Essences traditionally perform 3 metaphysical roles.

R1. As regulative principles. Thus the essence of a chicken prevents it
from
turning into a tree.

R2 As sortal kinds. The essence of a chicken is what allows us to
recognise
it as a chicken rather than something else.

R3 As an explanatory factor. The essence of a chicken is what makes it a
chicken rather than something else. It tells us why chickens are
chickens.

Now, I will pass over R3 with the comment that I find it rather
incomprehensible,
and focus mainly on R1 and R2.

There seems to be a significant tension between these two roles.
Consider the following
facts:
F1: Individuals (from a common-sense perspective) change considerably
in their properties over time.
F2: Sortally speaking, we tend to group things by their properties.

Consider a chicken and an egg. The individual chicken comes
from the individual egg, yet has almost no properties in common, with
it.
So do we say that there is a an essence which the individual egg
has in common with the individual chicken, which regulates the
transition
from egg to bird, or do we say that eggs are grouped with eggs
by common properties (meaning that there is a generic essence of
eggness),
and chickens likewise grouped with chickens.

There would seem to be three approaches to the issue:
MA The Middle-Ground approach,
AA The Atomic approach,
UA The Universal approach.

To be more specific:

MA Individual things have individual essences which survive
some but not all changes of property (eg Man's essential characteristic
of Rationality does not survive the loss of the property of life).
AA The essence of a thing changes whenever its properties change.
UA There is one essence, the Essence of the universe.

Let us take each approach in turn and examine what it has to offer.

MA Individual things have individual essences which survive
some but not all changes of property.

Advantages:
+1 Supports our common-sense intuitions about individuals.
Disadvantages:
-1 How is a thing's ceasing-to-be regulaated ? What happens in the gap
between one essence (chicken) and the next (sunday dinner) ?
-2 It necessitates a theory of natural kkinds; that is,
if things really are categorised by there essences,
we ought to be able to come up with a semantics,
or categorial schema which is inherently 'right'
and non-arbitrary. Yet this seems difficult in practice.
Does the Eskimo's seven kinds of snow mean there are
seven snow-essences, or only one as the European says ?

AA The essence of a thing changes whenever its properties change

Advantages:
+2 No problem of natural kinds.
Disadvantages:
The problem (-1) of 'regulative gaps' re-appears in a much stronger
form.
-3 Redundancy. Essences perform no episttemic or metaphysical
role that is not performed by properties.

UA There is one essence of everything

Advantages:
+2 No problem of natural kinds.
+3 No problem of regulative gaps.
Disadvantages:
-3 Performs no useful sortal role.

Now, MA is the traditional view of essentialism, and it has two things
that need fixing: An answer to 'regulative gaps', and a theory of
natural kinds.
AA is not properly an essentialist theory, since it has no proper
regualtive
role.
Moreover its sortal role is redundant, so it seems hopeless.
UA is not properly an essentialist theory, since it has no proper sortal
role.
However, that issue has an existing fix, namely a nominalist theoru of
kinds.
Since a nominalist theory can be defined how we like,
we need have no fears that it will not conform to our intuitions.

I propose that the universalist approach, in conjunction with the
nominalist fix is the best approach, and note that it is the
metaphysics of science. The Universal Essence, as a regulative principle
is the system of physical law. The metaphysics of science is
nomic realism and sortal nominalism.

PS I am aware that the above treatment is somewhat sketchy, and hope to
return to it at some point in
the future.

Peter D. Jones 10/8/01