Native fluency: only 1.5 MB of information needed?

[zenmonkey]

we process language as sounds, not as text.

There are lots of people who learn a second language only to read it (and perhaps to write it) but who rarely if ever hear or speak it. So sound is not really involved in storing the information. Yes, we process the sounds, but the knowledge and memory of language is not stored as a sound. There aren't thousands of little sounds buzzing around in our memories. The necessary knowledge and memory must be stored in the brain as some sort of code, and storage of code requires fairly accurately known numbers of "bits" of data. So I think the amount of "memory space" that it takes to store whatever we learn as language (even if we learn it by sound) is interesting. Besides the data for the vocabulary, we have to store knowledge of what things look like (an elephant, a tree, etc.) but that is not counted in the amount of memory required for fluency. A person would have all that recognition of things, etc. even if they grew up totally isolated and never learned a language. Agreed; there would be a difference in the amount of bits require to store the sound of a word compared to the amount required for only the spelling of a word. But there would be a correlation (long sounds correlate to long words). So if a person learns to read, write, listen and speak a language, perhaps the storage requirements are something like 4 x 1.5 MB, plus overhead for grammar, expressions, etc.

Maybe, but the article is not about those people. We don't store sound either. We store electrochemical potentials (which are understood to be managed as sounds, even from reading... there is a thread here somewhere. Edit: see dave's post. ). Bits are binary representations of information - once again, the brain does not store bits. This is what you get when you as a brilliant mathematician to theorise about how the brain function. The von Neumann paradigm was a nice theoretical exercise on how the mind might be considered as a computer - except physiologically speaking, this model is nonsense. There are those today (although I am not a strong believer of this) that the mind functions a quantum level and that, yes, some function is potentiated there. If you are interested in that... then a more modern reading would also include those estimations (which then toss out by several orders of magnitude) that entire paper.

That brain model is the one Penrose and Hameroff developed in the early 1990s and what they call the Orchestrated Objective Reduction (Orch OR) model. Penrose's book Emperor's New Mind talk about this.

When you write MB... do you mean Bytes? Are we talking about units that represent 8 bits, 16 bits or 64 bits? Knowing that a byte can be anyone of those depending on the hardware - I have here devices that have a byte of 64 and 8 bits. So saying 1.5 MB has no sense, if the brain "hardware" (synapse potential firing) is tristate (sometimes) or higher...

As to whether a person is able to recognise all things without language ... that's a totally different subject, and a fascinating one, to which, if memory serves correctly, Oliver Sacks answer with a mitigated "no". The experience of feral children suggests that in many cases, these individuals that do not use language also lose some primary abilities to create general inferences. Often it is impossible to teach them language later in life.

I think it's worth pointing out that this paper is written from an information theoretic approach.

Exactly. And is not about how information is actually managed by the brain/mind.

[tommus]

When you write MB... do you mean Bytes? Are we talking about units that represent 8 bits, 16 bits or 64 bits?

The original linked article says "The study uses the standard definition of eight bits to a byte". The main research article says 12.5 million bits and 1.5 megabytes.

[zenmonkey]

When you write MB... do you mean Bytes? Are we talking about units that represent 8 bits, 16 bits or 64 bits?

The original linked article says "The study uses the standard definition of eight bits to a byte". The main research article says 12.5 million bits and 1.5 megabytes.

So going back to von Neumann model and that the resulting back-of-the-envelope calc then gives you "most computational neuroscientists tend to estimate human storage capacity somewhere between 10 terabytes and 100 terabytes" (A human brain with about 100 billion neurons. Each neurons seems making about 1,000 connections, representing about 1,000 potential synapses => 100 trillion data points, or 100 terabytes of information...)

So ... according to that article, language (and the reference cited in that article) a native level language only used 0.00000015 % (1.5 x 10^-6 %) of the brain. So basically (according to that model), we have unlimited language storage space.

Oh, wait, we can have fun with this. The brain weighs about 3 lbs.
That means that the total weight of that native language storage area is 20.4 picograms (2.04 x 10^-5 g).

You know how small that is?

That's a tenth of the weight of one standard mosquito. 1/10 a mosquito.
That's 7 grains of sand. Seven.

So just taking this to its logical conclusion shows how silly that estimation is. Language uses orders of magnitude more information processing.

Now, Brocka's and Wernicke's areas are somewhere between 10-20 grams (my estimates) ... what the hell is going on with the rest of the area if < 1% of 1% of 1% of that area is involved in language storage?? Are they showing Waterworld on a big screen in there and no one is working? Drunken party boat?

(I apologize to WolframAlpha for the pain it must have felt to do these calculations.)

The amount of theoretical information needed and estimated within this type of information theory has zilch to do with what is really going on in actual language processing and actual memory.

[Serpent]

we process language as sounds, not as text.

There are lots of people who learn a second language only to read it (and perhaps to write it) but who rarely if ever hear or speak it. So sound is not really involved in storing the information.

The original article was about acquiring L1. Also, even if one learns L2 as read-only, they're likely to pronounce things in their head without much regard for the phonetical accuracy (ie using L1 phonetics, or using L2 rules for L3).

[Lianne]

Proof that English is easier.

Indeed. Some other languages like Dutch and German where long, compounded words are common, average word lengths must be considerably longer. Such as:

Dutch
Hottentottensoldatententententoonstellingsbouwterrein
meaning: "construction ground for the Hottentot soldiers' tents exhibition"

German
Bundespräsidentenstichwahlwiederholungsverschiebung
meaning: "deferral of the second iteration of the federal presidential run-off election"

Afrikaans
Tweedehandsemotorverkoopsmannevakbondstakingsvergaderingsameroeperstoespraakskrywerspersverklaringuitreikingsmediakonferensieaankondiging
meaning: "issuable media conference's announcement at a press release regarding the convener's speech at a secondhand car dealership union's strike meeting"

And of course, these are very common words that often occur in normal conversation!

One would tend to think that these long words don't put an extra burden on learning the language because they are made up of a bunch of short, simple words. That works to some extent in reading and listening. However, in writing or speaking, it is not so easy to know or remember which words to put together and in which order. So this just increases the difficulty for active usage versus passive usage. I often wonder about native speakers using such long compound words. Do they even think about them being run together, or do they just say them as an English native would use a series of separate words? I think it depends on the mostoftenusedenglishcommonwordcombinations.

Check out the Wikipedia longest words by language.

Ooh I just have to take this opportunity to share the Ojibwe word for blueberry pie: miinibaashkiminasiganibiitoosijiganibadagwiingweshiganibakwezhigan

It translates as "blueberry cooked to jellied preserve that lies in layers in which the face is covered in bread". Apparently it's generally considered the longest word in the language. And it actually is something one would want to be able to say.