reply on: Stacker Saloon \ stacker news

pull down to refresh

102 sats \ 5 replies \ @ek 26 Aug \ parent \ on: Stacker Saloon

what do you mean by that?

Alphabetic letters are a way of turning speech (which is a kind of fuzzy set of modulating audio waves) into a set of discrete digits. This was a surprising thought to me because "digital" feels like such a modern invention.

And really, alphabets aren't just digitizing speech, they're digitizing language, and since language plays such a big role in how we think these days, alphabetic letters become a way of digitizing meaning.

But humans may not have digital minds in the way that a computer does -- if what a computer is can be said to be a mind.

Another interesting place to take this is that non-alphabetic languages (like Chinese) might be better suited to human thinking? (this is now me wading way out into pure speculation without really thinking any of it through). Also, apparently numbers are considered ideograms.

123 sats \ 0 replies \ @nullama 17h

This made me think of oral languages.

These languages don't have a written form, so they might pose a challenge to AI. In theory a system could be trained to pick up the audio cues, but maybe it would miss the subtleties of the speaker.

142 sats \ 2 replies \ @adlai 26 Aug

I think the word that more accurately conveys what you probably mean when you write "digitizing", is "discretizing". The example of tonal languages is great, because musical signals are definitely decomposable (both by wetware brains and by electric circuitry) into discrete syntactic elements, regardless of any digitisation of the signal.

0 sats \ 1 reply \ @Scoresby 26 Aug

But can I say that alphabetic writing is discrete while ideographic writing is not? I wanted to get at the distinction that alphabets take our language and turn it into a series of digits.

162 sats \ 0 replies \ @adlai 26 Aug

I suspect that if you follow this distinction out all the way, you'll find that ideographic writing is also a collection of symbols that can be considered digits, although the base is much higher... while English might have an effective base around thirty (if you allow for some punctuation), Kanji has an accepted common base of a few thousand and if you held some professor's feet to the proverbial fire, you could probably establish some mathematical bound for this base.

Check out syllabaries. Some writing systems are really quite close to an extremely compressed trace of the analog signal, while still being a linear sequence of symbols.