I identify the DNA/gene as non-arbitrary and fixed only in the
limited sense that unlike language where there is no intrinsic
relation between signifier and signified, DNA at least has the
statistically significant tendency for its acids to pair only with
certain other acids. This is not to say that what parts of DNA or
chromosomes may function as a gene is not flexible, nor widely
variant, nor that there is not a wide range of alternatives always
changing in the biosphere.
I speak of determinism only on the molecular level. I would suggest
that everything else is quite a complex moving open system.
There can be no such thing as empiricism as long as language is
involved in any stage or human observers intercede in measurements
that are made with tools requiring interpretation. Social sciences
are particularly problematic in this regard since they often envision
issues of language and representation as countable incidents based on
protocols that do not take the vagaries of representation itself into
account. Of course, not all of them do this, but enough.
In so far as anyone ever conceived of DNA as a code, it was always
already paradigmatic, since Shroedinger referred to some agency as
that before DNA's structure was even discerned (an dby someone who
had read Schroedinger). Calling DNA a code certainly does not define
how it functions, but the paradigm does condition how we think about
DNA--and how some scientists approached it--i.e. "decoding." DNA is
not a code. It is a chemical. If we need an analogy, cipher is the
better one.