The Big Picture
Page 63
02
we think about quantum measurement. Penrose has some specific ideas
03
about what those alterations might be— quantum gravity is involved, and
04
filamentary structures in the brain called microtubules— but the upshot is
05
that the wave functions of structures in our brains collapse in just the right
06
way to grant human beings powers of insight and cognition that computers
07
will never achieve.
08
There are a number of objections one could raise, and people have had
09
fun raising them against Penrose for years now. The best ones center on the
10
leap from “Human cognition doesn’t work like a formal mathematical sys-
11
tem” to “The human brain doesn’t obey the known laws of physics.” What
12
we call “thinking” is a way of talking about a very high- level emergent phe-
13
nomenon. It may emerge out of underlying processes that are absolutely
14
rigid and logical, and yet itself not show those characteristics very much at
15
all. Indeed, rigid logic (or even the ability to multiply big numbers accu-
16
rately) is something that human beings are notoriously bad at. Our thoughts
17
leap around, we make mistakes, we have hunches. The fact that we can
18
reach conclusions that wouldn’t be reached by a specific formal system
19
doesn’t seem particularly surprising.
20
Gödel’s Incompleteness Theorem doesn’t quite say there are true state-
21
ments that can’t be proven. Rather, it says that such statements exist for any
22
consistent formal system. How do we know that some particular set of axi-
23
oms defines a consistent system? Or— putting it another way— how can we
24
be sure that we are accurately “perceiving” the truth of Gödel’s self-
25
referential sentences?
26
As Scott Aaronson has pointed out, it’s more accurate to say that we
27
believe certain systems are consistent, though Gödel has shown that we can
28
never prove it. If we allow a computer to assume that the system is consis-
29
tent, it would have no trouble at all proving statements like “This statement
30
cannot be proven.” (Proof: if it could be proven, the system would be incon-
31
sistent!) He quotes Alan Turing: “If we want a machine to be intelligent, it
32
can’t also be infallible. There are theorems that say almost exactly that.”
33
Humans certainly satisfy the criterion of not being infallible.
34
Putting on our Bayesian hats, the fact that the convoluted minds of hu-
35S
man beings naïvely seem to be able to perceive truths that can’t be directly
36N
proven by completely rigorous computer programs doesn’t seem nearly
370
Big Picture - UK final proofs.indd 370
20/07/2016 10:02:55
A R E PhO t O n S C O n S C I Ou S ?
strong enough to warrant modifying our best understanding of quantum
01
mechanics. Especially because the use to which such modifications are be-
02
ing put has nothing directly to do with the mysteries of quantum mechan-
03
ics themselves— it’s just a way to grant powers of insight and cognitive
04
wizardry to the human brain. And at the end of the day, there’s nothing
05
about the brain’s ability to see the truth of unprovable statements that helps
06
us understand the Hard Problem, the issue of inner mental experiences. If
07
you think the Hard Problem is hard, quantum mechanics is unlikely to
08
help you; if you think it’s not so bad, you probably don’t feel the need to
09
change the laws of physics to help us understand the brain.
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
S35
N36
37 1
Big Picture - UK final proofs.indd 371
20/07/2016 10:02:55
01
02
03
43
04
05
What Acts on What?
06
07
08
09
10
11
12
13
14
15
16
17
18
the idea that we are part of the natural world can lead to a sense of
profound loss if the reasons and causes for our actions aren’t what
we thought they were. We’re not human beings, equipped with in-
tentions and goals, so the worry goes; we’re bags of particles mindlessly
19
bumping into one another as time chugs forward. It’s not love that will keep
20
us together, it’s just the laws of physics. A version of this concern was ar-
21
ticulated by philosopher Jerry Fodor:
22
23
If it isn’t literally true that my wanting is causally responsible
24
for my reaching, and my itching is causally responsible for my
25
scratching, and my believing is causally responsible for my say-
26
ing . . . if none of that is literally true, then practically every-
27
thing I believe about anything is false and it’s the end of the
28
world.
29
30
Don’t worry! It’s not the end of the world.
31
We live in a reality that can be fruitfully talked about in many differ-
32
ent ways. We have an extravagant assortment of theories, models, vocabu-
33
laries, stories, whatever you prefer to call them. When we speak about
34
a human being, we can describe them as a person with desires and tenden-
35S
cies and inner mental states; or we can describe them as a collection of
36N
372
Big Picture - UK final proofs.indd 372
20/07/2016 10:02:55
W h At AC t S O n W h A t ?
biological cells interacting via electrochemical signals; or we can describe
01
them as an agglomeration of elementary particles following the rules
02
of the Core Theory. The question is, how do we fit these different stories
03
together? In particular, what acts on what? Does the existence of
04
the particle- physics description, in which “causality” is nowhere to be
05
found, imply that it is illegitimate to talk about scratching being caused by
06
itching?
07
The poetic- naturalist answer is that any of the stories we have stands or
08
falls on its own terms as a description of reality. To evaluate a model of the
09
world, the questions we need to ask include “Is it internally consistent?,” “Is
10
it well- defined?,” and “Does it fit the data?” When we have multiple distinct
11
theories that overlap in some regime, they had better be compatible with
12
one another; otherwise they couldn’t both fit the data at the same time. The
13
theories may involve utterly different kinds of concepts; one may have par-
14
ticles and forces obeying differential equations, and another may have hu-
15
man agents making choices. That’s fine, as long as the predictions of the
16
theories line up in their overlapping domains of applicability. The success
17
of one theory doesn’t mean that another one is wrong; that only happens
18
when a theory turns out to be internally incoherent, or when it does a bad
19
job at describing the observed phenomena.
20
Developing a theory of human thought and behavior in terms of neu-
21
ral signals or interacting particles doesn’t in any way imply that your
22
wanting is not responsible for your reaching. There is no obstacle to
23
that kind of vocabulary of desire and intentionality being “true,” as long
24
as its predictions are compatible with those of other successful vocabularies.
25
It’s possible that what Fodor means by “literally true” is something like
26
“an essential element of every possible description of nature,” or perhaps “of
27
our best and most comprehensive description of nature.” In other words,
28
there can’t exist any successful vocabulary that doesn’t include “wanting”
29
and “believing” as fundamental concepts. In that case, it is not literally
30
true— the physical and biological descriptions of human beings are per-
31
fectly adequate on their own terms, and don’t invoke concepts like wants
32
and beliefs.
33
But that’s an unnecessarily constraining notion of “literally true.” Ther-
34
modynamics and the fluid description of air didn’t stop being true once we
S35
N36
373
Big Picture - UK final proofs.indd 373
20/07/2016 10:02:55
T H E B IG PIC T U R E
01
discovered atoms and molecules. Both ways of talking are true. Likewise,
02
human thoughts and intentions haven’t disappeared just because we obey
03
the laws of physics.
04
•
05
06
This issue seems more complicated than it is because of an understandable
07
tendency, in a world described by multiple distinct but mutually compatible
08
stories, to jumble up the concepts of one story with those of another— to
09
cross the lines separating distinct ways of talking.
10
Rather than acknowledging that there is one way of talking about the
11
world in terms of the quantum fields and interactions of the Core Theory,
12
and another way in terms of electrochemical signals traveling between cells,
13
and yet another way in terms of human agents with desires and mental
14
states, we fall into the trap of using multiple vocabularies at the same time.
15
When told that every mental state corresponds to various physical states of
16
one’s brain, one wants to complain, “Do you really think the reason why I’m
17
scratching is only because of some synaptic signaling, and not because I feel
18
an itch?” The complaint is misplaced. You can describe what’s happening in
19
terms of electrochemical signals in your central nervous system, or in terms
20
of your mental states and the actions they cause you to perform; just don’t
21
trip up by starting a sentence in one language and attempting to finish it in
22
another one.
23
One of the most common arguments against Cartesian dualism (or
24
mental properties that influence physical ones) is causal closure of the physi-
25
cal. The laws of physics as we know them— the Core Theory, in the domain
26
we’re interested in— are complete and self- consistent. You give me a quan-
27
tum state of a system, and there are unambiguous equations that will tell
28
me what it will do next. (We’ve written down one such equation in the
29
Appendix.) There is no ambiguity, no secret fudge factors, no opportunity
30
for differing interpretations of what is happening. If you give me the precise
31
and complete quantum state corresponding to “a person feeling an itch,”
32
and I have the calculational abilities of Laplace’s Demon, I could predict
33
with extraordinary accuracy that the quantum state will evolve into a dif-
34
ferent state corresponding to “a person scratching themselves.” No further
35S
information is needed, or allowed.
36N
374
Big Picture - UK final proofs.indd 374
20/07/2016 10:02:55
W h At AC t S O n W h A t ?
•
01
02
In chapter 13 we discussed the idea of “strong emergence,” according to
03
which the behavior of a system with many parts is not reducible to the ag-
04
gregate behavior of all those parts. A related idea is downward causation:
05
behavior of the parts is actually caused by the state of the whole, in a way
06
not interpretable as due to the parts themselves.
07
Poetic naturalists tend to view downward causation as a deeply mis-
08
guided idea. Then again, they view upward causation as equally misguided.
09
“Causation,” which after all is itself a derived notion rather than a funda-
10
mental one, is best thought of as acting within individual theories that rely
11
on the concept. Thinking of behavior in one theory as causing behavior in
12
a completely different theory is the first step toward a morass of confusion
13
from which it is difficult to extract ourselves
.
14
It’s certainly possible that behavior in coarse- grained macroscopic theo-
15
ries might be entailed by features of more comprehensive theories, and we
16
certainly want them to be consistent with such theories when the descrip-
17
tions overlap. We might even, as long as we’re careful, say that features of an
18
underlying theory can help explain features of an emergent one. But we get
19
in trouble if we try to say that phenomena in one theory are caused by phe-
20
nomena in a different one. I know that I cannot use my mental powers to
21
reach across space and bend spoons, since the fields and interactions of the
22
Core Theory don’t accommodate that kind of capacity. But I can describe
23
that feature purely in the macroscopic language: human beings don’t pos-
24
sess the power of telekinesis. The microscopic explanation might aid my
25
understanding, but it’s not a necessary part of how I talk about human- scale
26
behavior.
27
And the converse, downward causation of human- scale properties influ-
28
encing the microscopic behavior of particles, is misguided. A standard ex-
29
ample is the formation of snowflakes. Snowflakes are made of water
30
molecules, interacting with other molecules to form a crystalline structure.
31
There are many possible structures, determined by the initial configuration
32
of the seed from which the snowflake grows. Therefore, it is claimed, the
33
macroscopic shape of the snowflake is acting “downward” to determine the
34
precise location of individual water molecules.
S35
N36
375
Big Picture - UK final proofs.indd 375
20/07/2016 10:02:55
T H E B IG PIC T U R E
01
It’s bad form to mix vocabularies in such a vulgar way. Water molecules
02
interact with other water molecules, and other molecules in the air, in pre-
03
cise ways that are specified by the rules of atomic physics. Those rules are
04
unambiguous: you tell me what other molecules any individual water mol-
05
ecule is interacting with, and the rules will say precisely what will happen
06
next. The relevant molecules may be part of a larger crystalline structure,
07
but that knowledge is of zero import when studying the behavior of the
08
water molecule under consideration. The environment in which the mole-
09
cule is embedded is relevant, but there is no obstacle to describing that en-
10