Spoken Ancient Greek was considerably slower than modern

[jeidsath]

A new paper suggests that human languages converge to 39.15 bits/second.

https://advances.sciencemag.org/content/5/9/eaaw2594

Italians deliver speech at about 9 syllables/second, but have fewer separate syllables than Germans, who speak at about 5-6 syllables/second. Looking at 17 languages, all seemed to converge at 39 bits/second.

The natural consequence for ancient Greek, is that as vowel collapse made syllables indistinguishable from each other, speakers would have have sped up. I would want to do some detailed calculations, but the vowel collapse would probably have taken Greek from the low end of the chart for syllables per second, up to the high end. And in fact, Modern Greeks currently speak at 8-9 syllables per second. The syllable rate of ancient Greek might have been more similar to a language like Vietnamese, with 4-5 syllables per second.

This suggests a pedagogical tip for modern Greeks learning the ancient language: Slow Down

EDIT: I know less about it, but the high syllable rate of Italian compared to other languages, and the loss of a vowel length distinction since Latin makes me suspicious that the same could be true there. Perhaps Bedwere could comment.

[bedwere]

It seems to me that English is faster than Italian, but it may be my impression. However, I was trying to do some "copyright violation" by copying the Julius Tomin's reading of the Gospel of St. Luke as soundtrack of this movie with the whole KJV and found it impossible: the English is too fast compared with the Greek.
In any case, let's try to observe length for pedagogical reasons.

[Lukas]

When you write "bits/second," do you mean computer bits or some other measure?

[jeidsath]

An information theory bit, which is the same as a computer bit, yes. Claude Shannon defines "bit" in his famous 1948 paper, giving credit to Tukey (inventor of the FFT) for the term.

[Callisper]

The natural consequence for ancient Greek, is that as vowel collapse made syllables indistinguishable from each other, speakers would have have sped up. I would want to do some detailed calculations, but the vowel collapse would probably have taken Greek from the low end of the chart for syllables per second, up to the high end. And in fact, Modern Greeks currently speak at 8-9 syllables per second. The syllable rate of ancient Greek might have been more similar to a language like Vietnamese, with 4-5 syllables per second.

What's the logic for the numbers?

For AG > MG loss of tonal complexity is another contributing factor.

This suggests a pedagogical tip for modern Greeks learning the ancient language: Slow Down

First you'd have to get them to pronounce the distinct AG vowels as distinct

[RandyGibbons]

It seems to me that English is faster than Italian, but it may be my impression.

Interesting, Roberto (bedwere). I experience it the other way around, which may just be that when we are learning a second language we want the native speakers to slow it down a bit?!

Though Joel's data seems to suggest Italians speak rapidly. To me that's not a function of vowels but of what I recently wrote in another post: Italians are hot-blooded .

[jeidsath]

Well, the point of the paper is that everyone, in Italian and English even, probably express ideas at the same rate, and that syllable-rich languages have a lower syllable rate per second in consequence.

The loss of tonal complexity form ancient Greek to Modern wasn't likely too big a factor. It changed to stress, and except for a very small number of cases, the difference remained "marked" versus "unmarked" syllables, which would have had the same information value.

However, the vowel collapse was pretty stark. The modern language only has 5 vowel sounds, while the ancient had 5 short + 5 long + 8 or so diphthongs. During this time there were not a bunch of new consonant combinations brought into the language. Someone would need to actually count up the number of different syllables in an average Modern Greek text, versus an average Ancient Greek text, but the vowel compression seems to make it likely that there are fewer distinct syllables in Modern.

Bedwere's point is a good one. If it's idea rate, we might be able to make the comparison between syllable number in translations and original texts (though "natural expression" is hard to achieve in translation). I think that I've seen M. L. West or Allen mention some numbers here, though I forget where.

[donhamiltontx]

Just the one study?

[jeidsath]

Just the one study?

This is a big replication of an earlier 2011 (?) study that found approximately the same thing, but now done with more data and precision.

The theory seems to be, that since people can listen to language input at much faster rates (you can speed up an audiobook to prove this to yourself), and since the human speech system is capable of producing faster sounds, the real limitation is how fast the brain can string ideas together.

EDIT: Here is the 2011 study