The
comma is a highly logical marker. When it demarcates sequential nounal
elements, or fails to demarcate them, it is doing so for a
meaning-making purpose. In this sentence:
My
son Andrew is a law student,
the
two nounal elements `my son' and `Andrew' are not demarcated, which
leaves them in apposition, such that they act upon one another as noun
and adjective: `Andrew' effectively describes `my son', or vice versa.
This adjectival capacity of each noun prevents either from
functioning as an undistributed naming element. So the logical
template of this sentence is: `Some things that are `my son' are some
things that are `Andrew'. Since some things that are `my son' are
some things that are `Andrew', and vice versa, Andrew is one of
my sons, not the only one.
However,
when the comma breaks the apposition, the nounal elements cease to impact
upon each other as a noun-adjective combination. Rather,
`my son' and `Andrew' are alternative names for the one entity:
My
son, Andrew, is a law student.
The
`one entity' logic of this naming sequence entails that `my son' is the
undistributed noun-subject, such that the logical template of this
sentence is: `All things that are `my son' are some things
that are `Andrew'. (Clearly, `Andrew' cannot be undistributed: it is not
an exclusive name.)
Now, since
all things that are `my son' are `Andrew' , I have no son but
Andrew.
(This
point is perfectly clear in propositional logic. Theory of syntax has
been rather tardy in picking it up, despite the fact that comma usage is
so highly dependent on the distributed/undistributed distinction.)