Root versus lexical languages.

[s455152 at aix1.uottawa.ca]

Dear List members--

I distinctly remember having been taught, as an undergraduate, that one
can distinguish ROOT LANGUAGES from LEXICAL LANGUAGES. The latter are the
more common type and consist of those languages (like English) where, as a
rule, lexical items stand in isolation from one another, i.e. derivation
is not regular or transparent. Root languages, on the other hand, are
languages like Sanskrit or early Semitic languages, where derivation is so
regular and transparent that, by taking a small number (800 to 1000) of
roots, one can generate the bulk of the lexicon: this transparency is such
that, for example, in an Arabic dictionary, roots, not words, are what is
listed.

My question is twofold:

1-Can anyone point me to published work comparing these two types of
language,

and

2-While the transformation of a root language into a lexical language is a
banal, commonly observed phenomenon (from Sanskrit to the modern
Indo-Aryan languages), is anything known about the reverse, i.e. how a
lexical language turns into a root language?

Many thanks in advance,

Stephane Goyette,
University of Ottawa.