Welcome, Chris.
The syllable structure of Japanese is very
rigorously constrained: a consonant that closes a syllable must be
identical with the consonant beginning the next syllable, or be a nasal agreeing
with the following consonant as to its place of articulation. Only the nasal
spelt "n" may occur word-finally. It follows that the only consonant
clusters allowed between vowels are sequences like -p.p- or -m.b- (periods
mark syllable boundaries), while -p.t-, -s.k- or -r.t- are
prohibited:
kit.te 'stamp'
gak.koo 'school'
tan.sen 'edge'
tom.bo 'dragonfly'
No Japanese syllable may begin (or end)
with a cluster (not even in loanwords, hence supootsu = English "sport"
with Japanese-style remodelling). But even languages that allow syllable-initial
or final sequences of consonants typically impose rather tight constraints on
them. In English, for example, a syllable-initial sequence can't contain two
obstruents (stops or fricatives) other than st-, sp- or sk-, and an obstruent
(or one of these three sequences) may only be followed by a non-obstruent. There
are also a number of more specific constraints, and their overall effect is
that syllable-initial br-, pl-, skw-, hj-, sm- or tw- are permissible, but
tn-, ks-, pt-, mn- (which are OK in Classical Greek, for example) are ruled
out.
Polish, Russian, and other languages which
seem to allow just about anything word-initially, have their own structural
limitations. In Polish, for example, a seeming monstrosity like pstr- is
permitted (and occurs in several words), whereas ptsr-, sptr- and ptrs- (or
any other of the remaining twenty permutations of these four segments) are
not allowed even potentially.
Well-formedness constraints on syllable
structure are arguably universal, though their relative strength
is language-specific. For example, all languages "dislike" syllables
without an initial consonant, and all languages "dislike" syllable-final
consonants, but these constraints do not function as absolute prohibitions
unless they happen to dominate all other structural requirements in a given
language (in which case we get a language in which all words have one of
the shapes CV, CVCV, CVCVCV, etc., while CVC, VC, VCCV, CVVC etc. are
banned). A language in which the preference
for open syllables and the horror of syllable-initial clusters are both
high-ranking will sound like Hawaii.
More often constraints manifest their
existence as relative preferences. They will be violated very frequently in
order to satisfy more important demands, but they may "adjudicate" between
competing word shapes that are minimally different, favouring one of them to the
exclusion of the other. For example,
statements under (1) are not universally true (in fact, they are violated
in most languages), but those under (2) are:
(1)
... A syllable must begin with a
consonant.
... A syllable mustn't begin with two
or more consonants.
... A syllable must be open (not
closed by a consonant).
(2) OTHER THINGS BEING EQUAL:
... A syllable with an initial
consonant is better formed than one that begins with a vowel.
... A syllable beginning with a single
consonant is better formed than one beginning with a cluster.
... An open syllable is better formed
than a closed syllable.
Piotr
----- Original Message -----
Sent: Saturday, November 25, 2000 5:30 PM
Subject: [tied] Consonant-verb regularities
I wonder if you all might indulge a neophyte question. Now,
some
languages, such as Russian, have lots of consonant combinations.
Others, such as Hawaiian, use vowels more heavily. Some seem to
insist
that every syllable be a consonant followed by a vowel. I
sense
a number
of vague patterns here. Are there any regularities in the
relative use of
vowels and consonants? Do most languages have some
sort of ironclad rule
regarding vowel-consonant relationships? For
example, Japanese, I believe,
seems to follow the rule that each
syllable consists of one consonant
followed by one vowel -- but there
are a few exceptions, I believe. Do such
exceptions tend to follow a
rule?
I suppose that the thrust of my
question is, are vowel-consonant
relationships fundamental to a language,
and therefore subject to
strict rules, or are they more a matter of local
fashion, with lots
of random exceptions?
Thank you for your
indulgence.
Chris
Crawford