Charity - Washington and Lee Universityhome.wlu.edu/~goldbergn/scholarship/Articles/Principle of...
Transcript of Charity - Washington and Lee Universityhome.wlu.edu/~goldbergn/scholarship/Articles/Principle of...
The
Pri
ncip
le o
f C
hari
ty�
NA
TH
AN
IEL
GO
LD
BE
RG
. M
ount
Sai
nt M
ary'
s U
nive
rsity
AB
ST
RA
CT
: T
he r
ecen
t pub
lica
tion
ofa
thi
rd a
ntho
logy
of D
onal
d D
avid
son'
s ar
ti
cles
, a
nd
ant
icip
ated
pub
lica
tion
of
two
mor
e, e
ncou
rage
s a
cons
ider
atio
n o
f th
emes
bi
ndin
g to
geth
er D
avid
son'
s li
feti
me
of
rese
arch
. O
ne s
uch
them
e is
the
pri
ncip
le o
f ch
arit
y (P
C).
In
ligh
t o
f th
e m
ilea
ge D
avid
son
gets
out
of
PC
, I
prop
ose
a ca
refu
l ex
amin
atio
no
fPC
itse
lf In
Par
t 1, I
con
side
r so
me
way
s in
whi
ch D
avid
son
arti
cula
tes
PC
. In
Par
t 2,
I sh
ow t
hat
the
arti
cula
tion
tha
t D
avid
son
requ
ires
in h
is w
ork
on e
pis
tem
olog
y is
unt
enab
le g
iven
wha
t D
avid
son
says
in h
is w
ork
on s
eman
tics
. I
conc
lude
th
at D
avid
son
can
use
PC
onl
y in
his
wor
k on
sem
anti
cs o
r no
t at
all.
RE
SU
ME
: L
a p
arut
ion
rece
nte
du t
rois
iem
e re
cuei
l d'a
rtic
les
de D
on
ald
Dav
idso
n,
lequ
el d
evra
it e
tre
suiv
i de
deu
x au
tres
, in
cite
a e
xam
iner
les
the
mes
qui
tra
vers
ent
tous
ses
trav
aux.
Par
mi c
es t
hem
es s
e tr
ouve
Ie p
rinc
ipe
de c
hari
te (
PC
) .C
on
sid
era
nt
tout
Iep
art
ique
Dav
idso
na
tire
du P
C, j
em
epr
opos
ed'
enfa
ire
un e
xam
enat
tent
if.
Dan
s la
pre
mie
re p
arti
e, j
'exa
min
e di
vers
es fo
rmul
atio
ns d
u P
C p
ar D
avid
son.
Dan
s la
sec
onde
part
ie,j
em
ontr
e qu
e la
form
ulat
ion
qu'e
xige
ntse
s tr
avau
x d
'epi
stem
olog
ie
esti
nte
nabl
eet
antd
onne
cequ
'lle
ndi
tdan
sse
str
avau
xde
sem
anti
que.
De
la,j
eco
ncl
us q
ue D
avid
son
ne p
eut s
e se
rvir
du
PC
que
dan
s se
s tr
avau
x de
sem
anti
que
ou p
as
du t
out.
The
rece
nt p
ubli
cati
on o
f a th
ird
anth
olog
y o
f Don
ald
Dav
idso
n's
arti
cles
, an
d an
tici
pate
d pu
blic
atio
n o
f tw
o m
ore,
1 en
cour
ages
a c
onsi
dera
tion
of
them
es b
indi
ng to
geth
er D
avid
son'
s li
feti
me
of
rese
arch
. O
ne s
uch
them
e is
the
pri
ncip
le o
f ch
arit
y (P
C).
In
ligh
t of
the
mil
eage
Dav
idso
n ge
ts o
ut
of
PC, I
pro
pose
a c
aref
ul e
xam
inat
ion
of
PC
itse
lf.
Hen
ce, t
houg
h R
am
berg
is r
ight
that
"pr
ecis
e ar
ticu
lati
on o
f th
e pr
inci
ple
of c
hari
ty tu
rns
ou
t to
be
an e
xtre
mel
y tr
icky
task
" (1
989,
p. 7
0),2
in P
art 1
of
this
art
icle
I co
n-
Dia
logu
e X
LII
I (2
004)
, 671
-83
© 2
004
Can
adia
n P
hilo
soph
ical
Ass
ocia
tion
/Ass
ocia
tion
can
adie
nne
de p
hilo
soph
ie
672
Dia
logu
e
side
r so
me
way
s in
whi
ch D
avid
son
does
art
icul
ate
pc.
In P
art
2 I
show
th
at th
e ar
ticu
lati
on t
hat
Dav
idso
n re
quir
es i
n hi
s w
ork
on e
pist
emol
ogy
is u
nten
able
giv
en w
hat
Dav
idso
n sa
ys i
n hi
s w
ork
on s
eman
tics
.
1.
Dav
idso
n (1
984a
) in
trod
uces
PC
as
a m
etho
dolo
gica
l pri
ncip
le in
sem
an
tics.
Acc
ordi
ng to
Dav
idso
n, a
rad
ical
inte
rpre
ter n
eeds
to
use
PC
to c
on
stru
ct a
Tar
ski-
styl
e th
eory
of
trut
h fo
r an
alie
n's
lang
uage
. Suc
h a
theo
ry,
Dav
idso
n cl
aim
s, a
mou
nts
to a
the
ory
of
mea
ning
for
tha
t la
ngua
ge.
Acc
ordi
ng t
o D
avid
son,
an
alie
n m
eans
by
a pa
rtic
ular
utt
eran
ce w
hat
ever
suc
h a
theo
ry e
ntai
ls th
at sh
e w
ould
mea
n by
it, a
nd b
elie
ves
wha
teve
r is
nec
essa
ry th
at sh
e be
take
n to
bel
ieve
in o
rder
to c
onst
ruct
suc
h a
theo
ry
of
mea
ning
in th
e fi
rst
plac
e.3
My
conc
ern
is n
ot w
heth
er D
avid
son
is r
ight
, but
how
, in
ligh
t of
its
use
spec
ific
ally
as
a m
etho
dolo
gica
l pri
ncip
le i
n se
man
tics
, D
avid
son
arti
cu
late
s pc
. R
elat
ive
to t
hat u
se,
Dav
idso
n of
fers
thr
ee,
som
etim
es o
verl
ap
ping
, ar
ticu
lati
ons
of
PC.
Fir
st, D
avid
son
says
tha
t to
con
stru
ct a
the
ory
of
mea
ning
for
an
alie
n's
lang
uage
, an
int
erpr
eter
max
imiz
es a
gree
men
t be
twee
n th
e al
ien
and·
hers
elf
as f
ar a
s po
ssib
le:
Cha
rity
in
inte
rpre
ting
the
wor
ds a
nd t
houg
hts
of
othe
rs i
s un
avoi
dabl
e· in
an
othe
r di
rect
ion
as w
ell:
just
as
we
mus
t m
axim
ize
agre
emen
t, o
r ri
sk n
ot
mak
ing
sens
e o
f w
hat t
he a
lien
is t
alki
ng a
bout
, so
we
mus
t max
imiz
e th
e se
lf
cons
iste
ncy
we
attr
ibut
e to
him
, o
n p
ain
of
not
unde
rsta
ndin
g hi
m.
(198
4a,
p. 2
7; m
y em
phas
is o
n "m
axim
ize
agre
emen
t,,)4
We
wan
t a t
heor
y th
at s
atis
fies
the
for
mal
con
stra
ints
on
a th
eory
of
trut
h, a
nd
th
at m
axim
izes
agr
eem
ent,
in th
e se
nse
of
mak
ing
[alie
ns]
righ
t, a
s fa
r as
we
can
tell,
as
ofte
n as
pos
sibl
e. (
1984
b, p
. 13
6; m
y em
phas
is)
Now
, ac
cord
ing
to D
avid
son
(198
4a,
I 984
b),
an i
nter
pret
er m
axim
izes
ag
reem
ent
on s
ente
nces
hel
d tr
ue,
and
sent
ence
s he
ld t
rue
are
belie
fs.
So
Dav
idso
n fi
rst
clai
ms
that
PC
adv
ises
an
inte
rpre
ter
to m
axim
ize
agre
em
ent o
n b
elie
fs s
hare
d by
the
ali
en a
nd h
erse
lf.
But
wha
t doe
s "m
axim
ize
agre
emen
t" m
ean?
In
the
lat
ter
quot
atio
n, D
avid
son
expl
icat
es i
t as
an
inte
rpre
ter'
s ta
king
an
alie
n to
be
righ
t by
the
inte
rpre
ter's
ligh
ts a
s of
ten
as p
ossi
ble.
T
his
expl
icat
ion
itse
lf s
eem
s to
be
a se
cond
art
icul
atio
n o
f P
C,
vari
ati
ons
of
whi
ch D
avid
son
offe
rs e
lsew
here
:
[pro
vidi
ng a
the
ory
of
mea
ning
] is
acc
ompl
ishe
d by
ass
igni
ng t
ruth
con
diti
ons
to a
lien
sen
tenc
es th
at m
ake
nati
ve s
peak
ers
righ
t whe
n pl
ausi
bly
poss
ible
, ac
cord
in
g, o
f cou
rse,
to
our
own
view
on
wha
t is
rig
ht.
(198
4b, p
. 13
7; m
y em
phas
is)
Pri
ncip
le o
f Cha
rity
67
3
The
gen
eral
pol
icy,
how
ever
, is
to c
hoos
e tr
uth
cond
itio
ns t
hat
do a
s w
ell a
s po
ssi
ble
in m
akin
g sp
eake
rs h
old
sent
ence
s tr
ue w
hen
(acc
ordi
ng t
o th
e th
eory
and
th
eory
bui
lder
's v
iew
on
the
fact
s) t
hose
sen
tenc
es a
re t
rue.
(19
84c,
p.
150;
my
emph
asis
)
We
get a
fir
st a
ppro
xim
atio
n o
f a
fini
shed
the
ory
by a
ssig
ning
to
sen
tenc
es o
f a
spea
ker
cond
itio
ns o
f tru
th t
hat
actu
ally
obt
ain
(in
our
own
opin
ion)
just
whe
n th
e sp
eake
r hol
ds th
ose
sent
ence
s tr
ue. T
he g
uidi
ng p
olic
y is
to
do th
isas
far
as
poss
ible
. (1
984d
, p.
196;
my
emph
asis
)
Fo
r Dav
idso
n, s
ince
a s
ente
nce
held
tru
e is
a b
elie
f, h
is s
econ
d ar
ticu
lati
on
of
PC
is t
hat
an
inte
rpre
ter
as fa
r as
pos
sibl
e ta
kes
an a
lien
to
have
bel
iefs
tr
ue b
y he
r in
terp
rete
r's
ligh
ts.
Dav
idso
n so
met
imes
use
s th
ese
firs
t tw
o ar
ticu
lati
ons
toge
ther
. B
ut
twic
e he
mod
ifie
s th
e fi
rst
from
"m
axim
ize"
to "
opti
miz
e":
The
met
hodo
logi
cal
advi
ce t
o i
nter
pret
in
a w
ay t
hat
opti
miz
es a
gree
men
t sh
ould
not
be
conc
eive
d as
res
ting
on
a ch
arit
able
ass
umpt
ion
abou
t hu
man
in
tell
igen
ce..
.. I
f w
e ca
nnot
find
a w
ay t
o in
terp
ret t
he u
tter
ance
s an
d ot
her
beha
viou
r of
a cr
eatu
re a
s re
veal
ing
a se
t of
belie
fs la
rgel
y co
nsis
tent
and
true
by
our
own
stan
dard
s, w
e ha
ve n
o re
ason
to
cou
nt th
at c
reat
ure
as r
atio
nal,
as
hav
ing
belie
fs, o
r as
say
ing
anyt
hing
. (1
984b
, p.
137;
my
emph
asis
)
The
bas
ic m
etho
dolo
gica
l pre
cept
is, t
here
fore
, tha
t a g
ood
theo
ry o
f in
terp
re
tati
on m
axim
izes
agr
eem
ent.
Or,
give
n th
at s
ente
nces
are
infi
nite
in n
umbe
r ..
. a
bett
er w
ord
mig
ht b
e op
timiz
e. (
1984
e, p
. 16
9)
The
se s
eem
to
exp
ress
a t
hird
art
icul
atio
n o
f P
C: a
n in
terp
rete
r op
tim
izes
ag
reem
ent
betw
een
the
alie
n an
d h
erse
lf.
No
w w
hy d
oes
Dav
idso
n sw
itch
fro
m "
max
imiz
e" t
o "
opti
miz
e"?
Th
ou
gh
Dav
idso
n m
ight
hav
e m
ore
than
one
rea
son,
the
latt
er q
uo
tati
on
m
akes
cle
ar t
hat
his
ch
ief r
easo
n in
volv
es h
is c
laim
th
at a
n a
lien
has
an
in
fini
te n
um
ber
of
beli
efs
and
th
at o
ne c
ann
ot
"max
imiz
e" a
n i
nfin
ite
num
ber
of
anyt
hing
.5
Why
doe
s D
avid
son
.cla
im t
hat
an
ali
en h
as a
n i
nfin
ite
nu
mb
er o
f be
lief
s? T
ho
ug
h m
y co
ncer
n is
no
t w
heth
er D
avid
son
is r
ight
, b
ut
rath
er
how
his
cla
imin
g th
is r
elat
es t
o h
is a
rtic
ulat
ing
PC
, le
t m
e no
neth
eles
s pr
esen
t D
avid
son'
s re
ason
for
cla
imin
g th
is.
Dav
idso
n do
es s
o, b
ecau
se i
t fo
llow
s fr
om t
he
way
in w
hich
he
uses
Tar
ski's
sem
anti
c th
eory
of
tru
th to
ge
nera
te a
the
ory
of
mea
ning
. Dav
idso
n ad
op
ts T
arsk
i's m
eth
od
of
recu
rsi
vely
gen
erat
ing
an in
fini
te n
umbe
r o
f T
-sen
tenc
es, e
ach
stat
ing
that
on
e o
f an
infi
nite
num
ber
of
recu
rsiv
ely
gene
rate
d ob
ject
-lan
guag
e se
nten
ces
is t
rue
if,
and
onl
y if
, on
e o
f an
inf
init
e nu
mbe
r o
f re
curs
ivel
y ge
nera
ted
met
alan
guag
e se
nten
ces
is t
rue.
And
, for
Dav
idso
n, e
ach
obje
ct-l
angu
age
674
Dia
logu
e
sent
ence
is
a se
nten
ce t
hat
an a
lien,
and
eac
h m
etal
angu
age
sent
ence
a
sent
ence
tha
t an
int
erpr
eter
, ho
lds
true
, re
spec
tivel
y.
Thu
s D
avid
son
wri
tes:
"T
he a
im o
f th
eory
will
be
an in
fini
te n
umbe
r of
sent
ence
s al
ike
in
trut
h,"
imm
edia
tely
con
tinu
ing:
"W
hat
the
[int
erpr
eter
] m
ust
do i
s fi
nd
out,
how
ever
he
can,
wha
t se
nten
ces
the
alie
n ho
lds
true
in
his
own
tong
ue"
(198
4a, p
. 27;
my
emph
asis
). N
ow, f
or D
avid
son,
sin
ce a
sen
tenc
e he
ldtr
ueju
stis
a b
elie
f, an
alie
nth
eref
ore
has
anin
fini
te n
umbe
r ofb
elie
fs.
One
mig
ht o
bjec
t th
at o
n D
avid
son'
s vi
ew a
n in
terp
rete
r ne
ed n
ot
attr
ibut
e to
an
alie
n an
infi
nite
num
ber o
f be
liefs
. Ins
tead
, she
can
take
the
alie
n to
spe
ak a
lan
guag
e w
ith
an i
nfin
ite n
umbe
r o
f se
nten
ces.
Yet
, fo
r D
avid
son,
a l
angu
age
just
is t
he s
et o
f se
nten
ces
(exp
ress
ed h
omop
honi
ca
lly) t
hat a
n al
ien
hold
s tr
ue. A
nd, f
or D
avid
son,
thes
e se
nten
ces
held
true
ar
e de
term
ined
by
an in
terp
rete
r w
hen
cons
truc
ting
a th
eory
of
mea
ning
. I
take
thi
s to
be
one
way
of
unde
rsta
ndin
g D
avid
son'
s re
mar
k: "
I co
ncl
ude
that
ther
e is
no
such
thi
ng a
s a
lang
uage
, no
t if
a l
angu
age
is a
ny
thin
g lik
e w
hat m
any
ling
uist
s ha
ve s
uppo
sed.
The
re is
the
refo
re n
o su
ch
thin
g to
be
lear
ned,
mas
tere
d, o
r bo
rn w
ith"
(19
86,
pp.
445-
46),
for
D
avid
son'
s po
int s
eem
s to
be
that
it is
a m
ista
ke to
und
erst
and
a la
ngua
ge
as s
omet
hing
exi
stin
g in
depe
nden
tly
of
part
icul
ar a
cts
of
inte
rpre
tati
on,
and
so p
arti
cula
r th
eori
es o
f m
eani
ng. N
ow s
ince
suc
h a
theo
ry c
onta
ins
an i
nfin
ite
num
ber
of
T-s
ente
nces
, co
rrel
atin
g an
inf
init
e nu
mbe
r o
f ob
ject
-lan
guag
e se
nten
ces,
whi
ch t
he a
lien
hold
s tr
ue,
wit
h an
inf
init
e nu
mbe
r o
f m
etal
angu
age
sent
ence
s, a
n al
ien
hold
s an
infi
nite
num
ber
of
sent
ence
s tr
ue.
And
so,
for
Dav
idso
n, t
he a
lien
has
an in
fini
te n
umbe
r o
f be
liefs
. O
ne m
ight
obj
ect
to a
n al
ien'
s ha
ving
an·
infi
nite
num
ber
of
belie
fs o
n it
s ow
n te
rms.
Aga
in,
my
conc
ern
is n
ot w
heth
er D
avid
son
is c
orre
ct.
Non
ethe
less
, le
t m
e su
gges
t w
hat
Dav
idso
n ha
s in
min
d. A
ccor
ding
to
Dav
idso
n, a
bel
ief i
s id
entif
ied
only
aga
inst
the
back
grou
nd o
f a
"sys
tem
" (1
984e
, p.
157
) o
r "d
ense
pat
tern
" (1
984f
, p.
200
) o
f be
liefs
. If
an
inte
rpr
eter
take
s an
alie
n to
bel
ieve
, e.g
., th
at it
is r
aini
ng, t
he in
terp
rete
r mig
ht
also
tak
e th
e al
ien
to b
elie
ve t
hat
rain
fal
ls,
that
rai
n fa
lls f
rom
the
sky
, th
at r
ain
falls
fro
m t
he s
ky t
o th
e gr
ound
, th
at r
ain
falls
fro
m t
he s
ky t
o th
e gr
ound
tod
ay,
and
that
the
grou
nd is
bel
ow t
he s
ky.
Fur
ther
ther
e is
in
pri
ncip
le n
o re
ason
why
an
inte
rpre
ter
need
sto
p. I
n fa
ct,
Dav
idso
n .
clai
ms
that
the
only
alt
erna
tive
to id
enti
fyin
g a
beli
ef a
gain
st s
uch
a ba
ck
grou
nd o
f be
liefs
is
to i
dent
ify
it a
s ex
pres
sing
an
anal
ytic
tru
th;6
but
D
avid
son
reje
cts
that
ther
e ar
e su
ch tr
uths
. R
egar
ding
one
's n
ot b
eing
abl
e to
"m
axim
ize"
an
infi
nite
num
ber
of
anyt
hing
, D
avid
son
(198
4b,
p. ·1
37;
1984
e, p
. 16
9) s
eem
s to
und
erst
and
"max
imiz
e ag
reem
ent"
as
som
ethi
ng li
ke t
akin
g th
e al
ien
and
inte
rpre
ter
to a
gree
on
mo
st b
elie
fs.
"Opt
imiz
e ag
reem
ent"
mig
ht t
hen
mea
n so
me
thin
g lik
e ta
king
the
alie
n an
d in
terp
rete
r to
agr
ee o
n as
man
y be
liefs
as
poss
ible
. Onl
y th
is e
xpla
ins
why
Dav
idso
n th
inks
tha
t max
imiz
ing
an in
fi
Pri
ncip
le o
f Cha
rity
67
5
nite
num
ber
of
thin
gs i
s un
inte
llig
ible
, an
d w
hy,
for
Dav
idso
n (1
984b
, p.
137
; 19
84e,
p.
169)
, "o
ptim
ize"
is p
refe
rabl
e to
"m
axim
ize"
: w
here
as
sens
e ca
n be
mad
e o
f ag
reei
ng o
n as
man
y o
f an
infi
nite
num
ber o
f be
liefs
as
pos
sibl
e, n
o se
nse
can
be m
ade
of
agre
eing
on
mo
st o
f an
infi
nite
num
be
r o
f be
liefs
. F
or a
ny n
umbe
r o
f be
liefs
less
tha
n th
e to
tal w
ould
be
infi
ni
tely
less
th
an th
e to
tal;
"m
ost"
can
not q
uant
ify
over
an
infi
nite
num
ber
of
anyt
hing
.7
Yet
, rec
all D
avid
son
(198
4b, p
. 13
6) e
xpli
cati
ng "
max
imiz
e ag
reem
ent"
as
taki
ng a
n a
lien
as
ofte
n as
pos
sibl
e to
be
righ
t by
the
inte
rpre
ter'
s lig
hts.
A
nd t
his
is c
onsi
sten
t w
ith
ther
e be
ing
an in
fini
te n
umbe
r o
f be
liefs
. In
fa
ct,
Dav
idso
n th
ere
seem
s to
und
erst
and
"max
imiz
e ag
reem
ent"
in
the
sam
e w
ay a
s he
els
ewhe
re (1
984b
, p. 1
37;
1984
e, p
. 16
9) u
nder
stan
ds "
opti
m
ize
agre
emen
t."
Non
ethe
less
, as
exp
lain
ed b
elow
, th
is a
mbi
guit
y do
es
not t
hrea
ten
my
unde
rsta
ndin
g hi
m.
Now
,· D
avid
son
atte
mpt
s to
use
PC
, as
it
func
tion
s se
man
tica
lly,
to
draw
tw
o co
nclu
sion
s in
epi
stem
olog
y:
that
sch
eme-
cont
ent
dual
ism
(1
984d
) an
d sc
epti
cism
(19
84f;
2001
b) a
re b
oth
unte
nabl
e.8
Rel
ativ
e to
thes
e us
es,
Dav
idso
n of
fers
two
diff
eren
t, co
nfli
ctin
g ar
tic
ulat
ions
of p
c. T
hese
art
icul
atio
ns s
till c
once
rn P
C a
s it
func
tion
s se
man
tic
ally
, bu
t, u
nlik
e th
e ot
hers
, ar
e m
eant
to
allo
w D
avid
son
to r
each
his
ep
iste
mic
con
clus
ions
. T
he fi
rst
new
art
icul
atio
n, t
he f
ourt
h in
tot
o, a
mou
nts
to t
he c
laim
that
to
con
stru
ct a
the
ory
of
mea
ning
, an
inte
rpre
ter
take
s m
ost
of a
n al
ien'
s be
lief
s to
be
true
by
her
inte
rpre
ter'
s lig
hts:
Cha
rity
is f
orce
d on
us;
whe
ther
we
like
it o
r not
, if
we
wan
t to
unde
rsta
nd o
th
ers,
we
mus
t cou
nt t
hem
rig
ht in
mos
t mat
ters
. (1
984d
, p.
197;
my
emph
asis
)
But
of
cour
se i
t ca
nnot
be
assu
med
tha
t sp
eake
rs n
ever
hav
e fa
lse
belie
fs ...
. W
e ca
n, h
owev
er, t
ake
it a
s a
give
n th
at m
ost b
elie
fs a
re c
orre
ct..
.. A
theo
ry o
f in
terp
reta
tion
can
not
be c
orre
ct. t
hat
mak
es a
man
ass
ent
to v
ery
man
y fa
lse
sent
ence
s: i
t m
ust
gene
rally
be
the
case
tha
t a
sent
ence
is t
rue
whe
n a
spea
ker
hold
s it
to b
e. (
1984
e, p
. 16
9)
So,
now
Dav
idso
n ex
plic
itly
talk
s o
f "m
ost"
of
an i
nfin
ite
num
ber
of
belie
fs.
And
this
fou
rth
arti
cula
tion
exp
licat
es t
he s
ense
of
the
firs
t ar
tic
ulat
ion
not e
xpli
cate
d by
the
sec
ond:
"m
axim
izin
g ag
reem
ent"
can
mea
n ei
ther
, as
per
the
sec
ond
arti
cula
tion
, ta
king
an
alie
n an
d in
terp
rete
r to
sh
are
as m
any
belie
fs a
s po
ssib
le,
or,
as p
er th
e fo
urth
, ta
king
mos
t of
an
alie
n's
belie
fs t
o be
tru
e by
her
inte
rpre
ter's
ligh
ts.
Dav
idso
n ne
eds
to a
rtic
ulat
e P
C in
the
latt
er w
ay t
o re
ach
his
epis
tem
ic
conc
lusi
ons.
For
Dav
idso
n (1
984d
) pur
port
s to
est
abli
sh h
is fi
rst e
pist
emic
co
nclu
sion
, th
e un
tena
bili
ty o
f sc
hem
e-co
nten
t du
alis
m,
by e
stab
lish
ing
the
impo
ssib
ilit
y o
f tw
o ki
nds
of f
ailu
res
of t
rans
lata
bili
ty, v
iz.,
"com
plet
e,
676
Dia
logu
e
and
part
ial,
fai
lure
s o
f tr
ansl
atab
ilit
y. T
here
wou
ld b
e co
mpl
ete
fail
ure
if
no s
igni
fica
nt ra
nge
of [
true
] sen
tenc
es in
one
lang
uage
cou
ld b
e tr
ansl
ated
in
to t
he o
ther
; th
ere
wou
ld b
e pa
rtia
l fa
ilur
e if
som
e si
gnif
ican
t ra
nge
coul
d be
tran
slat
ed a
nd s
ome
rang
e co
uld
not"
(198
4d, p
. 18
5).9
If
part
ial
fail
ure
of
tran
slat
abil
ity
is i
mpo
ssib
le,
then
no
sign
ific
ant
rang
e o
f tr
ue
sent
ence
s in
one
lang
uage
fai
ls t
o be
tra
nsla
tabl
e in
to a
noth
er la
ngua
ge.
Tho
ugh
Dav
idso
n is
not
exp
licit
on w
hat n
o si
gnif
ican
t ran
ge o
f su
ch s
en
tenc
es e
ntai
ls, h
e do
es w
rite
in th
e sa
me
artic
le, a
s qu
oted
abo
ve, t
hat "
we
mus
t cou
nt [o
ur in
terl
ocut
ors]
rig
ht in
mo
st m
atte
rs."
Thu
s, w
hate
ver
else
it
ent
ails
, no
sign
ific
ant r
ange
of s
uch
sent
ence
s en
tail
s th
at m
ost s
ente
nces
he
ld t
rue
in o
ne la
ngua
ge a
re tr
ansl
atab
le i
nto
anot
her.
So,
for
Dav
idso
n to
est
abli
sh th
e un
tena
bili
ty o
f sc
hem
e-co
nten
t dua
lism
, he
mus
t est
abli
sh
that
mos
t sen
tenc
es h
eld
true
in o
ne la
ngua
ge n
eed
to b
e tr
ansl
atab
le in
to
anot
her.
And
sin
ce D
avid
son
does
so
by r
elyi
ng o
n h
is a
ccou
nt o
f ra
dica
l in
terp
reta
tion
, mos
t sen
tenc
es h
eld
true
in o
ne la
ngua
ge n
eed
to b
e tr
ans
lata
ble
into
the
radi
cal
inte
rpre
ter's
lang
uage
. D
avid
son'
s se
cond
epi
stem
ic c
oncl
usio
n, t
hat
scep
tici
sm i
s un
tena
ble,
al
so r
equi
res
talk
of
"mos
t" b
elie
fs.
Dav
idso
n's
argu
men
t ag
ains
t sc
epti
ci
sm is
com
plic
ated
; it
invo
lves
arg
uing
fro
m P
C's
spe
cify
ing
trut
h by
an
inte
rpre
ter'
s lig
hts,
plu
s th
e po
ssib
ilit
y o
f an
"om
nisc
ient
inte
rpre
ter,
" 10 to
re
ach
the
follo
win
g co
nclu
sion
s:
But
now
it
is p
lain
why
mas
sive
err
or a
bout
the
wor
ld is
sim
ply
unin
tell
igib
le.
(198
4f, p
. 20
1; m
y em
phas
is)
Onc
e w
e ag
ree
to t
he g
ener
al m
etho
d o
f in
terp
reta
tion
I h
ave
sket
ched
, it
be
com
es im
poss
ible
cor
rect
ly t
o ho
ld th
at a
nyon
e co
uld
be m
ostl
y w
rong
abo
ut
how
thi
ngs
are.
(20
01b,
p.
151;
my
emph
asis
)
Fro
m th
e re
st o
f D
avid
son'
s ar
gum
ent,
it
beco
mes
cle
ar t
hat
his
conc
lu
sion
is n
ot m
erel
y th
at m
assi
ve e
rror
is u
nint
elli
gibl
e an
d th
at it
is i
mpo
ssi
ble
corr
ectl
y to
hol
d th
at a
nyon
e co
uld
be m
ostl
y w
ron
g a
bout
how
th
ings
are
. His
con
clus
ion
is t
hat o
nly
mas
sive
trut
h is
inte
lligi
ble,
and
that
it
is i
mpo
ssib
le t
o ho
ld t
hat
anyo
ne c
ould
not
be
mos
tly
righ
t ab
out h
ow
thin
gs a
re.
For
ruli
ng o
ut t
he p
ossi
bili
ty o
f m
assi
ve e
rror
or
bein
g m
ostl
y w
rong
stil
l allo
ws
one
to b
e ri
ght a
nd w
rong
the
sam
e am
ount
of t
ime,
and
su
ch a
sta
te o
f af
fair
s do
es n
ot
esta
blis
h th
e un
tena
bili
ty o
f sc
epti
cism
. R
egar
dles
s, a
s w
ith
the
case
of
"no
sig
nifi
cant
ran
ge,"
Dav
idso
n is
no
t ex
plic
it o
n w
hat
"mas
sive
err
or"
or
"mas
sive
tru
th"
and
"mos
tly
wro
ng"
or "
mos
tly
righ
t" e
ntai
l. N
onet
hele
ss, h
is u
ses
of
them
sug
gest
that
, wha
tev
er e
lse
they
ent
ail,
the
y en
tail
tha
t m
ost
belie
fs a
re e
rron
eous
or
true
, an
d w
rong
or
righ
t, r
espe
ctiv
ely.
For
oth
erw
ise
Dav
idso
n w
ould
no
t be
ab
le t
o es
tabl
ish
the
unte
nabi
lity
of
scep
ticis
m.
And
so
his
conc
lusi
ons
amou
nt t
o th
e cl
aim
tha
t it i
s ne
cess
ary
that
mos
t be
liefs
are
true
. R
ecal
l
Pri
ncip
le o
f Cha
rity
67
7
that
Dav
idso
n's
argu
men
t for
thi
s in
volv
es h
is r
elyi
ng o
n P
C's
spe
cify
ing
trut
h by
an
inte
rpre
ter'
s lig
hts.
But
then
to a
rgue
that
sce
ptic
ism
is u
nten
ab
le,
D'a
vids
on n
eeds
to
arti
cula
te P
C a
s an
inte
rpre
ter's
tak
ing
mos
t o
f an
ali
en's
bel
iefs
to
be t
rue
by h
er in
terp
rete
r's li
ghts
. B
ut h
ow c
an D
avid
son
talk
abo
ut "
mos
t" b
elie
fs?
For
Dav
idso
n is
rig
ht
that
no
sens
e ca
n be
mad
e o
f mos
t of
an in
fini
te n
umbe
r of
anyt
hing
, and
, re
call,
for
him
, an
ali
en h
as a
n in
fini
te n
umbe
r o
f be
liefs
. In
fac
t, re
call,
th
is w
orry
pus
hes
Dav
idso
n to
pre
fer
talk
of
"opt
imiz
ing"
to "
max
imiz
in
g" th
e nu
mbe
r o
f tr
ue b
elie
fs i
n th
e fi
rst
plac
e.
Dav
idso
n, c
ogni
zant
of
prob
lem
s co
ncer
ning
"m
axim
izin
g,"
is a
lso
cogn
izan
t of p
robl
ems
conc
erni
ng "
mos
t,"
for
Dav
idso
n tr
ies
to e
xplic
ate,
w
ith
a fi
fth
arti
cula
tion
of
PC
, wha
t he
mea
ns b
y "m
ost b
elie
fs."
Wri
ting
o
f a
cohe
renc
e th
eory
of
trut
h, D
avid
son
expl
ains
: "A
ll th
at a
coh
eren
ce
theo
ry c
an m
aint
ain
is t
hat
mo
st o
f th
e be
lief
s in
a c
oher
ent
tota
l se
t o
f be
liefs
are
tru
e" (2
001
b, p
. 13
8). B
ut D
avid
son
imm
edia
tely
qua
lifi
es th
is:
Thi
s w
ay o
f st
atin
g th
e po
siti
on c
an a
t bes
t be
take
n as
a h
int,
sin
ce th
ere
is n
o us
eful
way
to
coun
t bel
iefs
, and
so
no c
lear
mea
ning
to t
he id
ea t
hat m
ost o
f a
pers
on's
bel
iefs
are
true
. A
som
ewha
t bet
ter
way
to
pu
t the
poi
nt is
to
say
ther
e is
a p
resu
mpt
ion
in fa
vor
of
the
trut
h o
f a
beli
ef t
hat
cohe
res
wit
h a
sign
ific
ant
mas
s o
f be
lief.
(ibi
d., p
p. 1
38-3
9; m
y em
phas
is)
Thu
s, D
avid
son
here
cla
ims,
the
re is
a p
resu
mpt
ion
in f
avou
r o
f th
e tr
uth
of
a be
lief
tha
t coh
eres
wit
h a
sign
ific
ant m
ass
of
belie
f. T
his
quot
atio
n is
not
an
expr
essi
on o
f P
C p
er s
e b
ut a
n ep
iste
mic
con
se
quen
ce o
f P
C t
hat
Dav
idso
n us
es a
s an
int
erm
edia
ry s
tep
in a
rgui
ng
from
PC
, and
his
vie
ws
on s
eman
tics
gen
eral
ly, a
gain
st s
cept
icis
m. F
or P
C
as a
met
hodo
logi
cal p
rinc
iple
in s
eman
tics
doe
s no
t per
se
guar
ante
e th
at
mos
t be
liefs
are
tru
e si
mpl
icit
er,
even
if
unde
rsto
od a
s th
ere
bein
g a
pre
sum
ptio
n th
at a
bel
ief
appr
opri
atel
y co
heri
ng is
tru
e. I
t gu
aran
tees
onl
y th
at m
ost
belie
fs a
re t
rue
by a
n in
terp
rete
r's li
ghts
. In
oth
er w
ords
, th
ere
is m
ore
to D
avid
son'
s ar
gum
ent,
fro
m w
hich
thi
s m
ost
rece
nt e
xcer
pt is
ta
ken,
tha
n pc
.II
Non
ethe
less
, ke
epin
g th
is i
n m
ind,
I u
nder
stan
d th
e qu
otat
ion
as s
ugge
stin
g a
fift
h ar
ticu
lati
on o
f P
C it
self
: th
ere
is a
pre
sum
pti
on i
n th
e fa
vour
of a
n al
ien'
s be
lief
s be
ing
true
by
an i
nter
pret
er's
ligh
ts.
-2. Thu
s D
avid
son
offe
rs fi
ve,
som
etim
es o
verl
appi
ng, a
rtic
ulat
ions
of
PC
:
(i)
an
inte
rpre
ter
max
imiz
es a
gree
men
t on
bel
iefs
bet
wee
n he
r al
ien
and
hers
elf
as f
ar a
s po
ssib
le;
(ii)
an
int
erpr
eter
as
far
as p
ossi
ble
take
s an
ali
en t
o ha
ve b
elie
fs t
rue
by h
er in
terp
,ret
er's
ligh
ts;
678
D
ialo
gue
(iii)
an
int
erpr
eter
opt
imiz
es a
gree
men
t be
twee
n th
e al
ien
and
hers
elf
(= a
n in
terp
rete
r ta
kes
an a
lien
and
her
self
to
agr
ee o
n a
s m
any
belie
fs a
s po
ssib
le);
(iv)
an
int
erpr
eter
tak
es m
ost
of
an a
lien'
s be
liefs
to
be t
rue
by h
er
inte
rpre
ter'
s lig
hts;
and
(v)
ther
e is
a p
resu
mpt
ion
in th
e fa
vour
of
an a
lien
's b
elie
fs b
eing
true
by
an
inte
rpre
ter'
s lig
hts.
Whi
ch o
f th
ese
is b
est?
For
Dav
idso
n, s
ince
(i)
is a
mbi
guou
s, e
xpli
cabl
e in
the
sen
se o
f ei
ther
(ii)
or
(iv)
, le
t me
cons
ider
onl
y (i
i)-(
v).
Art
icul
atio
ns (i
i), (
iii),
and
(v)
are
all t
oo w
eak
to re
ach
eith
er o
f D
avid
so
n's
epis
tem
ic c
oncl
usio
ns.
Con
side
r (i
i).
If a
n in
terp
rete
r as
far
as p
os
sibl
e ta
kes
an a
lien
to
have
bel
iefs
tru
e by
her
inte
rpre
ter'
s li
ghts
, the
n an
al
ien
mig
ht h
ave
no b
elie
f tr
ue b
y he
r int
erpr
eter
's li
ghts
. 12
Hen
ce, m
ost
of
the
alie
n's
belie
fs n
eed
not
be
true
by
her i
nter
pret
er's
ligh
ts. B
ut th
en b
oth
of
Dav
idso
n's
epis
tem
ic c
oncl
usio
ns a
re b
lock
ed.
Fir
st,
tran
slat
ion
is
trut
h-pr
eser
ving
. So
if
mos
t o
f an
alie
n's
belie
fs,
and
so s
ente
nces
hel
d tr
ue in
her
lang
uage
, nee
d n
ot b
e tr
ue b
y he
r in
terp
rete
r's
ligh
ts, t
hen
mos
t se
nten
ces
held
true
in h
er la
ngua
ge n
eed
no
t be
tran
slat
able
into
her
inte
rpr
eter
's la
ngua
ge.
But
then
Dav
idso
n's
argu
men
t aga
inst
sch
eme-
cont
ent
dual
ism
fai
ls.
Sec
ond,
Dav
idso
n's
argu
men
t ag
ains
t sc
epti
cism
pur
port
s to
sho
w t
hat
mos
t o
f an
alie
n's
belie
fs n
eed
to b
e tr
ue,
by
show
ing
(int
er
alia
) th
at m
ost
need
to
be t
rue
by h
er in
terp
rete
r's
light
s. A
nd s
o D
avid
so
n's
argu
men
t aga
inst
sce
ptic
ism
fails
. H
ence
, w
ere
Dav
idso
n to
art
icul
ate
PC
as
(ii)
, th
en h
e w
ould
fai
l to
es
tabl
ish
that
eit
her
sche
me-
cont
ent
dual
ism
or
scep
tici
sm i
s un
tena
ble.
F
urth
er,
the
poin
t gen
eral
izes
to
(iii)
and
(v),
for
eac
h ar
ticu
lati
on is
als
o co
nsis
tent
wit
h an
alie
n's
havi
ng n
o be
lief
tru
e by
her
inte
rpre
ter'
s lig
hts.
A
nd s
o, w
ere
Dav
idso
n to
art
icul
ate
PC
as
eith
er o
f th
ese,
the
n he
wou
ld
likew
ise
fail.
T
hat
leav
es o
nly
(iv)
. D
avid
son'
s ep
iste
mic
con
clus
ions
, re
call,
req
uire
so
met
hing
at l
east
as
stro
ng.
Hen
ce, s
olvi
ng b
ackw
ard
from
the
se c
oncl
usi
ons
to th
eir p
rem
ises
, (iv
) se
ems
the
best
art
icul
atio
n o
f P
C fo
r D
avid
son.
N
onet
hele
ss, (
iv)
invo
lves
mak
ing
sens
e o
f "m
ost"
of
an in
fini
te n
umbe
r.
of
belie
fs, w
hich
Dav
idso
n hi
mse
lf c
laim
s im
poss
ible
. The
re a
re n
ever
the
less
at
leas
t si
x w
ays
in w
hich
Dav
idso
n m
ight
be
able
to
do s
o af
ter
all.
Fir
st,
he m
ight
use
"m
ost
belie
fs"
figu
rativ
ely.
In
fac
t, th
is s
eem
s D
avid
so
n's
stra
tegy
. B
ut th
e on
ly w
ay i
n w
hich
usi
ng "
mos
t bel
iefs
" fi
gura
tivel
y ca
n he
lp is
if
it a
llow
s D
avid
son
to m
ean
both
som
ethi
ng li
ke a
pre
sum
pti
on in
fav
our
of
a ce
rtai
n ki
nd o
f (v
iz.,
true
) be
liefs
an
d m
ost b
elie
fs.
For
on
ly s
omet
hing
like
the
for
mer
can
, by
Dav
idso
n's
own
light
s, a
void
uni
nte
lligi
bilit
y, w
hile
onl
y so
met
hing
lik
e th
at l
atte
r ca
n, a
s I
have
sho
wn,
Pri
ncip
le o
f Cha
rity
67
9
allo
w D
avid
son
to u
se P
C to
arg
ue f
or h
is e
pist
emic
con
clus
ions
. B
ut th
en
this
"fi
gura
tive
" us
e co
mm
its
Dav
idso
n to
bot
h am
bigu
ity
(inv
olvi
ng tw
o di
ffer
ent s
ense
s o
f "m
ost b
elie
f") a
nd
unin
tell
igib
ilit
y (i
nvol
ving
mos
t of a
n in
fini
te n
umbe
r o
f so
met
hing
). S
o, t
his
firs
t w
ay in
whi
ch D
avid
son
mig
ht
mak
e se
nse
of
"mos
t bel
iefs
" le
aves
him
in a
wor
se p
osit
ion
than
bef
ore.
S
econ
d, D
avid
son
mig
ht s
impl
y ob
serv
e th
at a
n in
terp
rete
r ta
kes
such
"o
rdin
ary"
bel
iefs
as
that
it is
rai
ning
, bu
t no
t ne
cess
aril
y su
ch "
extr
aor
dina
ry"
belie
fs a
s th
at e
utha
nasi
a is
per
mis
sibl
e, t
o b
e tr
ue b
y he
r lig
hts.
A
nd D
avid
son
mig
ht a
rgue
tha
t, w
hate
ver
diff
icul
ty t
here
is i
n re
gim
ent
ing
the
clai
m t
hat
ther
e ar
e "m
any
mor
e" o
rdin
ary
than
ext
raor
dina
ry
belie
fs, t
here
is n
o di
sput
ing
that
in s
ome
sens
e th
is c
laim
is t
rue.
But
then
it
is a
pla
in f
act
that
in
som
e se
nse
an i
nter
pret
er d
oes
take
mos
t o
f an
al
ien'
s be
liefs
to
be t
rue
by h
er li
ghts
. T
he p
robl
em w
ith
this
resp
onse
, bes
ides
its
leav
ing
the
sens
e o
f "i
n so
me
sens
e" o
paqu
e, is
tha
t on
Dav
idso
n's
view
it i
s n
ot
a pl
ain
fact
tha
t in
any
sens
e th
ere
are
"man
y m
ore"
ord
inar
y th
an e
xtra
ordi
nary
bel
iefs
. A
bove
, I
show
ed th
at, f
or D
avid
son,
if
an in
terp
rete
r tak
es a
n al
ien
to b
elie
ve th
at
it is
rai
ning
, th
en s
he m
ight
als
o ta
ke t
he· a
lien
to
belie
ve t
hat
rai
n fa
lls,
that
rai
n fa
lls f
rom
the
sky
, tha
t rai
n fa
lls f
rom
the
sky
to
the
grou
nd, e
tc.
Her
e I
sugg
est t
hat,
for
Dav
idso
n, if
an
inte
rpre
ter t
akes
an
ali
en to
bel
ieve
th
at e
utha
nasi
a is
per
mis
sibl
e, th
en s
he m
ight
als
o ta
ke th
e al
ien
to b
elie
ve
that
"eu
than
asia
"wou
ld d
escr
ibe
the
act
done
to
a,
b,c
, et
c.,
bu
t n
ot
to
p,
q, r
, etc
.; th
at e
utha
nasi
a w
ould
be
perm
issi
ble
in c
ase
one
beca
use
of
a,
b, c
, etc
., bu
t no
t bec
ause
of p
,q, r
, etc
., in
cas
e tw
o be
caus
e o
f a,
b, c
, etc
., bu
t no
t bec
ause
ofp
, q,
r, et
c., i
n ca
se th
ree
beca
use
of
a, b
, c, e
tc.,
but n
ot
beca
use
of
p,q,
r,
etc.
, et
c.;
that
eut
hana
sia
is f
row
ned
up
on
by
thos
e be
liev
ing
a, b
, c, e
tc.,
but n
ot t
hose
bel
ievi
ngp,
q, r
, etc
.; et
c. A
nd a
ll th
ese
mig
ht b
e be
liefs
tha
t th
e in
terp
rete
r w
ould
her
self
hol
d fa
lse.
But
the
n it
is
no
t a p
lain
fac
t th
at a
n in
terp
rete
r w
ould
take
an
alie
n to
bel
ieve
man
y m
ore
ordi
nary
tha
n ex
trao
rdin
ary
belie
fs,
so t
his
wou
ld n
ot
be a
way
of
Dav
idso
n's
expl
aini
ng h
ow a
n in
terp
rete
r co
uld
take
mo
st o
f an
ali
en's
be
liefs
to
be
true
by
her
inte
rpre
ter's
ligh
ts.
Thi
rd,
Dav
idso
n m
ight
arg
ue t
hat
the
num
ber
of
belie
fs t
hat
any
alie
n ha
s is
fin
ite.
Yet
, re
call,
thi
s co
ntra
dict
s a
cent
ral
tene
t o
f D
avid
son'
s ad
opti
ng a
Tar
ski-
styl
e th
eory
of
tru
th a
s a
theo
ry o
f m
eani
ng.
Fur
ther
, al
so r
ecal
l, D
avid
son
seem
s to
cla
im t
hat t
he o
nly
alte
rnat
ive
to id
enti
fy
ing
a be
lief
aga
inst
wha
t is
in p
rinc
iple
an
infi
nite
num
ber
of
othe
r be
liefs
is
id
enti
fyin
g it
as
expr
essi
ng a
n an
alyt
ic t
ruth
-an
alt
erna
tive
th
at
Dav
idso
n re
ject
s.
Fou
rth,
Dav
idso
n m
ight
arg
ue t
hat,
tho
ugh
in p
rinc
iple
the
num
ber
of
belie
fs t
hat a
ny a
lien
has
is in
fini
te, i
n pr
acti
ce a
n in
terp
rete
r tak
es a
n al
ien
to h
ave
only
a f
inite
num
ber
of
belie
fs. A
nd o
f th
ese
it d
oes
mak
e se
nse
to
say
that
mos
t are
true
by
her
inte
rpre
ter'
s lig
hts.
Yet
, acc
ordi
ng to
Dav
id
son,
rec
all,
an a
lien
mea
ns b
y a
part
icul
ar u
tter
ance
wha
teve
r a
theo
ry o
f
680
Dia
logu
e
mea
ning
for
her
lang
uage
ent
ails
that
she
wou
ld m
ean
by it
, and
bel
ieve
s w
hate
ver
is n
eces
sary
tha
t sh
e be
tak
en t
o be
lieve
in
orde
r to
con
stru
ct
such
a th
eory
of m
eani
ng in
the
firs
t pla
ce. N
ow a
theo
ry o
f mea
ning
, eve
n in
pra
ctic
e, e
ntai
ls a
n in
fini
te n
umbe
r o
f T
-sen
tenc
es,
each
par
ticu
lar
T
sent
ence
ent
aili
ng w
hat
an a
lien
wou
ld m
ean
by a
par
ticu
lar.
utte
ranc
e.
And
, to
con
stru
ct a
n in
fini
te n
umbe
r o
f T
-sen
tenc
es,
and
so a
the
ory
of
mea
ning
, eve
n in
pra
ctic
e, it
is n
eces
sary
to ta
ke a
n al
ien
to h
ave
an in
fini
te
num
ber
of
belie
fs.
So, f
or D
avid
son,
eve
n in
pra
ctic
e, a
n in
terp
rete
r tak
es
an a
lien
to h
ave
an in
fini
te n
umbe
r o
f be
liefs
. F
ifth
, D
avid
son
mig
ht a
rgue
tha
t, e
ven
thou
gh th
e to
tal s
et o
f be
liefs
is
inf
inite
, bec
ause
he
(197
0) c
laim
s th
at b
elie
fs a
re m
enta
l st
ates
tok
en
iden
tica
l w
ith
brai
n st
ates
, be
liefs
mus
t oc
cupy
phy
sica
l sp
ace
in b
rain
s.
And
, sin
ce b
rain
s ha
ve o
nly
fini
te s
pace
, the
re c
an b
e on
ly a
fin
ite n
umbe
r o
f be
liefs
"in
" an
y on
e o
f th
em.. T
houg
h th
e di
ffer
ence
bet
wee
n be
liefs
"i
n" a
nd n
ot
"in"
the
brai
n w
ould
nee
d ex
plan
atio
n, t
here
wou
ld b
e no
pr
oble
m w
ith
"mos
t" a
ppli
ed t
o th
ose
in t
he b
rain
. O
ne r
espo
nse
to s
uch
an a
rgum
ent
wou
ld b
e m
erel
y to
obs
erve
tha
t D
avid
son'
s vi
ews
on s
eman
tics
are
not
nece
ssar
ily c
onsi
sten
t w
ith
his
view
s on
min
d, f
or,
acco
rdin
g to
the
for
mer
, be
liefs
are
the
oret
ical
con
st
ruct
s at
trib
utab
le d
urin
g in
terp
reta
tion
, and
it is
· not
cle
ar h~w
theo
ret
ical
con
stru
cts
can
be t
oken
-ide
ntic
al w
ith
anyt
hing
phy
sica
l. E
vnin
e (1
991,
con
clus
ion)
and
Chi
ld (
1994
, ch
ap.
4) w
orry
abo
ut th
is i
ncon
sist
en
cy,
the
form
er g
oing
so
far
as t
o ar
gue
that
Dav
idso
n's
view
s on
min
d sh
ould
be
jett
ison
ed i
f ir
revo
cabl
y in
cons
iste
nt w
ith
his
view
s on
sem
an
tics..
At
the
very
lea
st,
this
rai
ses
the
wor
ry t
hat
Dav
idso
n's
view
s on
se
man
tics,
epi
stem
olog
y, a
nd
min
d m
ight
all
be
inco
nsis
tent
. A
noth
er
resp
onse
wou
ld b
e m
erel
y to
obs
erve
that
, reg
ardl
ess
of
whe
ther
ther
e ca
n be
onl
y a
fini
te n
umbe
r o
f be
liefs
"in
" a
brai
n, I
jus
t sh
owed
tha
t, f
or
Dav
idso
n, e
ven
in p
ract
ice,
an
inte
rpre
ter
none
thel
ess
take
s an
ali
en t
o ha
ve a
n in
fini
te n
umbe
r of
belie
fs. S
o, t
he p
robl
em o
f qu
anti
fyin
g ov
er a
n in
fini
te s
et r
emai
ns.
And
six
th,
Dav
idso
n m
ight
tra
nsla
te "
mos
t be
liefs
" in
to s
ome
non
quan
tifi
cati
onal
idio
m. P
erha
ps h
e m
ight
art
icul
ate
PC
as
som
ethi
ng li
ke:
(vi)
cet
eris
par
ibus
, an
int
erpr
eter
take
s an
alie
n's
belie
fs t
o be
tru
e by
he
r in
terp
rete
r's li
ghts
.
He
mig
ht t
hen
prov
ide
a no
n-qu
anti
fica
tion
al c
onst
rual
of
the
cete
ris
pari
bus
clau
se.
Yet
(vi
) is
no
bett
er t
han
(ii)
, (i
ii),
or
(v).
For
(vi
) is
con
si
sten
t wit
h no
bel
ief o
f an
alie
n's
bein
g tr
ue b
y he
r int
erpr
eter
's li
ghts
. No
r is
it
clea
r ho
w t
rans
lati
ng "
mos
t be
liefs
" in
to a
ny n
on-q
uant
ific
atio
nal
idio
m c
an a
void
thi
s pr
oble
m.
Thu
s, g
iven
Dav
idso
n's
view
s on
sem
antic
s, n
one
of
thes
e w
ays
of
han
dlin
g "m
ost b
elie
fs"
succ
eeds
. Non
ethe
less
, Dav
idso
n m
ight
try
a di
ffer
ent
Pri
ncip
le o
f Cha
rity
68
1
tack
. H
e m
ight
arg
ue th
at th
e pr
oble
m o
fmak
ing
sens
e o
f "m
ost
bel
iefs
" is
mer
ely
tech
nica
l and
no
t wor
th th
e at
tent
ion
that
I ha
ve p
aid
it. Y
et o
nly
by i
gnor
ing
the
prob
lem
can
Dav
idso
n es
tabl
ish
his·
sw
eepi
ngep
iste
mic
co
nclu
sion
s. A
nd
so
the
prob
lem
is w
orth
the
atte
ntio
n th
at I
have
pai
d it
, in
sofa
r as
one
take
s D
avid
son
seri
ousl
y th
at h
is v
iew
s o
n s
eman
tics
hav
e ep
iste
mic
con
sequ
ence
s. N
or
can
one
fail
to
tak
e D
avid
son
seri
ousl
y ab
out t
his
sinc
e, a
ccor
ding
to
Dav
idso
n, h
is "
met
hodo
logy
of
inte
rpre
ta
tion
is
. . .
noth
ing
bu
t ep
iste
mol
ogy
seen
in
the
mir
ror
of
mea
ning
" (1
984e
, p.
169)
. H
ence
(iv)
, th
e on
ly a
rtic
ulat
ion
of
PC
cap
able
of e
ntai
ling
Dav
idso
n's
epis
tem
ic c
oncl
usio
ns i
s in
cons
iste
nt w
ith
.. his
. vie
ws
on
sem
anti
cs.
Yet
be
caus
e su
ch v
iew
s sa
ncti
on D
avid
son'
s us
ing
PC
in
the
fir
st p
lace-P
C
is a
met
hodo
logi
cal p
rinc
iple
in s
eman
tics
-an
y a
rtic
ulat
ion
of
PC
inco
nsi
sten
t ,w
ith D
avid
son'
s se
man
tic
view
s in
vali
date
s hi
s us
ing
PC
gen
eral
ly.
Thu
s, D
avid
son
can
use
PC
eit
her
only
as
a m
etho
dolo
gica
l pr
inci
ple
in
sem
anti
cs o
r n
ot
at a
ll. 13
Not
es
1 T
hat
thir
d an
thol
ogy
is D
avid
son
2001
a, i
tsel
f an
tici
pati
ng p
ubli
cati
on o
f D
avid
son
2004
and
fort
hcom
ing,
bri
ngin
g th
e to
tal n
umbe
r o
f an
thol
ogie
s o
f D
avid
son'
s ar
ticl
es t
o fiv
e. (
See
Dav
idso
n 20
01a,
pp.
221
-25.
) 2
Non
ethe
less
, se
e M
alpa
s (1
992,
§5.
3.3)
for
a s
ympa
thet
ic d
iscu
ssio
n o
f th
e tr
icki
ness
of
the
task
. 3
"Wha
t a f
ully
inf
orm
ed in
terp
rete
r cou
ld le
arn
abou
t wha
t a s
peak
er m
eans
is
all
ther
e is
to
lear
n; t
he s
ame
goes
for
wha
t th
e sp
eake
r be
lieve
s" (
Dav
idso
n 20
01b,
p.
148;
my
emph
asis
).
4 A
ccor
ding
to
Dav
idso
n, a
n in
terp
rete
r he
rsel
f ha
s la
rgel
y se
lf-c
onsi
sten
t be
liefs
; oth
erw
ise,
acc
ordi
ng to
him
, the
inte
rpre
ter c
ould
not
her
self
be
inte
rpr
eted
, an
d s
o co
uld
not
have
an
y be
liefs
. T
hus,
the
int
erpr
eter
's m
axim
izin
g th
e se
lf-c
onsi
sten
cy th
at s
he a
ttri
bute
s to
an
alie
n ca
n be
reg
arde
d as
a f
urth
er
inst
ance
of
her m
axim
izin
g he
r agr
eem
ent w
ith
the
alie
n.
5 N
onet
hele
ss,
see
n.7
for
a fu
rthe
r re
ason
. 6
In fa
ct, D
avid
son
(199
0, p
. 319
; 19
91, p
. 19
5) m
akes
this
poi
nt in
term
s o
f PC
. F
or h
e sa
ys t
hat P
C, i
n th
e co
ntex
t of
a T
arsk
i-st
yle
theo
ry o
f tr
uth,
an
d so
in
the
cont
ext
of
attr
ibut
ing
an i
nfin
ite
num
ber
of
belie
fs,
is a
n al
tern
ativ
e to
id
enti
fyin
g an
y be
lief
as
expr
essi
ng a
n an
alyt
ic t
ruth
. 7
The
re m
ay b
e a
furt
her
reas
on w
hy D
avid
son
switc
hes
from
"m
axim
ize"
to
"opt
imiz
e."
Tho
ugh
in la
ter w
riti
ng D
avid
son
cont
inue
s to
em
ploy
PC
"ac
ross
th
e bo
ard"
(198
4g, p
. xvi
i), h
e al
so a
rgue
s th
at s
ome
agre
emen
t is
mor
e im
por
tant
than
oth
ers.
And
, in
the
cont
ext o
f see
king
wei
ghte
dag
reem
ent,
rath
er th
an
mer
ely
coun
ting
the
num
ber
of
belie
fs o
n w
hich
the
re is
agr
eem
ent,
Dav
idso
n so
met
imes
spe
aks
of
"opt
imiz
ing.
" (I
tha
nk a
n an
onym
ous
revi
ewer
for
bri
ng
ing
this
to
my
atte
ntio
n.)
Non
ethe
less
, thi
s w
ould
be
a fu
rthe
r re
ason
, for
, as
all
682
Dia
logu
e
thes
e pa
ssag
es (i
nclu
ding
200
1b, p
p. 1
38-3
9, q
uote
d ab
ove)
mak
e cl
ear,
the
chie
f re
ason
for
the
switc
h is
the
wor
ry th
at a
n al
ien
has
an in
fini
te n
umbe
r of b
elie
fs.
8 D
avid
son'
s (1
984d
) ar
gum
ent a
gain
st s
chem
e-co
nten
t dua
lism
is a
lso
anar
gu
m
ent
agai
nst
conc
eptu
al r
elat
ivis
m.
Dav
idso
n ta
kes
thes
e co
nclu
sion
s to
be
rela
ted
to s
uch
furt
her
conc
lusi
ons
as th
at th
ere
is n
o su
ch th
ing
asla
ngua
ge,
in t
he t
radi
tion
al s
ense
of
"lan
guag
e" (
1986
); t
hat
ther
e is
no
such
thi
ng·.a
s ex
clus
ivel
y su
bjec
tive
know
ledg
e (1
988)
; an
d th
at o
ne s
ees
thro
ugh
lang
uage
, in
Dav
idso
n's
sens
e o
f "l
angu
age,
" to
the
wor
ld it
self
(199
7).
9� D
avid
son
(198
4d) f
ocus
es o
n th
e tr
ansl
atab
ilit
y o
f tru
e se
nten
ces.
His
om
itti
ng
"tru
e" h
ere
seem
s a
mer
e li
ngua
laps
a.
10�
One
par
ticu
lar
com
plic
atio
n in
Dav
idso
n's
.arg
umen
t aga
inst
sce
ptic
ism
is it
s in
voki
ng th
e po
ssib
ilit
y o
f an
om
nisc
ient
inte
rpre
ter.
For
the
clas
sic
resp
onse
to
this
invo
cati
on, s
ee F
oley
and
Fum
erto
n (1
985)
.. 11
� In
fac
t, th
is e
x:ce
rpt
is t
aken
fro
m a
pas
sage
ant
icip
atin
g D
avid
son'
s ar
guin
g fr
om P
C a
ndth
epos
sibi
lity
of a
n om
nisc
ient
inte
rpre
ter t
o th
e co
nclu
sion
that
sk
epti
cism
is u
nten
able
. 12
Q
lne
mig
ht o
bjec
t tha
t thi
s is
an
unch
arit
able
con
stru
al o
f "a
s fa
r as
poss
ible
."�
Non
ethe
less
, it i
s un
clea
r w
hat e
lse
Dav
idso
n m
ight
mea
n by
the
phr
ase.
� 13
I
than
k W
ayne
Dav
is,
Mic
hael
Fer
ry,a
nd
sev
eral
ano
ny
mo
us
revi
ewer
s fo
r�
sugg
esti
ons
and
som
e o
f th
e ob
ject
ions
con
side
red
abov
e.
Ref
eren
ces
Chi
ld, W
illi
am
1994
C
ausa
lity,
Int
erpr
etat
ion
and
the
Min
d N
ew Y
ork:
Oxf
ord
Uni
ver
sity
Pre
ss.
Dav
idso
n, D
onal
d 19
70
"Men
tal E
vent
s."
In E
xper
ienc
e an
d T
heor
y. E
dite
d bu
Law
renc
e F
oste
r an
d J
.. W
. S
wan
son.
Am
bers
t, M
A:U
nive
rsit
y o
f M
assa
ch
uset
ts P
ress
, pp
. 79
-101
. R
epri
nted
in
Dav
idso
n 19
80, p
p. 2
07
24.
1980
E
ssay
s on
Act
ions
and
Eve
nts.
New
Yor
k: C
lare
ndon
Pre
ss.
1984
a�
"Tru
th a
nd M
eani
ng."
In
Inqu
irie
s in
to T
ruth
and
Inte
rpre
tati
on.
New
Yor
k: C
lare
ndon
Pre
ss,
pp.1
7-36
. O
rigi
nall
y pu
blis
hed
in
Syn
thes
e, 1
7 (1
967)
: ·30
4-23
. 19
84b�
"R
adic
al I
nter
pret
atio
n."
In I
nqui
ries
int
o T
ruth
an
d I
nter
pret
atio
n. N
ew Y
ork:
Cla
rend
on P
ress
, pp.
125
-40.
Ori
gina
lly
publ
ishe
d in
Dia
lect
ica,
27
(197
3):
313-
28.
1984
c�
"Bel
ief
and
the
Bas
is o
f M
eani
ng."
In
Inq
uiri
es i
nto
Tru
th a
nd
Inte
rpre
tati
on. N
ew Y
ork:
Cla
rend
on P
ress
, pp.
141
-54.
Ori
gina
lly
publ
ishe
d in
Syn
thes
e,27
(19
74):
309-
23.
1984
d�
"On
the
Ver
y Id
ea o
f a
Con
cept
ual
Sche
me.
" In
Inq
uiri
es i
nto
Tru
th a
nd In
terp
reta
tion
. N
ew Y
ork:
Cla
rend
on P
ress
, pp.
183
-98.
O
rigi
nall
y pu
blis
hed
in P
roce
edin
gs a
ndA
ddre
sses
oft
he A
mer
ican
P
hilo
soph
ical
Ass
ocia
tion
, 47
(197
4):
5-20
.
Pri
ncip
le o
f Cha
rity
68
3
1984
e�
"Tho
ught
and
Tal
k."
In I
nqui
ries
int
o T
ruth
and
Int
erpr
etat
ion.
N
ew Y
ork:
Cla
rend
on P
ress
, pp.
155
-70.
19
84f�
"T
he M
etho
d o
f Tru
th in
Met
aphy
sics
." I
nIn
quir
ies
into
Tru
th a
nd
Inte
rpre
tati
on.
New
Yor
k: C
lare
ndon
Pre
ss,
pp.
199-
214.
Ori
gi
nall
y pu
blis
hed
in M
idw
est
Stud
ies
in P
hilo
soph
y, V
ol.
2: S
tudi
es in
th
e P
hilo
soph
y o
f Lan
guag
e. E
dite
d by
Pet
er A
. F
renc
h, T
heod
ore
E.
Ueh
ling
, Jr
., an
d H
owar
d K
. W
etts
tein
. M
orri
s, M
N:
Uni
ver
sity
of
Min
neso
ta P
ress
, 19
77, p
p. 2
44-5
4.
1984
g In
quir
ies
into
Tru
th a
nd In
terp
reta
tion
. New
Yor
k: C
lare
ndon
Pre
ss.
1986
� "A
Nic
e D
eran
gem
ent
of
Epi
taph
s."
In T
ruth
and
Int
erpr
etat
ion:
P
ersp
ecti
ves
on
the
Phi
loso
phy
of
Don
ald
Dav
idso
n.
Edi
ted
by
Ern
est
Lep
ore.
New
Yor
k: B
lack
wel
l, pp
. 43
3-46
. 19
88�
"The
Myt
h o
f th
e Su
bjec
tive.
" In
Bew
usts
ein,
Sp
rach
e un
d di
e K
unst
. E
dite
d by
M.
Ben
edik
t an
d R
. B
urge
r. V
ienn
a: E
diti
on
S. V
erla
g de
r O
ster
reic
hisc
hen
Sta
atsd
ruck
erei
, pp.
45-
54.
1990
"T
he S
truc
ture
an
d C
onte
nt o
f T
ruth
." Jo
urna
l of P
hilo
soph
y, 8
7:
279-
328.
19
91
"Epi
stem
olog
y E
xter
nali
zed.
" D
iale
ctic
a, 4
5: 1
91-2
02.
Rep
rint
ed
in D
avid
son
2001
a, p
p. 1
93-2
04.
1997
"S
eein
g th
roug
h L
angu
age.
" In
Tho
ught
and
Lan
guag
e. E
dite
d by
Jo
hn P
rest
on. N
ew Y
ork:
Cam
brid
ge U
nive
rsit
y Pr
ess,
pp.
15-
22.
2001
a Su
bjec
tive,
Int
ersu
bjec
tive,
Obj
ectiv
e. N
ew Y
ork:
Oxf
ord
Uni
vers
ity
Pres
s.
2001
b�
"A C
oher
ence
The
ory
of
Tru
th a
nd K
now
ledg
e."
In S
ubje
ctiv
e,
Inte
rsub
ject
ive,
O
bjec
tive.
N
ew Y
ork:
O
xfor
d U
nive
rsit
y Pr
ess,
pp
. 13
7-53
. O
rigi
nall
y pu
blis
hed
in K
ant
oder
Heg
el.
Edi
ted
by
D.
Hei
nric
h. S
tutt
gart
: K
lett
-Cot
ta,
1983
, pp.
423
-38.
20
04
Pro
blem
s o
f Rat
iona
lity
. N
ew Y
ork:
Oxf
ord
Uni
vers
ity
Pres
s.
For
thco
min
g Tr
uth,
Lan
guag
e an
d H
isto
ry.
New
Yor
k: O
xfor
d U
nive
rsit
y Pr
ess.
E
vnin
e, S
imon
19
91
Don
ald
Dav
idso
n. S
tanf
ord,
CA
: S
tanf
ord
Uni
vers
ity
Pres
s.
Fole
y, R
icha
rd,
and
Ric
hard
Fum
erto
n 19
85
"Dav
idso
n's
The
ism
?" P
hilo
soph
ical
Stu
dies
, 48
: 83
-90.
M
alpa
s, J
effr
ey E
. 19
92
Don
ald
Dav
idso
n an
d th
e M
irro
r o
f Mea
ning
: H
olis
m,
Tru
th,
Inte
rpr
etat
ion.
New
Yor
k: C
ambr
idge
Uni
vers
ity
Pres
s.
Ram
berg
, B
jern
T.
1989
D
onal
dD
avid
son'
sP
hilo
soph
y o
fLan
guag
e:A
nIn
trod
ucti
on O
xfor
d:
Bas
il B
lack
wel
l.