The Big Picture

Home > Other > The Big Picture > Page 14
The Big Picture Page 14

by Carroll, Sean M.


  h r

  e p o

  r p

  o o

  p s

  o istiton

  o s

  n . S

  s y

  . S m

  y b

  m o

  b l

  o ilcal

  ca ly,y

  18

  19

  Cre

  Cr d

  e e

  d n

  e c

  n e i

  c n p

  e i

  r

  n p o

  r p

  o os

  p i

  os tiio

  t n

  io X

  n

  Lik

  L e

  ik l

  e iho

  l

  o

  iho d

  o o

  d f

  o ob

  f

  s

  ob e

  s r

  e v

  r ati

  v on

  ati

  on D

  Pr

  P io

  r r cr

  io

  e

  r cr d

  e e

  d n

  e c

  n e

  c

  gi

  g v

  i e

  v n o

  e

  bs

  n o e

  bs r

  e v

  r a

  v t

  a io

  t n

  io D

  n

  20

  

  20

  

  gi

  g v

  i e

  v n p

  e

  r

  n p o

  r p

  o os

  p i

  os tiio

  t n

  io X

  

  

  

  n X

  

  

  

   in p

  i

  r

  n p o

  r p

  o os

  p i

  os tiio

  t n

  io X

  n 

  21

  22

  Tha

  Th t

  a ’ts B

  s a

  B y

  a e

  y s

  e ’ss The

  s Th o

  e r

  o e

  r m i

  em n a n

  in a u

  n t

  u sh

  s e

  h l

  e l. The s

  e symbol “

  l “ ” m

  ” e

  m a

  e n

  a s “

  ns i“s p

  is r

  p o

  r p

  o o

  p r

  o -r‑

  23

  titon

  o a

  n l

  a to

  l t .o” It

  ” I ’ts ju

  s j s

  u t a r

  s

  e

  t a r m

  e i

  m n

  i d

  n e

  d r t

  e h

  r t a

  h t w

  a e s

  t w h

  e s o

  h u

  o l

  u d ma

  d m k

  a e s

  k u

  e s r

  u e a

  r l

  e a l o

  l f o

  l o u

  f o r c

  u r

  r c e

  r d

  e e

  d n

  e c

  n e

  c s

  e

  24

  ad

  a d u

  d p t

  d u o 1 a

  p t

  t t

  o 1 a h

  t t e e

  h n

  e e d o

  n f t

  d o h

  f t e d

  h a

  e d y

  a .y

  25

  •

  [

  26

  27

  It f

  I e

  t f e

  e les n

  l a

  s n t

  a u

  t r

  u a

  r l t

  a o a

  l t s

  o a ssisgin n

  g u

  n n m

  u e

  m r

  e ircal c

  ca r

  l c e

  r d

  e e

  d n

  e c

  n e

  c s i

  e n c

  s i e

  n c r

  e tra

  t ian ca

  i

  s

  n ca e

  s s

  e , l

  s i

  , l k

  i e p

  k o

  e p k

  o e

  k r

  e r

  28

  ha

  h n

  a d

  n s

  d or fl

  s o i

  r fl ps

  p of a

  s o co

  f a c i

  o n

  i ,

  n wh

  , w e

  h r

  e e w

  r

  e

  e w ca

  e c n s

  a i

  n s m

  i p

  m l

  p y co

  y c u

  o n

  u t

  n al

  t a l t

  l h

  l t e p

  h o

  e p s

  o ssisb

  i i

  b liiltites

  e .s

  29

  We

  W ’ere a

  re l

  a slo

  s fa

  o f m

  a i

  m liilair

  a wi

  r w th

  t us

  h u isn

  i g

  n g pr

  p o

  r b

  o a

  b b

  a i

  b liilty

  it -

  y t‑ a

  t lak w

  lk h

  w e

  h n r

  e e

  n r f

  e e

  f r

  e r

  r irn

  i g

  n to f

  g t u

  o f t

  u u

  t r

  u e

  r e

  30

  ev

  e e

  v n

  e t

  n st: “

  s Th

  : “ e

  Th r

  e e i

  r s l

  e i e

  s l s

  e s t

  s h

  s t a

  h n a 1 p

  a

  e

  n a 1 p r

  e c

  r e

  c n

  e t ch

  n

  a

  t ch n

  a c

  n e t

  c h

  e t a

  h t t

  a h

  t t
e o

  h n

  e o c

  n o

  c m

  o i

  m n

  i g a

  n s

  g a tse

  t r

  e o

  r i

  o d

  d

  31

  wi

  w lil i

  l m

  l i p

  m a

  p c

  a t t

  c h

  t t e E

  h a

  e E r

  a trh a

  t n

  h a d ca

  n

  u

  d ca s

  u e a m

  s

  a

  e a m s

  a s e

  s x

  s e t

  x itn

  i c

  n t

  c iton

  o .

  n”

  32

  The B

  Th a

  e B y

  a e

  y s

  e isain a

  a p

  n a p

  p r

  p o

  r a

  o ch i

  a

  s m

  ch i

  u

  s m ch m

  u

  o

  ch m r

  o e g

  r e

  e g n

  e e

  n r

  e a

  r l t

  a h

  l t a

  h n t

  a h

  n t i

  h si, h

  s o

  , h w

  o e

  w v

  e e

  v r

  e . I

  r t

  . I t

  33

  re

  r m

  e i

  m n

  i d

  n s u

  d s t

  s u h

  s t a

  h t w

  a e a

  t w s

  e a ssisgin p

  g r

  n p iror c

  o r

  r c e

  r d

  e e

  d n

  e c

  n e

  c s

  e , a

  s n

  , a d u

  n p

  d u d

  p a

  d t

  a e t

  t h

  e t e

  h m a

  e

  p

  m a p

  p r

  p o

  r p

  o r

  p irait

  a e

  t ley,

  y

  34

  to

  to ev

  e e

  v rey f

  r a

  y f c

  a tcu

  t a

  u l p

  a r

  l p o

  r p

  o o

  p soistiitoin t

  o h

  n t a

  h t m

  a a

  t m y o

  a r m

  y o a

  r m y n

  a o

  y n t b

  o e t

  t b r

  e t u

  r e a

  u b

  e a o

  b ut t

  o h

  ut t e w

  h or

  e wo lrd

  l .

  d

  35S

  Do

  D e

  o s G

  e

  o

  s G d e

  o x

  d e i

  x sits? C

  t

  a

  ? C n o

  a

  u

  n o r i

  u n

  r i n

  n e

  n r c

  e o

  r c n

  o s

  n c

  s ico

  i u

  o s e

  u x

  s e p

  x e

  p r

  e irein

  e c

  n e

  c s b

  e e e

  s b x

  e e p

  x l

  p alian

  i e

  n d i

  e n

  d i

  n

  36N

  78

  Big Picture - UK final proofs.indd 78

  20/07/2016 10:02:40

  9780525954828_BigPicture_i-x_1-470_4P.indd 78

  3/30/16 10:10 PM

  u PdA t I n g Ou R K nOW l E d g E

  purely physical terms? Are there objective standards of right and wrong?

  01

  All of the possible answers to such questions are propositions for which

  02

  each of us has a prior credence (whether we admit it or not), and which we

  03

  update when relevant new information comes in (whether we do so cor-

  04

  rectly or not).

  05

  Bayes’s Theorem allows us to be quantitative about our degrees of belief,

  06

  but it also helps us keep in mind how belief works at all. Thinking about

  07

  credences in this way provides a number of useful lessons.

  08

  Prior beliefs matter. When we’re trying to understand what is true

  09

  about the world, everyone enters the game with some initial feeling about

  10

  what propositions are plausible, and what ones seem relatively unlikely. This

  11

  isn’t an annoying mistake that we should work to correct; it’s an absolutely

  12

  necessary part of reasoning in conditions of incomplete information. And

  13

  when it comes to understanding the fundamental architecture of reality,

  14

  none of us has complete information.

  15

  Prior credences are a starting point for further analysis, and it’s hard to

  16

  say that any particular priors are “correct” or “incorrect.” There are, needless

  17

  to say, some useful rules of thumb. Perhaps the most obvious is that simple

  18

  theories should be given larger priors than complicated ones. That doesn’t

  19

  mean that simpler theories are always correct; but if a simple theory is

  20

  wrong, we will learn that by collecting data. As Albert Einstein put it: “The

  21

  supreme goal of all theory is to make the irreducible basic elements as sim-

  22

  ple and as few as possible without having to surrender the adequate repre-

  23

  sentation of a single datum of experience.”

  24

  Simplicity is sometimes easy to gauge; sometimes it is less so. Consider

  25

  three competing theories. One says that the motion of planets and moons

  26

  in the solar system is governed, at least to a pretty good approximation, by

  27

  Isaac Newton’s theories of gravity and motion. Another says that Newto-

  28

  nian physics doesn’t apply at all, and that instead every celestial body has an

  29

  angel assigned to it, and these angels guide the planets and moons in their

  30

  motions through space, along paths that just coincidentally match those

  31


  that Newton would have predicted.

  32

  Most of us would probably think that the first theory is simpler than the

  33

  second— you get the same predictions out, without needing to invoke

  34

  vaguely defined angelic entities. But the third theory is that Newtonian

  S35

  N36

  79

  Big Picture - UK final proofs.indd 79

  20/07/2016 10:02:41

  T H E B IG PIC T U R E

  01

  gravity is responsible for the motions of everything in the solar system ex-

  02

  cept for the moon, which is guided by an angel, and that angel simply

  03

  chooses to follow the trajectory that would have been predicted by Newton.

  04

  It is fairly uncontroversial to say that, whatever your opinion about the first

  05

  two theories, the third theory is certainly less simple than either of them. It

  06

  involves all of the machinery of both, without any discernible difference in

  07

  empirical predictions. We are therefore justified in assigning it a very low

  08

  prior credence. (This example seems frivolous, but analogous moves become

  09

  common when we start talking about the progress of biological evolution

  10

  or the nature of consciousness.)

  11

  Some people don’t like the Bayesian emphasis on priors, because they

  12

  seem subjective rather than objective. And that’s right— they are. It can’t

  13

  be helped; we have to start somewhere. On the other hand, ideally the

  14

  likelihoods of making certain observations can be objectively deter-

  15

  mined. If you have a certain theory about the world, and that theory is

  16

  precise and well-defined, you can say with confidence what the chances are

  17

  of observing various bits of data under the assumption that your theory is

  18

  correct. In realistic circumstances, of course, we are often stuck trying to

  19

  evaluate theories that aren’t so rigorously defined in the first place. (“Con-

  20

  sciousness transcends the physical” is a legitimate proposition, but it’s not

  21

  sufficiently precise to make quantitative predictions.) Nevertheless, it’s our

  22

  job to try to make our propositions as well-defined as possible, to the point

  23

  where we can use them to objectively establish the likelihoods of different

  24

  observations.

  25

  Everyone’s entitled to their own priors, but not to their own likelihoods.

  26

  Evidence should move us toward consensus. You might worry that

  27

  having subjective priors could make it hard for some people to ever reach

  28

  agreement. If I assign a prior credence of 0.000001 to an idea like “God

  29

  created the universe,” and you assign a prior credence of 0.999999 to the

  30

  same proposition, it would require some serious updating on the basis of

  31

  observations before one of us changed our view.

  32

  In practice, that’s a real problem. People have certain views that they’re

  33

  just never going to change, which in Bayesian language corresponds to pri-

  34

  ors set to 0 or 1. That’s too bad, and something we need to learn to deal with

  35S

  in the real world.

  36N

  80

  Big Picture - UK final proofs.indd 80

  20/07/2016 10:02:41

  u PdA t I n g Ou R K nOW l E d g E

  But in principle, if we are all trying to be fair and open- minded

  01

  and willing to change our beliefs in the face of new information, evidence

  02

  will win out in the end. You can assign a very high prior credence to some

 

‹ Prev