SETI

"Brian Tung" wrote in message
...
Zane wrote:
BUT, we can make the axiomatic statement that if there
are enough opportunities, anything not prohibited by the laws of
physics will certainly happen. This is so basic a concept that most
people, at least, don't think that it has a burden of proof.

Davoud wrote:
By definition, an axiom has no burden of proof; its truth is taken to
be self evident. Thus, what you have asserted is not an axiom. The
"will certainly happen" is the problematic part. _May_ happen works,
though.

The problem is that one can show by quite simple and rigorous math that
the probability of occurrence must be either zero or one, provided the
trials are independent and identically distributed (i.i.d.). As a simple
example, if you have a chance at all of winning the lottery (that is, you
play and the game isn't rigged to prevent you from winning), then given
an infinite number of trials, you are bound to win. What's more, you
are bound to win an infinite number of times. (However, that doesn't
mean your expected gain is infinitely positive. On the contrary, it's
infinitely negative. But I digress.)

So in order to conclude that "may" is preferable to "will," one must
show that the trials (of which the Earth is the only one known to yield
a successful result) are either not independent or not identicallly
distributed.

Dependence is difficult to argue. How could we prevent the development
of life even several light-years away? Affect it once it had produced
a technical civilization, yes; cause it to "commit suicide," even--not
likely, but just possible. But prevent it altogether? Hard to see how
that could happen.

Besides, as far as we know, our influence can only spread out at the
speed of light--no faster. The Earth has harbored life for no more than
about 4 billion years, and our effects could only extend for 4 billion
light-years in any direction. Beyond that is a vast arena for life to
develop, and if the question is predicated on an arbitrarily large
number of trials, then a 4-billion-light-year-radius sphere is, quite
literally, nothing. (Although we'll have to be listening for a very
long time to pick up anything more distant!)

All right, then, if dependence won't invalidate the principle, we can
examine whether or not the trials are really identically distributed.
Right away, we can say that they clearly are not. Mercury, for example,
is one such trial, and the odds for life developing there (at least in
its current state) are clearly not very exciting.

However, we know this because we know something a priori about Mercury.
In particular, we know that Mercury is very close to the Sun, and that
its rotation period is very long, so that the surface temperature
undergoes wild swings. (Nor is there, contrary to the hopes of 1950s
science fiction stories, a "twilight zone," because the tidal lock
acquired by the Sun is not 1:1, but 3:2.) It is not a planet hospitable
to life.

But if we knew nothing about Mercury other than that it was a planet
orbiting a Sun-like star, we would make different estimates altogether
about the planet's opportunities for developing life--in fact, we would
make it the same as for any other such planet. In other words, in the
absence of a priori knowledge about other planets, the trials *do*
become essentially identically distributed. To say that they are not
would require us to find out that (for instance) Earth-like planets
around Sun-like stars can only develop within a finite space in the
universe. That goes against the Copernican principle, and so it does
not seem to me plausible at all. It's true that the Copernican
principle is only a principle and not a hard and fast law, but just the
same, we're really arguing philosophy (since the amount of space we can
*listen* to is finite, not infinite), so I think it's perfectly
applicable.

In short, it seems hedging to me to say that given an essentially
infinite number of trials, we can only say that life "might" develop.
It *must* develop, absent evidence of dependence or a non-uniform
distribution, because it *has* happened, demonstrating that the
probability per trial is non-zero. Not only that, but it must develop
any number of times, no matter how unlikely it might be in any
individual case.

Brian Tung
The Astronomy Corner at http://astro.isi.edu/
Unofficial C5+ Home Page at http://astro.isi.edu/c5plus/
The PleiadAtlas Home Page at http://astro.isi.edu/pleiadatlas/
My Own Personal FAQ (SAA) at http://astro.isi.edu/reference/faq.txt

In an infinite universe, everything is not only possible but inevitable.

Tony Turner wrote:
In an infinite universe, everything is not only possible but inevitable.

Not quite. There is also the impossible, and that which is possible but
can happen only a finite number of times.

Brian Tung
The Astronomy Corner at http://astro.isi.edu/
Unofficial C5+ Home Page at http://astro.isi.edu/c5plus/
The PleiadAtlas Home Page at http://astro.isi.edu/pleiadatlas/
My Own Personal FAQ (SAA) at http://astro.isi.edu/reference/faq.txt

"Andrew Walker" wrote:

I'd call it useful science if they were doing more with the data,
like looking for unusual radio emitting stars or galaxies. As far
as I can tell if is no detection, which is extremely likely,
they'll have nothing to show for it, ie a waste of time.
Start searching for interesting astronomical objects and it's a
different story.

Great news x2! SETI@home data is currently being used in a
hydrogen survey of our Galaxy:

http://www.planetary.org/html/UPDATE...te_012403.html

SETI@home II will use the BOINC platform, which will enable users
to choose between many projects (e.g. Folding@home) using a
unified client. Besides SETI@home II, one of the first projects will
be Astropulse:

http://boinc.berkeley.edu/