Scriptio continua | Hacker News Reader

Scriptio continua

en.wikipedia.org

38 pointswolfi113 comments

Does spoken language have word boundaries? In my experience learning foreign languages (at the simplest levels) it sure feels like part of the difficulty comes at picking words out individually, and the pattern recognition only really kicks in around the phrase or sentence level.

Sort of, sort of not. It depends on the language. English for example is mostly a "stress-timed" language, which means the time between each stressed syllable is more or less perceived to be constant. Since there is usually 1 stress per morpheme, you can recognize when you are now in a new morpheme, though not at the "boundary". Morpheme is not quite the same as word, though.

here's a rabbit hole for you: https://en.wikipedia.org/wiki/Word#Definitions_and_meanings

There's often some indicator. E.g. some consonant clusters might only appear at the beginning of the word, never in the middle of it. Or stress might be regular enough to tell you where the last word boundary was or where the next one will be.

But in most natural languages, there are usually enough exceptions that you still have to rely on context to parse.

One thing I've noticed when changing TV channels was that I never seem to hear a half-word. I always seem to just register the first full word I hear, and my brain doesn't even pass onto the concious layers that incomplete stub of the word I caught in the middle of being spoken.

There must be some word boundary, that is hard to define though.

There are languages with regular stress, for example in Polish stress almost always falls on second-to-last syllable. So when you hear stress you know the word is about to end and the next word will soon begin.

In practice it's not 100% true, there are regional differences (in Kraków people stress the first syllable instead), so it's not the only factor.

Huh. I was just thinking something about how terrible that seemed, when I got to "Modern Latin script" which points out that we do this with ex. domain names - thisisanexample.mycompanyname.com is intelligible without spaces, even if it does take a bit more work to read (also a decent example of the problems with it - glancing back up, my brain tries to pick out "i sane" or something about a SAN before deciding that "is an" is the right segmentation)

Worked out well for expertsexchange.com…

That’s how they do it on Pen Island.

This also lead to some confusion https://en.m.wiktionary.org/wiki/busillis

Relevant: https://en.m.wikipedia.org/wiki/Boustrophedon

One little annoyance I have with reading books is when the text is small it's hard to track back from the end of a line to the next line's start. A bit of a shame boustrophedon isn't used.

Small nitpick that Japanese is not exactly scriptio continua, as the particles serve as word separators.

Reminded me of the game Ni No Kuni. This was referenced in its in-game magic book.

Reply on news.ycombinator.com