Waterfall Skepticism
Eric Schwitzgebel
Apr. 19, 2013
Yesterday morning
around dawn I sat hypnotized by “Paradise Falls”. I had hiked there from my parents’ house
while my family slept, as I often do when we visit my parents.
Although at first it
didn’t feel that way, I wonder if I have been harmed by philosophy. I gazed at the waterfall thinking about
Boltzmann brains – thinking, that is, about the possibility that I had no past,
or a past radically unlike what I usually suppose it to be, instead having
randomly congealed by freak chance from disorganized matter. On some ways of thinking about cosmology,
there are many more randomly congealed brains, or randomly congealed
brains-plus-local-pieces-of-environment, than there are intelligent beings who
have arisen in what we think of as the normal way, from billions of years of
evolution in a large, stable environment.
If such a cosmology is true, then it might be much more likely that I
have randomly congealed than that I have lived a full forty-five years in human
form. The thought troubled me, but also
the filament of doubt felt comfortable in a way. I enjoy my skeptical reflections.
(Paradise Falls,
Thousand Oaks, California; 6:25 a.m., April 28, 2013)
Of course, most cosmologists flee from the Boltzmann brains
hypothesis. If a cosmology implies that
you are very likely a Boltzmann brain, that’s normally taken to be a reductio ad absurdum of that cosmology. But I wonder if such dismissals arise more
from fear of skepticism than from sound reasoning. I am no expert cosmologist, with a view very
likely to be true about the origin and nature of the universe and thus about
the number of Boltzmann brains vs. evolved consciousnesses in existence – but
neither are any professional cosmologists sufficiently expert to claim secure
knowledge of these matters, the field is so uncertain and changing. As I gazed around Paradise Falls, the
Boltzmann brains hypothesis started to seem impossible to assess. This seemed especially so given the limited
tools at hand – not even an internet connection! – though
I wondered whether having such tools would really help after all.
Still, the world did not dissolve around me, as I suppose it
must dissolve around most spontaneously congealed brains. So as I endured, I came to feel more
justified in my preferred opinion that I am not a Boltzmann brain. However, I also had to admit the possibility
that my seeming to have endured over the sixty seconds of contemplating these
issues was itself the false memory of a just-congealed Boltzmann brain. My skepticism renewed itself, though somehow
this second time only as a shadow, without the force of genuine doubt.
I considered the possibility that I was a computer program
in a simulated environment. If
consciousness can arise in programmed silicon chips, then presumably there’s
something it’s like to be such a computerized consciousness. Maybe such computer consciousnesses sometimes
seem to dwell in natural environments, fed with simulated visual inputs (for
example of waterfalls), simulated tactile inputs (for example of sitting on a
stone), and false memories (for example of having hiked to a waterfall that
morning). If Nick Bostrom
is right, there might be many more such simulated beings than naturally evolved
human beings.[1]
I considered Dan Dennett’s argument against skepticism:
Throw a skeptic a coin, Dennett says, and “in a second or two of hefting,
scratching, ringing, tasting, or just plain looking at how the sun glints on
its surface, the skeptic will consume more information than a Cray
supercomputer can organize in a year.”[2] Our experience, he says, has an informational
wealth that cannot realistically be achieved by computational imitation. In graduate school, I had found this argument
tempting. But it seemed to me yesterday
that my vision of the waterfall was not as high fidelity as that, and could
easily be reproduced on a computer. I
fingered the mud at my feet. The
complexity of tactile sensation did not seem to me the sort of thing beyond the
capacity of a computer artificially to supply, if we suppose a future of
computers advanced enough to host consciousness. We are so eager to reject skepticism that we
satisfy ourselves too quickly with weak arguments against it.
Now maybe John Searle is right and no computer could ever
host consciousness.[3] Or maybe, though computer consciousness is
possible, it is never actually achieved, or achieved only so rarely that the
vast majority of conscious beings are organically evolved entities of the sort
I usually consider myself to be. But I
hardly felt sure of these possibilities.
The philosophers who most prominently acknowledge the
possibility that they are simulated beings instantiated by computer programs
don’t seem very worried by it. They
don’t push it in skeptical directions.
Nick Bostrom seems to think it likely to if we
are in a simulation, it is a large stable one.[4] David Chalmers emphasizes that if we are in a
simulation scenario like that depicted in the movie The Matrix, skepticism needn’t follow.[5] Maybe the easiest and most common way to
create an artificial consciousness is to evolve it up through a billion or a
million years in a stable environment; and maybe the easiest, cheapest way to
create seeming conversation partners is to give those seeming conversation
partners real consciousness themselves, rather than making them ELIZA-like
shells of simple response patterns.[6] But, on the other hand, if I take the
simulation possibility seriously, then I feel compelled to take seriously also
the possibility that my memories are mostly false, that I am instantiated
within a smallish environment of short duration, perhaps inside a child’s
game. I am the citizen to be surprised
when Godzilla stomps through; I am the victim to be rescued by the child’s
swashbuckling hero; I am the hero himself, born new and not yet apprised of my
magic. Nor did I have, at that moment, a
clever conversation partner to convince me of her existence. I might be Adam entirely alone.
Philosophers Fred Dretske and
Alvin Goldman say that as long as my beliefs have been reliably enough caused
by my environment, by virtue of well-functioning perceptual and memory systems,
then I know there’s a real waterfall there, I know that I have hiked the two
kilometers from my parents’ house.[7] But this seems to be a merely conditional
comfort. If my beliefs have been reliably enough caused…. But have they? And I was no longer sure I fully believed, in
any case. What is it, to believe? I still would have bet on the existence of my
parents’ house, but my feeling of doubtless confidence had evaporated. Had everything dissolved around me at that
moment, I would have been surprised but I would not have felt utter shock. I was no longer seamlessly sure that a
familiar world existed beyond the ridge.
I turned to hike back, and as I began to mount the slope, I
considered Descartes’s madman. In his Meditations
on First Philosophy, Descartes seems to say that it would be madness
seriously to consider the possibility that one is merely a madman, like those
who believe they are kings when they are paupers or who believe their heads are
made of glass. But why is it madness to
consider this? Or maybe it is madness,
but then, since I am now in fact considering it, should that count as evidence
that I am mad? Am I a philosopher who
works at U.C. Riverside, whom some readers take seriously, or am I indeed just
a madman lost in weird skepticism, with merely confused and stupid
thoughts? Somehow, this skepticism felt
less pleasantly meditative than my earlier reflections.
I returned home. That
afternoon in philosophical conversation, I told my father that I thought he did
probably exist other than as a figment of my
mind. It seemed the wrong thing to
say. I wanted to jettison my remnants of
skepticism and fully join the human community.
I felt isolated and ridiculous.
I thought about writing up this confession of my thoughts. I thought about whether readers would relate
to it or see me only as possessed by foolish, laughable doubts. The ancient philosopher Sextus
Empiricus was wrong; skepticism does not lead me to equanimity.
[1] Nick Bostrom (2003), “Are
we living in a computer simulation?”, Philosophical Quarterly, 53, 243-255.
[2] Daniel C. Dennett (1991), Consciousness Explained (Boston, MA: Brown, Little & Co.), p.
6.
[3] John R. Searle (1980), “Minds, brain, and programs”, Behavioral and Brain Sciences, 3, 417-457.
[4] See Nick’s and my correspondence about this matter,
posted on The Splintered Mind on August 30 and September 2, 2011: http:// schwitzsplinters.blogspot.com .
[5] David Chalmers (2010), The Character of Consciousness (Oxford: Oxford), ch.
13.
[6] On the program ELIZA, see Joseph Weizenbaum
(1966), “ELIZA – a computer program for the study of natural language
communication between man and machine”, Communications
of the ACM, 9 (1), 36-45.
[7] For an overview of such “reliabilist”
views in epistemology, see Alvin I. Goldman (2008), “Reliabilism”,
Stanford Encyclopedia of Philosophy:
http://plato.stanford.edu .