On Sep 19, 1:32 am, Ray Martinez <
pyramid...@yahoo.com> wrote:
> On Sep 18, 2:27 pm, Greg Guarino <
gdguar...@gmail.com> wrote:
>
>
>
>
>
>
>
>
>
> > On 9/18/2012 4:47 PM, pnyikos wrote:
>
> > >> >Natural selection*is falsifiable* and*testable*. Witness the
> > >> >role of counter-intuitive examples like the elaboration of
> > >> >sexual selection traits like male guppy tails. The less predation,
> > >> >the more elaborate the tails of future generations of males
> > >> >become (tail size and color is highly hereditary); the more predation,
> > >> >the duller and smaller the tails of future generations become.
> > > And if it had been the other way around, what then? Perhaps the
> > > environment could furnish clues; long tails might blend in with
> > > vegetation with long stalks, etc. and there is no falsification of
> > > natural selection, but an "additonal confirmation."
>
> > You confuse the potential falsification of the *concept* of Natural
> > Selection as an agent of biological change with the falsification of a
> > particular hypothetical selective factor as the agent of change in a
> > specific circumstance. If we hypothesize that long tails prevent
> > predation by being good camouflage, but find that in the presence of
> > predators tails shrink, we have falsified our hypothesis about tails,
> > but not the general concept of Natural Selection.
>
> What falsifies "the concept of Natural Selection as an agent of
> biological change"?
>
Lots of pretty good examples have been given by me and others
previously to that question
One of my favourites: If we observed that species, just like the
individuals that make them up, come into existence, grow and then die
(out) according t more or less fixed periods of time, independent from
the environment. E.g. if we observed the pattern that species always
only exists for 100 times the length of the average lifespan of its
members, then NS would be falsified. The idea that species have such
an inbuild "life span" was taken serious as a hypothesis at one point
in time.
Observing that all organisms have the same number of offspring would
be another falsification, finding that all traits are neutral and
their pattern corresponds to that of shift only another, or that all
variation of traits are not inheritable but purely epigenetic.
> > > The popular literature is full of alleged counterexamples to natural
> > > selection that get shot down with more sophisticated arguments.
>
> > > In other words, the alleged falsifications themselves get falsified.
>
> > >> >All due to the relative reproductive success of males in the
> > >> >face of differing levels of predation.
> > > This says nothing about how it could possibly be falsified.
>
> > > Nor have I seen anything on this thread that says how it could be
> > > falsified.
>
> > >> >Natural selection has not*been* falsified despite*being*
> > >> >falsifiable.
> > > So, let's see a hypothetical observation that you would consider as
> > > falsifying it.
>
> > If a trait were to become more prevalent (presumably by some heretofore
> > unknown mechanism) in a population despite the lower reproductive
> > success of those individuals that possess that trait, Natural Selection
> > as an agent of change in that population would be falsified. Given what
> > we know of heredity, that seems unlikely, but "unlikely" is not the same
> > as "unfalsifiable in principle".
>
> I think Greg is actually making an argument, without any awareness,
> that natural selection is NOT falsifiable. He has isolated specific
> instances as not harming conceptual existence as an agent causing
> change. What his argument accomplishes is reduction of change caused
> by natural selection. The concept remains intact, explaining a little
> less.
>
OK, if we try to translate this into English, you do have something
like a point: What you are saying essentially is: whenever we make an
observation that contradicts a theory, we can argue that there is
something specific to _this_ observation that creates an exception.
This way, we can reformulate the original theory, keeping most of its
concepts, and making it just a bit more complex, and less valuable for
making predictions.
That is absolutely true - but not specific to NS, or even biology, but
to all theories, scientific or otherwise. That a theory is falsifiable
(a structural property of the theory) does not tell us what to do
when a theory is actually falsified. Popper himself was far from clear
about that distinction, and it is one of the reasons why few people
these days are falsificationists in the Popperian sense.
Generally, belief revision is conservative - when we get new data, we
try to change as little of our beliefs as possible, and only as much
as is necessary. That is a general feature of belief revision and
applies to all our theories, scientific and otherwise. When we
discovered e.g. tha objects on the subatomic level did not obey the
Newtonian laws of particle mechanics, we did not throw out all the
Newtonian concepts, we simply restricted them to the larger objects
for which they still work perfectly well, and deal with the others
through quantum mechanics.
The same we do in everyday life. I belief for instance the theory that
the times on the bus timetable tell me when buses will arrive at my
stop. Now, the timetable says there ought to be a bus at 10.15, it is
already 10.17 and no bus in sight. At this point, the theory:
"timetables times correspond to real times" is falsified. What is the
rational reaction to this? Should I give up the belief that buses and
timetables exist at all? Or should I give up the idea that timetable
times and real times ever correspond, and that i can use the timetable
to predict at all when a bus will be coming? Or should I make a much
more conservative adjustment to my beliefs: timetable times tell me
_within a margin or error_ when buss will come? (that is, more of less
accurately)? Or maybe: Sometimes a driver gets ill, or a bus breaks
down, so in most cases, but not always, do timetables tell me when a
bus will come?
I'd say rational people will opt for one of the last two revisions in
this case. But the doctrine of falsification does not tell me what to
do, which parts of the theory I should revise and which ones to keep.
Now, if I get more data than a single observation, and find that in
Edinburgh, buses are as likely to show up when the timetable says they
should as not, I eventually give up the timetable as useless. But what
then if i travel to Glasgow? Should I trust the timetable there? Well,
I'd say that would be perfectly reasonable. in this case, i revised my
original theory of: "Timetables predict the times when buses come"
to "Timetables predict the times when buses come, but not in
Edinburgh where the bus operator is incompetent".
That illustrates a more interesting philosophical point: the
distinction between observations and theories is simply not as clear
cut as the Popperian model claims. Take again the example of me
waiting for a bus at the busstop. No bus shows up. What _exactly_
have i observed, what does it falsify? Have I observed
- no buses show up
- no buses show up at 10.15
- no buses show up in the morning
- no buses show up in Craighouse Gardens Busstop
- no buses show up in Edinburgh
- no buses show up in Craighouse gardens at 10.15
etc etc etc
all these statements are valid and mutually consistent descriptions
of my observation. None of them is "more true" than the other, just
more or less specific. However, the more specific the description of
my observation, the less impact it will have on any theory they
falsify. Narrowly described obs4rvations falsify only very specific
theories, widely described observations can falsify much more general
theories. Which description is the best to chose in a given situation
is not a question of what i observe, bit of all the other theories I
have, and my research question. If I have good reasons to belief e.g.
that bus timetables in Edinburgh are by and large correct, but suspect
that those in Morningside do not, because the busoperator dislikes
people in Morningside, the appropriate formulation - for that
specific situation and research question, would be: "no buses show up
in Craighouse Gardens Busstop " That is as specific as it needs be,
and as wide as it can be.
General lesson: how we describe an observation depends on our theories
and on our research question, there are no "pure" observation
statements independent from interpretation and theories.
So, the examples of falsification for natural selection that I gave
above are "as good an example" of the falsification of a theory, any
theory, as you can get. The concept is just as falsifiable as those
in any other theory, not more so, but also not less.
> Ray