It turns out that a language with only six phonemes is certainly possible,
if a little awkward (assuming that you can detect word boundaries by their
indirect effect on the phonology). Three vowels and three consonants seems
to be better than two vowels and four consonants. Choosing consonants that
cluster well is also an advantage. I've been playing around with this idea
for the last few days, and the result is a language called Kisuna.

Two dissimilar vowels or consonants can cluster together, but the initial
and final consonant clusters are somewhat restricted: only sk- and sn-
appear initially, and -ks, -nk, -ns, -sk finally. In a few cases, I've
allowed three vowels to come together: word final -iau (originally -ial) is
the most common.

There are a number of difficulties with this limited phonology. If you want
words to be reasonably short, you end up using most of the possible words,
which increases the likelihood for misunderstanding if a sound is misheard.
To pick a word at random, here are all the words that differ by only one
phoneme from the word "kansai" (chalk):

kinsai "wine"
kaksai "cart"
kankai "church"
kansui "to control"
kansau "insult"

However, because there are so few similarities between the Kisuna phonemes
to begin with, confusion is unlikely. And because there are so few sounds,
there are fewer *possible* ways to mishear a word than in a language like
English. But even though the phonemes themselves don't sound similar at
all, the words generated by combining these phonemes end up looking
confusingly alike.

Still, a quick check of JWPce's handy Japanese dictionary lookup reveals
that Japanese has two words pronounced "kansai", one "kankai", and FOUR
"kansui". In addition, Japanese has a "nansai" and three "sansai". If
Japanese can get by with this degree of packing in the same arbitrarily
picked area of "phonology space", perhaps Kisuna isn't as unnatural as it
seems at first.

Another difference between Kisuna and English is that long words aren't
analyzable into components as their English equivalents often are.

English               Kisuna

tele-vis-ion          kaninus
tele-scope            kaukia
micro-scope           kusnins
civil-iz-ation        nasakis
earth-quake           nauni
hour-glass            nisisin
for-ward              kinian
back-ward             sisnai
basket-ball           inisku
foot-ball             kanuinai
base-ball             kusuisi

Even with the difficulties of working with such a restricted phonology,
there are some exciting new developments. By chance, a number of the
randomly generated words appeared to be derived from related words by
adding the suffix -sa. With a few minor adjustments, it was possible to
generalize this to a suffix that derives nouns from adjectives! (To do
this, it was necessary to make it illegal for adjectives to end in -s.) By
similar adjustments, it should be possible to allow for additional
derivative suffixes, such as -ksu or -nus.

Additionally, due to the way the sound system evolved as I switched from
four consonants and two vowels to three of each, certain initial sequences
of phonemes are unused. Thus, derivational prefixes such as suku- or akia-
are also possible. My initial plan was to create an isolating grammar for
Kisuna, but the ability to add prefixes and suffixes to words could allow
for a certain degree of inflection.

Compounds of two or more words are also possible, although care must be
taken with words of less than two syllables as elements of compounds. As it
is, the components of compounds are likely to be ambiguous. Here are a
couple of examples:

"kananusink": "kanan" (flat) + "usink" (horn) = "moose"

"nausinasis": intended to be "nausi" (long) + "nasis" (neck) = "giraffe",
but could also be: "nausin" (reject) + "asis" (shave)

languages of Azir------> ----<>---
    h i l r i . o         "If all Printers were determin'd not to print any
     m l e @ o c m       thing till they were sure it would offend no body,
   (Herman Miller)       there would be very little printed." -Ben Franklin