we take English by rrrkakushi


									     If we take English to be a thirty-phoneme
language        (it     has     more       than
thirty, no matter how you figure them), permit
no           morpheme          shape         of
more than five phonemes {glimpse/glimps/
actually has six), and assume
that only one out of every 1,000 possible
sequences can be used, we still end
up with a total of 25,137 possible morpheme
shapes (the above 25,137,930
divided by 1,000)—enough to take care of any
    If   we   remind   ourselves   that
English words can easily consist of
three or more morphemes (e.g. un-
friend-li-ness), it is clear that we are
also      provided        with       an
overabundance of possible word
shapes—more       than    enough    for
Lewis Carroll to invent "slithy toves
did gyre and gimble in the wabe,"
using a few of the thousands of
available word shapes which had not

previously been claimed.


     In the preceding paragraphs we have
assumed, for purposes of presentation^ that a
message is neatly encoded first semantically,
then grammatically, and then phonologically.
But since normal speech is full of false starts,
hesitations, grammatical slips, and the like, it
seems clear that we behave a good deal more
like thef young lady who, when told that she
should "thinksbefore she spoke," replied with
rare honesty: "But 1 can't do that! How do I
know what Fm going to say until I start
     Ilf we do not normally plan out our entire
message before we start sending it, then we
must possess some sort of feedback device
which permits us to "monitor" the message as
it is sent and to make necessary adjustments
as we proceed—adjusting a present tense
verb to agree with its singular subject, for

To top