Curtis Clark lists at curtisclark.org
Wed Nov 16 13:42:40 CST 2011

```On 11/15/2011 9:35 AM, Richard Jensen wrote:
> OK - I was not precise enough: In the realm of Euclidean geometry, an
> obtuse triangle is a two-dimensional construct defined by the
> possession of a specific set of attributes.

Now you've turned it into a human construct, by restricting it to
Euclidean geometry. But I digress.

> The attributes determine membership, not vice-versa.  We cannot
> recognize a set unless we specify the attributes that members of that
> set must have (i.e., have defined the set). Once we have defined the
> set, we can then ask if any particular entity belongs to that set.  But
> we can't assign membership until the set is defined.

I don't disagree that attributes determine membership; I disagree that
those attributes constitute a definition. In formal deductive logic,
sets are defined, but in practice the attributes often shift in an ad
hoc manner, so that the "definition" arises only post hoc, when the need
arises to commit something to words.

> That's exactly why definitions matter. As you note, "certain
> definitions" may require recognition of HeLa cells as a separate
> species. Other definitions would not require that view. If we want to
> know if we are talking about the same thing, we have to agree on the
> definition a priori (or, a posteriori, we come to agree on one of the
> available definitions so we can make sure that our future
> conversations avoid the confusion that might arise). Regardless of the
> situation, definitions matter.

I think we may be using different definitions of "definition". I'm
looking at the Wikipedia article now, and it matches what I remember

A definition of a species can only provisionally be intensional: the
nature of science is such that the necessary and sufficient conditions
can never be known (since they can never fully be discovered by
induction). A definition of a species can quite nicely be extensional
(that's what I meant about the membership defining the species), but we
lack the breadth of knowledge to ever enumerate all the members, and we
use attribute-based decisions to decide whether individuals are members
or not.

It seems to me that taxonomists make heavy use of ostensive definitions.
The type method in nomenclature is based on that, and identification
guides use  intensional pseudo-definitions in the keys and ostensive
definitions in the illustrations.

I don't regard "definition" and "hypothesis" to be synonyms, so I still
assert that "definitions" of taxa don't meet the formal criteria of
definition, and are rather restatements of hypotheses.

> Isn't that what taxonomists are doing when declaring one species to be
> a later homonym of another? Or, when transferring a species to a new
> or different genus? A hypothesis has been falsified and rejected. In
> most cases, it's not that the attributes have changed, it's that our
> interpretation of the attributes has changed, or that we have new
> information about the group - the definition is more specific.

The hypothesis is more specific.

--
Curtis Clark                  http://www.csupomona.edu/~jcclark/
Professor, Biological Sciences                   +1 909 979 6371

```