Re: Random Name Generator

zaimoni@xxxxxxxxxxx wrote:

STEP 1a: DECIDE WHICH PHONEMES SOUND THE SAME TO A NATIVE SPEAKER. ;)

That way you can have passphrases that the @ can have a hard time
getting right (e.g., lollapalooza for Japanese-native @).

Well, the way to do that is to decide what phonemes are actually used
in languages the speaker knows. Japanese has only one voiced alveolar
approximant consonant; since 'l' and 'r' are both voiced alveolar
approximant consonants, they both map to that same Japanese phoneme.

This doesn't happen at random; the sounds that are indistinguishable
to someone are sounds that are 'alike' for purposes of their own
language, which means they're formed approximately the same way in
the mouth.

Consonants are formed by blocking the airflow from the lungs. Which
consonant, at least in English and European languages, depends on
whether nasalization (passage of air through the nose) is allowed,
where the blockage takes place, and how much the air is blocked.

Here's a table of English consonants: (use a fixed-size font).
From left to right it goes further back in the vocal tract, and
from top to bottom it describes degree of stoppage.

labial l-dental dental alveolar palatal velar glottal
stop p t k
voiced__________b_______________________d________________g________
fricative f th s sh h
voiced_________________v_________th_____z________zh_______________
affricate ch
voiced____________________________________________j_______________
approximant
voiced_______w_______________________r,l__________y_______________
nasal_____________m______________________n_______________ng_________


Notes.

Nasals are always voiced; otherwise they're just silent breathing.

I have abbreviated 'labio-dental' (formed between lips and teeth).

Dental, alveolar, and palatal consonants are formed with the tongue
against the teeth, a point where the shape of the palate changes shape
about a centimeter behind the front teeth, and a point further back in
the palate than that, respectively.

Some languages differentiate a set of consonants formed further back
than the 'palatal' consonants of English, and linguists also call
these 'palatal.' In that context the English palatal consonants
are called 'alveolar-palatal'.

A stop is a stoppage of all airflow. Since this is silent, the
consonant is formed when air is released. These consonants are
also called plosives. Fricatives impede airflow enough to cause
audible friction. Approximants barely impede airflow; just enough
to be heard. an Affricate starts with a stop, like a plosive. But
instead of a sudden release of air, it transitions into a fricative.

You can see what gives Japanese speakers such a problem in English;
We distinguish two different voiced consonants formed at the same
point of stoppage and with approximately the same degree of stoppage!
both 'l' and 'r' are voiced palatal approximants, but the The 'l'
is more closed than the 'r'; to pronounce it you have to have the
tongue nearer the roof of the mouth. Russian and Gaelic use this
distinction a lot more than English, so much that you pretty much
need an extra category in the 'stoppage degree' series to describe
them.

We have two different 'th' sounds, one voiced and one unvoiced.
Although we think of these as the same, they are not, and this
is a point where English speakers face the same problem in other
languages that Japanese speakers face in English.

Anyway, different languages have different sets of consonants,
some the same consonants that English has and some not. A lot
of the boxes that English leaves 'empty' in the table above are
used in other languages. (for example, the velar fricative, or
guttural 'ch' that actual knights used to pronounce the word
'knight' (middle English 'knicht') is now conspicuously absent
in English, but German still uses it (as in 'Bach' for example)).
And whatever sounds a language doesn't use, tend to get grouped
with whatever 'nearby' sounds that language does use. Almost
always, sounds that get confused will be sounds adjacent to
one another in the above chart.

Also, some languages use distinctions between consonants that
English does not recognize at all. Chinese and Hindi use
Aspiration, whether a stop is released with a 'puff' or just
released, and this gives them additional 'stop' consonants
that English speakers cannot usually distinguish from 'p',
'b', 'k', 'g', 't', and 'd'.

Languages tend not to have more than 20 or 30 consonants;
otherwise they start running together and it begins to be very
hard to pronounce words so that they remain distinct from
other words. In this chart, almost every consonant in English
has 'empty' space on at least two sides. That's why you can
understand English in a high wind, mumbled by a drunk, shouted
from a hundred meters away, or in the middle of radio static.
The restricted choices mean sounds that are lost or distorted
have only a limited number of ways to be resolved; in context
rarely more than two. Lots of different consonants are a
luxury for languages that rarely face interruption, noise,
stressed speakers, shouting, slurring, or distortion.

All this is supplemental information provided by a confessed
hardcore language geek; actually using it in a roguelike game
would please hardcore language geeks the world over, but is
probably a lot more work than gameplay justifies.


Bear

(ps. Vowels are much more complicated than consonants, but
also interpreted much more forgivingly by most listeners.)


.

Loading