Re: low entropy text

[Dennis <ixohoxi@xxxxxxxxxxxxx>]

Jorge Stolfi wrote:
> 
>     > [Dennis:] Twenty characters sounds low for a Romance language.
> 
> Well, Italian, Spanish, and Portuguese don't use "k", "y", "w" (except
> for foreign names and derivatives), so the "official" alphabet has
> only 23 letters. Moreover, Italian hardly ever uses "x" and "j",
> uses "q" only in historical spellings, and is quite stingy on
> diacritics.
> 
> So a 20-letter "effective" alphabet for Rhaeto-Romance (whose spelling
> is based on the Italian one) is not so out of range. (Assuming, of
> course, that space and punctuation is not counted.)

	Quite so.  I think that the Romance languages actually
have fewer vowel phonemes than the Germanic ones.  I
believe that Italian and Spanish only have the five
Latin vowels (but you mention open and closed vowels
for Italian) and do not have the distinction of vowel
length that Latin had.  French has more, about nine. 
By contrast, English has about 11, German has 14, Dutch
probably the same as German, etc.  The Romance
languages look vowel-rich because they don't have the
consonant clusters of German etc.  (I was thinking
about all this in working on Hamptonese.  Hamptonese
seems to have ~11 vowels and ~23 consonants.  So it's
probably phonetic English.)

> 
>     > Could it be due to a lossy spelling?
> 
> For Italian, it is partly a case of lossy spelling (no indication of
> stress; no distinction between "open" and "closed" vowels, as in
> "bello/nero", "sotto/notte"); and partly the use of digraphs for some
> phonemes ("sci" for "sh", "gi" for "j", etc.).

	This isn't terribly lossy.  English orthography is far
worse, in every respect.
 
> F'rinstance, it would be a chance to start paying that pizza bet,
> which I am now afraid that I will eventually lose...

	I don't know about that.  I'm starting to incline
towards Chinese, although I'll check out my medieval
French hypothesis as soon as I can.

Dennis