News for hackers. Stuff that matters.
Re:Pirahã: a non-Turing-complete human language by newhouseb on Wednesday January 06, 2010 @12:27AM (#1034231)
I was talking to a friend and they said that in order to quantify objects this tribe counts on a logarithmic scale and not an ordinal one, which I found quite cool, since there's really no reason why we must count in one or the other (logarithmic or ordinal).

This seems counterintuitive at first, but logarithms are just ratios and henceforth sometimes actually more useful than ordinal counts - for example, when comparing say, land mass sizes and one plot is 1sq mi larger than the other. Without the context of the size of the plots, this is relatively useless (if they are huge, they're about exactly the same, if one is small then the other is massively bigger). If we say plot B is 1 log unit larger than plot A then it is twice (if base 2) bigger than plot A, no extra context required.

I thought it was cool...

Re:Pirahã: a non-Turing-complete human language by yannis on Wednesday January 06, 2010 @01:07AM (#1034282)
It is surprising, but the human brain is wired to think logarithmically, it is only by years of training that our brains get re-wired to do otherwise. Stanislas Dehaene has done a lot of research in this area (http://en.wikipedia.org/wiki/Stanislas_Dehaene). The study of how the brain interprets continuous and discrete signals (for example sound and numbers) has led to the discovery that the neural representation of number is comparable to the slide rule calculators, which was also graduated with a logarithmic scale. See (http://www.unicog.org/publications/Dehaene_WeberFechnerNiede...)

It is very efficient to do calculations logarithmically and evolution appears to have optimized our brains (and that of a few other animals) in this respect.

Re:Pirahã: a non-Turing-complete human language by stcredzero on Wednesday January 06, 2010 @01:27PM (#1035297)
It is surprising, but the human brain is wired to think logarithmically

Then again, we perceive forces, brightness, and loudness logarithmically. (This is not just the subjective experience. The way we are wired to our sound, light, and kinesthetic receptors is indeed logarithmic.) Why shouldn't we also naturally think that way?

Re:Pirahã: a non-Turing-complete human language by gphil on Wednesday January 06, 2010 @12:56AM (#1034265)
There is an episode (entitled "Numbers") of WNYC's Radiolab (available for free as a podcast at http://www.wnyc.org/shows/radiolab/) that talks about the possibility of an innate ability in infants to interpret quantities on a logarithmic scale, which is later supplanted by the ability to count ordinally. I haven't looked into how much evidence there is to support this theory, but the anecdote offered in the podcast is thought-provoking.
Re:Pirahã: a non-Turing-complete human language by elblanco on Wednesday January 06, 2010 @01:09AM (#1034283)
I actually came here to post the same link but you beat me to it! I also remember seeing some of the research related to humans estimated comparisons between two things, and that we naturally use a logarithmic scale to do this...but I couldn't remember the details of the paper.