Results of Poll by Email No. 9

        There was a slight bit of confusion over this poll, due to a lack of
explanation and my confusion of phonemic vs. phonetic (hey, I write these
polls on Saturdays, when all my little grey cells start to resemble a frat
party gone horribly, horribly awry), but that did not prevent 28 of you from
responding. When asked how large your phonologies are, you answered thus:
        My conlang's phonemic inventory is:
        A. Gigantic (+100) (1 responses, 4%)
        B. Enormous (+75) (1 responses, 4%)
        C. Huge (+60) (0 responses, 0%)
        D. Expansive (+45) (2 responses, 7%)
        E. Large (+30) (9 responses, 32%)
        F. Average (+20) (11 responses, 39%)
        G. Small (+15) (3 responses, 11%)
        H. Tiny (+10) (1 responses, 4%)
        I. Wee (+5) (0 responses, 0%)
        J. Binary (0 responses, 0%)

        Andreas Johansson wanted to know, "Is assigning the 20-29 span as Average
based on linguotypological data?" To which I respond: I don't know. For some
reason, I had that number in the back of my head as "average," so either I
heard that somewhere, or I was very lucky, seeing as the greatest number of
responses landed in that category. Of course, if we tried to find the true
average (which I can't do, since not everyone sent in exact numbers), given
the large number of "Large" responses, the true average would probably be
around 30 or so.

        Roger Mill expressed his preferences: "I would tend to stay in the 25-35
range (Gwr e.g. has 24 C, 9 V).  I'd prefer to work on a language with a
small inventory rather than a large one (typing problems at the least;
romanization problems, and I don't like X-SAMPA. I don't at the moment have
terribly elaborate allophonic rules, but morphophonemics is fun...within
limits."

        Dan Sulani revealed his process of phoneme creation: "Unlike many other
conlangers, when designing rtemmu, I didn't develop a complete phonological
system as a  first step. I preferred to concentrate on the morpho-syntactic
and semantic aspects and to give the phonology pretty much a free rein. What
I did choose was my basic sound inventory. Since there turned out to be 33 of
these, my answer to the poll is "E".
        "The (allo)phonology is in a constant state of flux because I structure new
words mostly according to whim. Since what is allowed in a word mostly
depends upon what I like at any given time, allophonic rules for rtemmu as a
whole are likely to change with each new vocabulary item."

        And Rosta had this to say: "The number of phonemes is in the eye of the
analyst, and I often suspect those who begin a conlang with a plain list of
phonemes of having a certain phonological crudity -- first invent the
phonetics and then, through analysis, discover a phonology. The point is
this: with a natlang, the data are far more certain and definite than the
linguist's analysis of them, and this is the way I like conlangs to be too.
        "Anyway, to answer your question, Livagian, by my analysis,
comes in at (E).
        "But to answer Jesse's questions:
        "The phonology is large in that, by conlang standards it is elaborate,
though not by natlang standards. I prefer languages with phonetic richness.
Livagian uses elaborate allophonic rules -- the relatively parsimonious
phoneme inventory gives rise to a much much larger range of phonetic
exponence."

        Thanks to everyone who responded, and stay tuned for Poll by Email No. 10!
        :Peter