) The "no selection pressure for that". (Variant of above) Is there actually selection pressure on humans to be intelligent, indeed is there any selection pressure on humans at all? Humans aren't significantly predated, in the developed world, humans rarely die before the end of their reproductive age.

Free efficiency gains in intelligence is never a bad thing; if nothing else, evolution can just starve the brain of resources, maintaining a constant intelligence level, and sending the free-up resources to other fitness-increasing things like muscles or a gut that can handle more kinds of food (see footnotes). There will always be selection pressure for efficiency.

There certainly may be dysgenic pressures now, but they can't have been operating for more than a century or so and are correspondingly small in effect, inasmuch as Gregory Clark has documented superior reproduction by the rich in England up to fairly recently. Even by the most pessimistic estimate I've seen, there's not been more than 20-30 points of potential average lost. So here too we would not expect to see any cheap easy big IQ gains.

I'm not sure if you clicked through from the first article to the second which it's based on (it's on wayback engine) -- so some of this was in response to that as well (it's a slightly bonkers freeform ramble about intelligence increases). The whole point was answering the question can we make some major (non evolutionary) changes which would increase human intelligence by a vast degree.

So for what you say about (1) Yes, large mutations are rare. That is my point. A human created invention might be able to create something that evolution could not -- a huge increase in intelligence with no loss of "evolutionary fitness" which evolution could not find because it requires such a major change that it's incredibly unlikely. (If you've ever played much with genetic algorithms, unless you take some fairly "whacky" cross breeding, they get stuck in local minima easily).

2) If you don't believe neanderthals had a lower IQ, consider a point where something like a chimpanzee (or of equivalent intelligence, I've no idea which hominid would be in that point on our evolutionary journey but there must have been one) was the most intelligent animal on earth. Now consider the argument "clearly being more intelligent than a chimpanzee will lead to a loss of evolutionary fitness or chimpanzees would be more intelligent".

3) You have no evidence whatsoever that we have "hit the evolutionary ceiling". If you want to make the argument that evolution cannot make us more intelligent without a loss of evolutionary fitness you must first establish whether we are currently doing so. The best evidence available is that currently we are getting more intelligent at a startlingly rapid rate. A rate so rapid that in evolutionary terms, if it continues we would have astounding intelligence on an evolutionary timescale. A lot of people have hypothesised that this slowing is due to various nutritional factors but this is a long way from established. At the moment, then, the bald statement that evolution cannot on its own improve our intelligence seems a plainly silly one to make. Not only is there no evidence for it, the evidence seems against it.

4) If we're to reason about whether surgery or other interventions can increase our intelligence without a decrease in fitness because evolution can't we must first establish that evolution can't. This is another attack on that hypothesis. It turns out a grey squirrel was a much better "fit" for the environment than a red squirrel but in some parts of the world a grey squirrel did not evolve. Similarly it may be that a much more intelligent person would be much more evolutionary fit and, if we reran the clock of evolution, we'd all be ten times as smart and sitting here reading articles about how it's impossible to make us a hundred times as smart.

5) If we're reading the same thing (the first article) loophole 3 appears to be "the intervention may be simple, give major enhancements, but result in a net loss of fitness" -- I have no idea how that is relevant. If our task is "moving over flat terrain" a wheel is a simple intervention that increases fitness for task which evolution has not come up with. So I don't see how that is the same.

6) I think I must be looking at the wrong article loophole 2 here appears to be "the simple interventions may not lead to a major enhancement". I have no idea how that is relevant. What I'm talking about is that evolution is not optimising smartness. It could well be that evolution could make us ten times as smart but that has no gain or loss in evolutionary fitness. If that's the case then it probably won't happen through evolution.

7) "Free efficiency gains in intelligence is never a bad thing" In evolutionary terms neither are they a good thing unless they are needed in the environment. You seem to have a weird idea that evolution increases "efficiency". This is simply untrue... evolution increases only the fitness to replicate genes. This is not the same as increasing fitness. Why on earth would evolution tinker around making us better at digesting things or more muscular unless a really significant proportion of people were starving or dying because they were insufficiently muscular?

(Suspicious comment)
OK -- let's briefly think about the Flynn effect. One central claim necessary before your argument is that evolution is not currently increasing our intelligence. Before you can claim that evolution cannot increase our intelligence you must certainly be sure that this is not currently happening. If evolution was currently increasing our intelligence then let's think about this on an evolution timescale. Let's imagine an increase which would make "average" into "genius" -- say a 50 IQ point increase. On an evolutionary timescale let's say it will take 50,000 years (which is still actually pretty fast in mammalian evolution terms without selective breeding). So we should be looking for an 0.1 point increase every hundred years. Now, can you really say that you're sure this isn't happening? The Flynn effect (whatever it is) is certainly much larger than this. So before you begin your essay on "why evolution cannot increase our intelligence" you must first show it is not doing so. You really cannot -- if there were an underlying evolutionary change on an evolutionary timescale it would be so small it would be absolutely swamped by the astoundingly rapid changes in human IQ.

Now, that red squirrel. The guy who wrote the essay "no simple changes to an animal could possibly make it more hippity-hoppity eat-ity-nutity than a red squirrel because otherwise nature would have evolved the red squirrel to be better" is looking pretty dumb now. Nature had evolved something much better for that ecological niche, it had just done it elsewhere. It turns out that there was a pretty similar animal which was much better at doing the things that red squirrels do but nature had not evolved it. Why not? Because evolution is not optimising what you think it is. It is not in the business of building the perfect creature to occupy a niche efficiently (despite what discovery channel documentaries say). Only a moderate knowledge of biology will find you huge numbers of examples of invasive species which turn out to be much better in the niche than the invasive species. Take home message: what on earth makes you think evolution is good for optimising "smarts"?

Besides, an increase in intelligence by 10 times being completely fitness neutral is about as likely as flipping a coin 10 times and it landing on edge each time

Why -- what is your evidence for this? And remember, fitness here is evolutionary fitness, not fitness for being smart, not fitness for earing money, not fitness for qualifying for Harvard. Do you have any evidence whatsoever that more intelligent people have more children? Unless that huge intelligence increase actually results in more propensity to succesfully rear children to breeding age it's not increasing evolutionary fitness. It may increase a lot of other types of fitness.

> One central claim necessary before your argument is that evolution is not currently increasing our intelligence.

That's not necessary at all. The argument is simple: any increase to intelligence will have one of a few properties, or else evolution would already have increased it.

Intelligence could be increasing... as long as the increases had one of the properties.

> The Flynn effect (whatever it is) is certainly much larger than this. So before you begin your essay on "why evolution cannot increase our intelligence" you must first show it is not doing so. You really cannot -- if there were an underlying evolutionary change on an evolutionary timescale it would be so small it would be absolutely swamped by the astoundingly rapid changes in human IQ.

Already addressed Flynn. So your whole objection comes down to 'there might be some process operating now', which is just an argument from ignorance?

> Why -- what is your evidence for this? And remember, fitness here is evolutionary fitness, not fitness for being smart, not fitness for earing money, not fitness for qualifying for Harvard. Do you have any evidence whatsoever that more intelligent people have more children?

I don't. All my evidence points strongly in the other direction - exactly consistent with what I just said, about the fitness neutrality being extremely unlikely, and increases either being fit or unfit. If merely somewhat smart people *already* suffer big fitness penalties as evidenced by low fertility, then that makes it even *less* likely that being 10x smarter would be exactly fitness neutral!

If evolution is already increasing intelligence then your argument becomes, there may well be simple ways to increase intelligence with small changes and evolution is currently finding them. As the true situation is that we do not know, the only possible argument is "we cannot possibly tell from evolution if there are simple ways to increase intelligence because we do not know if evolution is currently increasing or decreasing it."

I do not think we are going to agree on the other matter of the fitness neutrality of intelligence.

> 7) "Free efficiency gains in intelligence is never a bad thing" In evolutionary terms neither are they a good thing unless they are needed in the environment. You seem to have a weird idea that evolution increases "efficiency". This is simply untrue... evolution increases only the fitness to replicate genes. This is not the same as increasing fitness. Why on earth would evolution tinker around making us better at digesting things or more muscular unless a really significant proportion of people were starving or dying because they were insufficiently muscular?

Because muscles are *metabolically expensive*. Because digestion is *metabolically expensive*. See the footnotes on the chimpanzees, muscles, and guts!

(For god's sake, I'm not putting in all those references, links, and footnotes because they look pretty.)

This whole "metabolically expensive" part highlights how you're thinking of the problem incorrectly. You are absolutely correct that they metabolically expensive are but if they were not, evolution would not go "wahey, free muscles, let's bulk up". If muscles were metabolically free and carried no penalty whatsoever humans would not aquire more unless there were an evolutionary reason to have them... unless they somehow increased our ability to propagate our genes. In the developed world almost nobody is in the situation where more muscle will do this. In the developing world possibly arguably this currently happens (remember we're not talking about situations where someone's life is better or easier because they're strong, we're talking about situations where a significant proportion live to propogate genes because they are).

I think this highlights your confusion between "fitness" in a weird role-playing game sense "Wow, that thing is smarter and stronger, it's obviously better" and "fitness" in the evolution sense "that thing sure can propogate its genes". This is why you automatically believe that a change in intelligence if it has no other costs will change evolutionary fitness.

> If muscles were metabolically free and carried no penalty whatsoever humans would not aquire more unless there were an evolutionary reason to have them... unless they somehow increased our ability to propagate our genes. In the developed world almost nobody is in the situation where more muscle will do this.

Fitness gradients don't have to be big. Even in the developed worlds, pressure could come from various things: from slightly decreased homicide, from the ability to resort to manual labor to earn one's bread, to slight reduction in injury rates.

As people in finance say, a free option is never a bad thing. If using it would not be better than your existing options, well, you just let it expire unused. Muscles are a lot like an option: if you don't need them, you can just digest them and use the protein or calories for something else. See my previous point about additional efficiency in brain construction being useful even if no additional intelligence would be useful.

Well I guess I will never convince you that evolution is not what you think it is.

Algernon's Law - can anyone spot the obvious flaw?

The clear and often quite profound difference between personal disadvantage and evolutionary disadvangate is one flaw that leaps clearly out at me. For example, a species where individuals didn't suffer and effects from aging and didn't die of old age would either have to deal with overpopulation or (if reproduction rates were sufficiently slow) problematically slow evolutionary change. OTOH, from a personal perspective not aging would be awesome. Did you mean that obvious flaw, or another one?

I can see a few other only somewhat less glaring flaws, such as assuming that evolution must have hit upon the global optimum intelligence, rather than merely a local optimum that could be greatly improved upon, or the rather obvious fact that most people with very high IQs don't tend to resemble RPG characters who have to pay for their high intelligence with a host of mental disads.

"The clear and often quite profound difference between personal disadvantage and evolutionary disadvangate is one flaw that leaps clearly out at me"

That's discussed in the article, no?

Indeed; EOC loophole #2. One of the most common.

The argument in Algernon's Law is basically "If there was a simple way to increase human intelligence then nature would already have delivered it."

The flaw is that dogs could say the same thing. Or monkeys could. Or humans could say it about their sense of smell.

We arrived where we are, evolutionarily speaking, through a series of tiny adjustments to fit in well in a specific situation. If the situation has changed (and things like The Flynn Effect and our massively different living conditions compared to our ancestors indicate it has) then small changes might viably improve us in ways that weren't viable in the past.

But there wasn't a simple way for any of those animals to improve their intelligence. There was only a complex way fraught with downsides. Apes paid a huge price to increase their intelligence.

There must have been many, many, simple ways for them to improve their intelligence, one tiny step at a time. They didn't change all at once, after all.

Incidentally, it is worth pointing out also that the Flynn effect is incredibly rapid in evolutionary terms. If we take a really modest estimate of 10 IQ points in 100 years... (some studies have seen more than 3 times that)... now project that onto an evolutionary timescale of (say) 100,000 years. That makes 10,000 points of IQ increase in that span.

OK, it's a totally silly projection (ludicrously so) but the point is that the best current evidence is for an absolutely startlingly quick increase in IQ when we're thinking about evolutionary timescales.

Of course there's lots of "what does IQ testing really measure" sort of arguments to be made.

From my own reading, my best guess is that there is a genetically strong component of maximal intelligence, and then whether we get anywhere near that point is down to our education/experience/food. I suspect The Flynn Effect is about improving these three things.

You may well be right on this. It would be absolutely startling if IQ were improving at that rate for other reasons.

I blame oestrogen in the water supply. It's turning us into women, who are clearly smarter.

I'd make a dreadful woman... the stubble for a start.

> Or humans could say it about their sense of smell.

I'm glad you picked that example. Do you know how dogs get such a good sense of smell? By devoting a huge chunk of their expensive calorie-sucking protein-built brain to it. And keep in mind, dogs already have smaller brains than their wolf forebears - so the instant they no longer needed as much brainpower (because the humans were doing some thinking for both), their brains shrunk. And you are arguing that their brains should have *increased* under their pressures?

The argument in Algernon's Law is basically "If there was a simple way to increase human intelligence then nature would already have delivered it."

But it DID.

I'm not sure that the argument from Jewish genetic defects holds up-- those defects are very rare.

I'm inclined to think that the Jewish advantage is mostly cultural-- because Talmudic study was highly valued for a long time, parents were more likely to see their children's intelligence as an advantage rather than a threat to their status.

It's possible to find unintellectual Jews, though in my experience it's not easy. I've been keeping an eye out for anti-intellectual Jews, and haven't found any.

The defects *have* to be relatively rare for heterozygote advantage to be fitness-improving in the postulated Ashkenazi historical context. If half your infants die of genetic diseases, you need the other half to be basically Supermen who bed Lois Lanes every other day to make up for your losses. At a rate like 1% or whatever the real rate is, then it becomes more possible for a small increase in IQ to have enough of an expected-value to pay for the genetic defects.

In any case, it's just one example.

Why do people leave their religion the Church of England?
I really can't care very much. I'm not a Christian, which makes it easier to see the parochialism of most Christian churches. How many Christians laugh at a "non-denominational church"?

Oddly I'd not heard much about 24/192 downloads until I read the article and then a few hours later, one of my favourite bands spams FB trying to get me to re-buy albums I already own in 24/96 format.

Ta for the perfectly timed heads-up, Andy :)

You are viewing andrewducker