Species as "Hypotheses"

Nico Mario Franz nmf2 at CORNELL.EDU
Thu Jul 15 01:57:33 CDT 2004

Hi Rich:

   this is why it's fun arguing with you on the list. Just the right mix
of shooting from the hip vs. backing it up. Thanks! Hope I can

NF: > Same is true for "a hypothesis must be falsifiable to be >

RP: > If it's not falsifiable, then why call it an hypothesis? Are there
kinds of scientific hypotheses that are not falsifiable? If so, shouldn't
some other word be used to refer to them, other than "hypothesis"?

NF (new): So far as I've learned, falsifiability is of course desirable.
There's a strong notion of it, perhaps applicable in mathematics and logic
(in the philosopher's sense of the subject area of deductive Logic); and a
more realistic version, applicable in our sciences - let's call it weak
falsification. "Hypothesis A should be considered falsified for now, given
the evidence I claim to have and claim is relevant, until the next round
of falsifying."

That's really induction. See Putnam, H. (1974). The 'corroboration' of
theories. In The Philosophy of Karl Popper, Book 1 (ed. P. A. Schilpp),
pp. 221-240. Open Court Publishing Company, La Salle.

When you use the term "falsifying", it sounds to me like one of those more
or less subtle attempts to lend more authority to an action (such as
naming, describing, and type-designating a species) than it needs in order
to stand out as good science. Those who use words like "logic",
"falsifying", "objective" too much, suffer (at least in my view) from a
kind of strange physics envy that won't help taxonomy in the long run.
Don't use high-flying philosophical terms to make your science look like
something it's not. That's an ideal which I'm personally trying to attain.

So: yes, weak falsification is nice, testability is perhaps even more
generally desirable. But hypotheses can be *potentially* testable, just
not now given our instruments, or our current theoretical limitations.
Does that mean they're non-scientific? Apparently no. Putnam reviews I
believe a major theory by Newton (it's in there, I won't say anything more
[embarrassing]). The theory wasn't even testable in a stricter sense for
about 100 years. Yet it swept through the entire field of physics, because
- as Putnam thinks - of its considerable explanatory successes, or
inductive reliability. Few cared whether it could be strictly falsified.

All of the above just to say: no, the falsifiable criterion doesn't do
good demarcation all the time. And no, falsification isn't any more
objective or powerful than induction. We can still do good work without
those two bullies in our corner.

I can't reply to everything, but one of the things that strike me about
your position is the following tension: (1) a stated desire to attain
objectivity, mixed with the idea that (e.g.) species are merely
"constructs." I don't think that mix could ever be worked into a
consistent position, but don't have the capacity to lay out why. If I were
to put the positions into two schools, one would be logical empiricism,
and the other constructivism. I guess my point would be you might not
realize how very (very) different bits and pieces of "lessons from
philosophy" alternately affect how you present your view, and that it
might take some work to sort it all out.

RP: > When I was discussing the merits of a multi-anchored taxonomic units
(analogous to Phylocode taxa), what I meant was that the biological scope
such names have objective boundaries, and while they do not themselves
constitute testable hypotheses, they do provide an unambiguously objective
method for testing whether any particular organism belongs to the taxon or
not.  This is not true of Linnaean taxa, [...]

NF (new): I disagree, compassionately and respectfully. First, and again,
the use of the term "objective" here is unfortunate. Taxonomy is a
language, to be developed, used, and understood by people.
Inter-subjective. If we could only agree on this. I'm quite sure that as
soon as you can say whether or not a certain "object in the world" is
subsumed under the multi-anchored definition - without ALSO relying on
taxonomic interpretations about the correctness of the use of names,
assignments of specimens to taxa, all sorts of inferences about how the
classification came about (and no, it's not even necessary to talk about
properties yet) - THEN you're not doing empirical work. That's a price for
"objectivity" that luckily under Linneus we've never had to pay. I'm
always amazed when I listen to serious taxonomists buying into the
lack-of-respect, quick-fix attitude of the Phylocode. What "phylogenetic
definitions" do (to a large degree) is disconnect nomenclature from
evidence. Full stop.

RP: > And we now return to this idea of "natural", that we (over)discussed
on this list recently. It's my turn to ask you to be more precise in your
meaning of the word of "natural" here. Do you mean to suggest that one of
the three definitions for "Centropyge fisheri" (sensu stricto, sensu lato,
and sensu ultra-lato) is intrinsically more "natural" than the other two,
and that the accumulation of evidence will eventually point us to the
"most natural" grouping? Or, are all three groupings equally "natural"
(assuming that all three are monophyletic), and the decision of where to
draw the "species line" is a purely subjective one? If the latter, then I
would describe the three notions of C. fisheri as three different
"definitions" of the species, none of which is testable or falsifiable
(except for monophyly -- which is not what I was talking about); and
therefore none of which could be thought of as an "hypothesis" (by any
definition of that term that I am familiar with).

NF: When I say "natural" this reflects my personally favored flavor of
thinking, according to which there IS a natural order among living things
out there in nature. We may sometimes have successes and other times
failures in our attempts to approximate that order with our taxonomy. How
do we know? Certainly not just because any definition or concept says so.
But we tend to think that (and rightly so I feel), when species
"hypotheses" are tested over and over against variously overlapping sorts
of (so-perceived) evidence), and thereby inter-subjectively confirmed, and
then used to do ecology, evolutionary biology, biocontrol, environmental
politics, etc. - all apparently successfully (stated predictions hold true
enough); it means the hypotheses we're "on to" something. Empirical
adequacy, approximate truth, operational functionality. Airplanes fly so
maybe our laws of physics aren't so bad. Whales are mammals so maybe our
notion of homology isn't so bad. I don't sense that that's a radical
analogy, though you may think it is.

Can I export my species hypothesis into other closely related yet also
somewhat independent fields, and then use it fruitfully to find out yet
other things? If so, then I'd say the hypothesis may be a natural one.
Looks like the jury is still out in your example. (BTW this makes me an
adherent of Wiley's of Wheeler and Platnick's species concept, the latter
only if "character" is turned into a pretty complex notion).

With a bit of good will and work, you can use this approach and also
account for cases where humans "wrongly" uphold hypotheses for much longer
than they should have. Going all the way to "species are merely
constructs" is a less attractive position to me.



More information about the Taxacom mailing list