Nik Taylor wrote:

> Eric Christopherson wrote:
> > I'm really fascinated by the idea but I
> > can't figure out how the mind would allow a morpheme to be modified from
> > the inside -- just seems like morphemes "should" be concrete, unbreakable
> > elements to me. It's a bias in my conlanging instinct I guess :)

I'm by no means a morphologist, but I imagine that there are a number
of  ways for this to come about.  Suppose, for example, that you have
two forms ['aSum] and  ['aSulum] (for "hand", say), where [-um] is a basic
syntactic morpheme (say, nominative) and [-ul] is a pluralizer.  That much
is pretty basic, straightforward morphology:  a root [aS-] with one or two
suffixes,  depending on the circumstance.  Let's further assume that some basic
phonological rule comes through that deletes all nuclear vowels syllables in
word final position where accent is not immediately adjacent.  In this case,
that would only apply to the second [u] in ['aSulum] to produce ['aSulm].
Here we are presented with an irregularity:  ['aSum] yet ['aSulm].   There
are two ways a child learning this language can approach this:  he can assume
that the base form of the morphemes realized as [-um] or [-m] is underlyingly
/-um/ and allow for that kind of allomorphy, or, if the pattern is persistent
enough across a wide variety of lexemes, that the [-l-] is in reality a morpheme
/-l-/ that is stuck right into the middle of what, to all other appearances, seems
to be a homogenous suffix [-um].   Either way is possible.  Now then, why don't
we see more infixing in languages?  Because we do see them more frequently
than one might think --  it is extremely common in languages of Mexico and
Central America (though there it is admittedly an areal feature).   If the
phenomenon is mostly absent from modern European languages, that could be
owned up to a dearth of morphology in general, because to have that kind
of complicated process shown above you need at least a certain level of
morphological resources for it to work with.

>I'm not sure either, perhaps it's just spontaneous or something.  After

> all, colloquial vulgar English uses infixes quite frequently, as in
> abso-f***ing-lutely, or a line I read somewhere "Down in
> Tumba-bloody-rumba shooting kanga-bloody-roos"

Frequently?  Certainly not on a daily basis.  Not for me or anyone I know,
at least.

> Perhaps sometimes it's simple metathesis.

Metathesis is only extremely rarely a regular phenomenon.  If a language
makes *regular* use of infixation, then some other process will also have
to be invoked; and the more processes you have to invoke to get your result,
the flimsier the theory must be, by Occam's Razor.

> Another guess is that perhaps it started with only some words.

Well, all linguistic change begins like this.  The question is: what are the
processes driving that linguistic change?  Reanalysis is, as you say, a good
way for that to occur.

Tom Wier <[log in to unmask]>
"Cogito ergo sum, sed credo ergo ero."