Session 14 Natural Language:
Semantics and Parsing
JACK AND JANET IN SEARCH OF A THEORY OF KNOWLEDGE
Eurene Charniak
A r t i f i c i a l I n t e l l i g e n c e Laboratory
Massachusetts I n s t i t u t e of Technology
Abstract
w o r l d such as:
In o r d e r to answer q u e s t i o n s about c h i l d r e n ' s
s t o r i e s one needs a g r e a t deal of "common
sense" knowledge. A model Is presented which
gives a rough o r g a n i z a t i o n to t h i s knowledge
along w i t h s p e c i f i c a t i o n s as to how the
I n f o r m a t i o n will be accessed. This rough
model Is then used as a basis f o r
tight
arguments about narrow Issues ( p r i m a r i l y
using examples concerning piggy banks.) The
paper Is intended as an i l l u s t r a t i o n of how
one might go about c o n s t r u c t i n g a theory of
knowledge.
Acknowledgements
This paper is based on p o r t i o n s of an KIT
Ph.D. t h e s i s submitted to the department of
E l e c t r i c a l Engineering.
The t h e s i s Is
reproduced as Al Technical R e p o r t , 2 6 6 , As In
my t h e s i s I would l i k e to thank a l l the
people at the MIT A r t i f i c i a l I n t e l l i g e n c e
Laboratory who l i s t e n e d to and argued w i t h me
on many o c c a s i o n s .
The work r e p o r t e d h e r e i n V/PS conducted
at the A r t i f i c i a l I n t e l l i g e n c e L a b o r a t o r y , a
Massachusetts I n s t i t u t e of Technology
research program supported in p a r t by the
Advanced Research P r o j e c t s Agency of the
Department of Defense and monitored by the
O f f i c e of Naval Research under C o n t r a c t
Number
N00014-70-A-0362-0003.
1
Introduction
Let us consider the problem of
c o n s t r u c t i n g an a b s t r a c t model of s t o r y
comprehension. To determine what the model,
or program, has " u n d e r s t o o d " about what it
has r e a d , we w i l l ask It q u e s t i o n s . So a
t y p i c a l s t o r y might s t a r t :
(1)
Janet needed some money. She got
her piggybank (PB) and s t a r t e d to
shake it. F i n a l l y some money came
out.
Some t y p i c a l q u e s t i o n s would be:
(2)
(5)
(4)
Why d i d Janet get the PC?
Did Janet get the money?
Why was the PB shaken?
Questions (2) - (4) are not answered
e x p l i c i t l y I n the t e x t .
That I s , the s t o r y
d i d not say " J a n e t got her PS because she
. . . " The s t o r y does not even c o n t a i n a f u l l
i m p l i c i t answer; one cannot l o g i c a l l y deduce
an answer from the statements in the s t o r y
w i t h o u t u s i n g general knowledge about the
(5)
(G)
(7)
One can o f t e n get money from PCs.
The hard p a r t of g e t t i n g money from
a PR is g e t t i n g It o u t . Once t h a t
Is done one can be s a i d to have the
money.
Shaking helps set money out of a PB.
So In order to understand a c h i l d r e n ' s
s t o r y we need a theory of every day
knowledge.
T h i s t h e o r y would have to answer
q u e s t i o n s l i k e "What is the knowledge we
have?" and "How Is It o r g a n i z e d so we can get
at the necessary I n f o r m a t i o n when It Is
needed?"
Note t h a t t h i s l a t t e r quest Ion
assumes t h a t we have some s p e c i f i c task or
tasks In mind, in our case answering
q u e s t i o n s about c h i l d r e n ' s s t o r i e s .
The r e s t of t h i s paper d i v i d e s i n t o two
parts.
In the f i r s t p a r t a rough d e s c r i p t i o n
of a model of c h i l d r e n ' s s t o r y comprehension
w i l l be p r e s e n t e d .
In the second s e c t i o n we
w i l l assume the model presented In the f i r s t
and look at some narrow q u e s t i o n s concerning
the o r g a n i z a t i o n and content of our knowledge
about piggy banks,
2
A Model of C h i l d r e n ' s Story Comprehension,
The model presented here Is s o l e l y
concerned w i t h deduction and does not
consider problems of n a t u r a l language per se.
In p a r t i c u l a r It does not deal w i t h syntax or
those problems on the boundary between syntax
and d e d u c t i o n l i k e d i s a m b i g u a t i o n of word
senses and d e t e r m i n a t i o n of noun phrase
referents.
(However, my Ph.D. t h e s i s
c o n s i d e r s the noun phrase problem in some
detail.)
So we w i l l assume t h a t as the s t o r y
comes i n t o the program It is immediately
t r a n s l a t e d I n t o an i n t e r n a l r e p r e s e n t a t i o n
which Is convenient f o r doing d e d u c t i o n .
The
i n t e r n a l r e p r e s e n t a t i o n w i l l b e a group o f
" a s s e r t i o n s " each a s s e r t i o n being a p r e d i c a t e
on an a r b i t r a r y number of arguments.
Putting
an a s s e r t i o n i n t o the data base Is to
"assert" i t .
The model w i l l t r y to " f i l l in
the b l a n k s " of the s t o r y on a l i n e by l i n e
basis.
That I s , a s I t goes a l o n g . I t w i l l
t r y to make connections between events in the
s t o r y ( u s u a l l y causal connect ions) and f i l l
In m i s s i n g f a c t s which seem important such as
J a n e t ' s now having the money in ( 1 ) .
2.1
Demons and Base Routines
Consider a f a c t l i k e :
(8)
I f " I t ( s Cor w i l l be) r a i n i n g " and
If "person P Is o u t s i d e "
then *'P wi 11 get w e t "
body (an a r b i t r a r y program). We w i l l execute
the body of the f a c t o n l y when an a s s e r t i o n
Is In the data base which matches the
pattern.
(We w i l l a l s o say t h a t the
a s s e r t i o n " e x c i t e s " the f a c t . )
I n (8) the
p a t t e r n would be "someone o u t s i d e . " Then In
(11) when we i n t r o d u c e (8) no a s s e r t i o n
matches the p a t t e r n .
But the next l i n e
c r e a t e s a matching a s s e r t i o n , so the f a c t
w i l l be e x c i t e d . We w i l l say t h a t a f a c t is
" l o o k i n g f o r w a r d " when I t s t o p i c concept
appears b e f o r e the a s s e r t i o n which matches
the p a t t e r n .
When the a s s e r t i o n which
matches the p a t t e r n comes f i r s t we w i l l say
t h a t the f a c t I s " l o o k i n g backward" (as i n
10).
We can see how Important l o o k i n g f o r w a r d
Is w i t h a few examples.
We have an I n t u i t i v e b e l i e f t h a t (8) Is a
f a c t about " r a i n " , r a t h e r t h a n , say, a f a c t
about " o u t s i d e . " Many t h i n g s happen o u t s i d e
and g e t t i n g wet Is o n l y one of them.
On the
o t h e r hand o n l y a l i m i t e d number of t h i n g s
happen when It r a i n s .
We w i l l enbody t h i s b e l i e f In our system
b y a s s o c i a t i n g (8) w i t h " r a i n " s o t h a t o n l y
when " r a i n " comes UP In the s t o r y w i l l we
even consider u s i n g r u l e ( 8 ) . We w i l l say
t h a t r a i n Is the " t o p i c , c o n c e p t " of ( 8 ) .
To
put t h i s another way, when a concept Is
brought up In a s t o r y , the f a c t s a s s o c i a t e d
w i t h it are "made, a v a i l a b l e " f o r use In
making d e d u c t i o n s .
(We w i l l a l s o say t h a t
the f a c t s are "out in" or " a s s e r t . " )
So,
if " c i r c u s " , say, has never come up, the
program w i l l not be able to make deductions
u s i n g those f a c t s a s s o c i a t e d o n l y w i t h
"circus."
Mote however t h a t we are not s a y i n g that
" r a i n " has to be mentioned e x p l i c i t l y in the
s t o r y b e f o r e we can use ( 8 ) .
It is o n l y
necessary t h a t t h e r e be a " r a i n " a s s e r t i o n
put i n t o the data base.
Other p a r t s of the
s t o r y may p r o v i d e f a c t s which cause the
program to a s s e r t t h a t it is r a i n i n g .
For
example:
(12)
"Janet was t h i n k i n g of g e t t i n g Jack
a ball for his birthday.
When she
t o l d Penny, Penny s a i d , ' D o n ' t do
t h a t . Jack has a b a l l . ' " Here we
I n t e r p r e t e d the l i n e "Jack has a
b a l l " as meaning t h a t he d i d not
want a n o t h e r . The common sense
knowledge Is the f a c t t h a t In many
cases having an X means t h a t one
w i l l not want another X.
T h i s piece
of i n f o r m a t i o n would probably be
f i l e d under " t h i n g s t o c o n s i d e r when
about to get something f o r somebody
else."
N a t u r a l l y I t was a n e a r l i e r
l i n e which mentioned t h a t Janet was
t h i n k i n g o f g e t t i n g Jack a b a l l .
(13)
" B i l l o f f e r e d t o t r a d e h i s pocket
k n i f e f o r J a c k ' s dog T i p .
Jack s a i d
') w i l l ask J a n e t ,
T i p Is her dog
t o o . ' " The l a s t l i n e I s I n t e r p r e t e d
as the reason Jack w i l l ask J a n e t .
T h i s r e q u i r e s i n f o r m a t i o n about the
r e l a t i o n between t r a d i n g and
ownership.
(14)
" J a n e t wanted to get some money.
She found her piggy bank and s t a r t e d
to shake I t .
She d i d n ' t hear
a n y t h i n g . " The l a s t l i n e means t h a t
t h e r e was n o t h i n g In the piggybank
on the b a s i s of f a c t s about
plggybanks.
In each of these cases it Is an e a r l i e r l i n e
which c o n t a i n s the I n f o r m a t i o n which is used
to a s s i g n the i n t e r p r e t a t i o n . So in (12)
t h e r e i s n o t h i n g I n h e r e n t I n the l i n e "Jack
has a b a l l " which means " d o n ' t g e t him
another."
I f t h e r e were, something I n the
l i n e would a l s o have to key a check f o r the
following situations:
338
(15)
B i l l and Dick wanted to p l a y
b a s e b a l l . When Jack came by B i l l
s a i d "There Is J a c k . He has a
ball."
(16)
Tom asked h i s f a t h e r If he would buy
him a b a l l .
"Jack has a b a l l , " s a i d
Tom.
(17)
B i l l ' s b a l l o f s t r i n g was stuck i n
the t r e e . He asked Jane how he
c o u l d set it o u t .
Jane s a i d "You
should h i t I t w i t h something.
Here
comes Jack. Me has a b a l l . "
...?"
A good t e s t as to whether a g i v e n f a c t
should be p a r t of a base r o u t i n e or a demon
is whether we need several l i n e s to set it up
or whether we can I l l u s t r a t e the f a c t by
presenting a single l i n e .
( n a t u r a l l y several
l i n e s c o u l d be made I n t o one by p u t t i n g
" a n d ' s " between them, but t h i s Is dodging the
point.
I am o n l y suggesting an I n t u i t i v e
t e s t . ) So we saw t h a t "Jack has a b a l l " was
not enough by I t s e l f to t e l l us t h a t Jack
does not want another b a l l .
Hence t h i s
r e l a t i o n Is embodied by a demon, not a base
r o u t i n e . On the o t h e r hand, o f t e n a s i n g l e
l i n e can t e l l us q u i t e a b i t as In "Jack was
on second base." This i n d i c a t e s t h a t the
base r o u t i n e f o r "second base" can o f t e n t e l l
us t h a t we are t a l k i n g about a baseball game.
Those f a m i l i a r w i t h Planner might n o t i c e
t h a t our " f a c t s " look q u i t e s i m i l l a r t o
Planner antecedent theorems, w i t h the
e x c e p t i o n t h a t our f a c t s can " l o o k hack" as
well as " l o o k f o r w a r d . " Antecedent theorems
are o n l y designed to look f o r w a r d .
I
I n i t i a l l y f o r m u l a t e d f a c t s a s antecedent
theorems because 1 was so Impressed w i t h the
need to " l o o k f o r w a r d . " However, r a t h e r then
c a l l the f a c t s antecedent theorems, I c a l l
them "demons" s i n c e It is a s h o r t e r name.
S p e c i f i c a t i o n and Removal of Demons.
It should be emphasized t h a t the model does
not " l e a r n " the i n f o r m a t I o n c o n t a i n e d in the
demons.
This i n f o r m a t i o n is put In by the
model maker. On the o t h e r hand, the demons
are not s p e c i f i c to the s t o r y in the sense
t h a t they mention Jack, o r " t h e red b a l l . "
Rather, they t a l k about "a person X" who at
one p o i n t in the s t o r y could he Jack, at
a n o t h e r . B i l l . We w i l l assume a mechanism
f o r b i n d i n g some of the v a r i a b l e s of the
demon ( " s p i c i f y i n g " the demon) at the time
the demon Is a s s e r t e d .
We want demons to be a c t i v e o n l y w h i l e
they are r e l e v a n t to the s t o r y ,
A s t o r y nay
s t a r t by t a l k i n g about g e t t i n g a present f o r
Jack, but u l t i m a t e l y r e v o l v e around the games
played at h i s p a r t y . We w i l l need some way
to remove the " p r e s e n t g e t t i n g " demons when
they have o u t l i v e d t h e i r u s e f u l n e s s . (An
i r r e l e v a n t but a c t i v e demon not o n l y wastes
time and space, but can cause us to
m i s i n t e r p r e t a new l i n e . ) As a f i r s t
a p p r o x i m a t i o n we w i l l assume t h a t a demon is
d e c l a r e d I r r e l e v a n t a f t e r a g i v e n numer of
l i n e s have gone by.
2.2
BooKKeeping
and
Fact Finders
Undating. and. Bookkeeping.
Up to t h i s
p o i n t we have Introduced two p a r t s of the
model, demons and base r o u t i n e s .
In t h i s
s e c t i o n we w i l l i n t r o d u c e the remaining two
parts.
Again l e t us consider the s i t u a t i o n when
Jack had a b a l l , B i l l a t o p , and they t r a d e d .
When we say t h a t B i l l now has the b a l l . It
i m p l i e s t h a t Jack no longer does.
That is to
say, we must somehow remove the f a c t t h a t
Jack has the b a l l from the data base.
A c t u a l l y we d o n ' t want to remove it, since we
may be asked "Who had the b a l l before P i l l
d i d . " I n s t e a d , we want to mark the
a s s e r t i o n in some way to i n d i c a t e t h a t It has
been updated. We will assume t h a t t h e r e is a
separate s e c t i o n , p r e t t y much Independent of
the r e s t of the model, which is r e s p o n s i b l e
f o r doing such u p d a t i n g . W e w i l l c a l l t h i s
section "bookkeeping."
Fact F i n d e r s .
But even d e c i d i n g t h a t
one statement updates another r e q u i r e s
s p e c i a l knowledge. Suppose we have;
Base Rout i n e s .
So f a r we have s a i d
t h a t demons are a s s e r t e d when the proper
concept has been mentioned.
But t h i s Implies
t h a t t h e r e is something attached to the
concept name t e l l i n g us what demons should be
put in,
if we look at a p a r t i c u l a r example, say
( 1 3 ) , i t i s B i l l ' s o f f e r t o t r a d e which sets
up the c o n t e x t f o r the r e s t of the fragment.
I w i l l assume t h a t the I n f o r m a t i o n to do so
is in the form of a program.
Such r o u t i n e s ,
which are a v a i l a b l e to set up demons, w i l l be
c a l l e d "base r o u t i n e s . "
These base r o u t i n e s w i l l he r e s p o n s i b l e
f o r more than s e t t i n g up demons. Suppose we
are t o l d t h a t Jack had a b a l l , and B i l l a
top.
Then Jack traded h i s b a l l t o B i l l f o r
the t o p . One q u e s t i o n we might ask is "Who
now has the t o p ? " N a t u r a l l y s i n c e q u e s t i o n s
of "who has w h a t " are Important in
u n d e r s t a n d i n g s t o r i e s we w i l l want to keep
tabs on such I n f o r m a t i o n . in t h i s p a r t i c u l a r
case, it must again be the " t r a d e " statement
which t e l l s us to s w i t c h possession of the
objects.
Every time a t r a d e occurs we w i l l
want to exchange o b j e c t s , so whenever we see
" t r a d e " we execute the " t r a d e " base r o u t i n e .
Of c o u r s e , the program c a n ' t be too s i m p l e minded, s i n c e I t must also handle " I w i l l
t r a d e . . . " and perhaps even " W i l l you t r a d e
(18)
Jack was In the house.
Sometime
l a t e r he was at the s t o r e .
If we ask " I s Jack In the house?" we want to
answer "No, he Is at the s t o r e . " Cut how is
bookkeeping going to f i g u r e t h i s out?
There
is a simple r u l e which says t h a t ( < s t a t e > A
B) updates ( < s t a t e > A C) where C Is not the
same as B. So (AT JACK FARM) would update
(AT JACK NEW-YORK). But in (18) we c a n ' t
simply look f o r Jack AT <someplace which is
not the s t o r e > , s i n c e he is in the house. To
make t h i n g s even worse, we c o u l d have:
(19)
Jack was in the house.
Sometime
l a t e r he was In the k i t c h e n .
To s o l v e t h i s problem we w i l l need:
339
340
p r o b a b l y more c o m p l e x t h a n we have i n d i c a t e d
so f a r .
Consider:
(26)
A
J a n e t was g o i n g t o buy some c a n d y .
She was a l s o g o i n g to buy some
fruit.
In (26) both occurences of "buy" w i l l
a c t i v a t e BUY-NEED-MONEY demons.
( T h o u g h we
d i d n o t comment o n t h i s e a r l i e r , t h e I d e a o f
s p e c i f y i n g demons a s m e n t i o n e d i n 2 . 1
o b v i o u s l y r e q u i r e s s e p a r a t e c o p i e s o f a demon
to be able to e x i s t s i m u l t a n e o u s l y . )
H o w e v e r , ( 2 6 ) does n o t I m p l y t h a t J a n e t
r e a l l y needs money.
For a l l w e know she has
as much as she needs In h e r p o c k e t .
If
demon-demon I n t e r a c t i o n w e r e a s s i m p l e a s w e
have made i t o u t t o b e , t h e two i n s t a n c e s o f
BUY-NEED-MONEY w o u l d j o i n u p t o p r o d u c e a
"need money" a s s e r t i o n .
So It Is not
s u f f i c i e n t f o r two demons t o b e l o o k i n g f o r
t h e same p a t t e r n .
L o o k i n g a t e x a m p l e ( 2 5 ) w e n o t e t h a t one
o f t h e demons c a v e a r e a s o n why J a n e t
might
need money, and t h e s e c o n d s u g g e s t e d t h a t
n e e d i n g money w a s t h e c a u s e o f a c e r t a i n
action.
So we h a v e :
Will
buy - - > Need money - - > W i l l
get
P u t t i n g Money
Penny s a i d t o J a n e t , " D o n ' t t a k e
y o u r money w i t h y o u t o t h e p a r k .
(You w i l l l o s e I t . )
G o and g e t y o u r
PB!"
(28)
A f t e r J a n e t h e l p e d Ms. Jones w i t h
h e r g r o c e r i e s Ms. Jones gave h e r a
dime.
J a c k came a l o n g and s a i d .
"Cone w i t h m e t o t h e p a r k , J a n e t . "
"OK," said Janet.
"Hut f i r s t I an
g o i n g home to f i n d my PB,
I do n o t
w a n t t o t a k e t h e money t o t h e p a r k . "
(29)
Janet
Mother
there
Let's
p u t some money o n t h e s i n k ,
s a i d , " I f you l e a v e t h e money
i t may f a l l I n t h e d r a i n .
f i n d your PB."
(30)
Janet s a i d , "l am g o i n g to put my
money away.
I w i l l g e t my P B . "
(31)
J a n e t h e l p e d Ms. Jones w i t h her
groceries.
Ms.
Jones gave J a n e t a
dime.
J a c k came a l o n g and s a i d ,
" J a n e t , l e t ' s go to the p a r k . "
"OK," said Janet.
" B u t i want t o
p u t my money In a s a f e p l a c e .
I am
g o i n g to get my PC."
it
flow t h e r e i s n o t h i n g s a y i n g t h a t o u r demon
needs t o a c c o u n t f o r ( 3 0 ) and ( 3 1 ) .
However,
i t seems q u i t e o b v i o u s t h a t w e a r e u s i n g t h e
same i n f o r m a t i o n I n a i l t h e e x a m p l e s a b o v e ,
The o n l y d i f f e r e n c e I s t h a t I n ( 2 7 ) - ( 2 9 ) w e
a r e e x p r e s s i n g t h e need f o r a " s a f e p l a c e " b y
m a k i n g n e g a t i v e comments a b o u t a n o t h e r
location.
If t h i s Is a s i n g l e f a c t we would
l i k e a s i n g l e demon t o e x p r e s s I t .
The
t r o u b l e i s f i n d i n g what ( 2 7 ) - ( 3 1 ) have I n
common.
Capturein Generalizations Before
moving on I s h o u l d p o i n t out t h a t the k i n d of
argument used i n t h i s s e c t i o n (and the n e x t
a l s o ) Is a " c a p t u r e the g e n e r a l i z a t i o n " type
a r g u m e n t commonly f o u n d I n l i n g u i s t i c s .
We
c o u l d have c r e a t e d a new demon t o e x p l a i n
(25).
I t w o u l d have s a i d , " I f a p e r s o n g e t s
h i s P B l o o k f o r him p l a n n i n g t o buy
something."
However, t h i s would be m i s s i n g
t h e g e n e r a l i z a t i o n t h a t " m o t i v e s " and
" r e s u l t s " a l w a y s a c t t h i s v/ay.
S o f a r I have
o n l y g i v e n one e x a m p l e to' s u p p o r t t h e " d e m o n demon i n t e r a c t i o n g e n e r a l i z a t i o n " , b u t i n I n
t h e n e x t s e c t i o n w e w i l l see a n o t h e r .
3.2
(27)
I n each c a s e t h e n a t u r a l q u e s t i o n i s ,
"Why s h o u l d J a n e t g e t h e r PB?"
Now we m i g h t
t r y t o c o n s t r u c t a " p i g g y b a n k " demon w h i c h
r e s p o n d s t o some common e l e m e n t i n ( 2 7 ) ( 2 9 ) and t h e n make t h e n e c e s s a r y a s s e r t i o n s .
A c l o s e l o o k at the examples even g i v e s a
s t a r t a t w h a t such a common e l e m e n t m i g h t b e .
s a y " a p a r t i c u l a r l o c a t i o n f o r t h e money i s
negatively evaluated."
We w i l l c a l l t h i s
demon PB-BAD-PLACE. The t r o u b l e w i t h such a
s o l u t i o n would be t h a t It would not account
for:
PR
To put t h i s in everyday terms, in (25) we
have b o t h a m o t i v e f o r n e e d i n g money
( b u y i n g ) , and a r e s u l t o f n e e d i n g t h e money
(go and g e t P B ) .
I n ( 2 6 ) w e have two
motives.
The n a t u r a l s u g g e s t i o n i s t h a t
demon-demon i n t e r a c t i o n b e r e s t r i c t e d t o
c a s e s w h e r e w e have b o t h m o t i v e and r e s u l t .
How do we r e c o g n i z e when we have b o t h
m o t i v e and r e s u l t ?
A s I t s t a n d s now one
demon l o o k s p r e t t y much l i k e any o t h e r .
We
m i g h t j u s t t r y t o l a b e l a l l demons a s
"motive" or " r e s u l t " w i t h respect to t h e i r
pattern.
O n t h e o t h e r hand I t m i g h t b e
p o s s i b l e t o d e r i v e " m o t i v e " and " r e s u l t " f r o m
more b a s i c c o n s i d e r a t i o n s .
At any r a t e ,
seems p r e m a t u r e t o f o r m a l i z e s u c h c o n c e p t s a t
this point.
W e s i m p l y d o n ' t know e n o u g h .
PIGGY
Bank P r o b l e m .
One f a c t we know
a b o u t P g ' s i s t h a t t h e y a r e good p l a c e s t o
keep money.
T h i s f a c t seems t o come I n t o
play In:
A Non-PJfrsv Bank P r o b l e m .
in
c o u r s e of l o o k i n g at examples l i k e
( 3 1 ) 1 n o t e d examples l i k e ;
(32)
Penny s a i d t o J a n e t , " D o n ' t t a k e
y o u r money w i t h y o u t o t h e p a r k .
Put I t o n t h e s h e l f . "
(33)
A f t e r J a n e t h e l p e d Ms. Jones w i t h
h e r g r o c e r i e s Ms. Jones gave her a
dime.
J a c k came a l o n g and s a i d
"Come w i t h m e t o t h e p a r k , J a n e t . "
"OK," said Janet.
"But f i r s t I am
g o i n g t o p u t m y money i n t h e h o u s e ,
I d o n o t w a n t t o t a k e t h e money t o
the p a r k . "
i n t o a P i g g y Bank
In t h i s s e c t i o n we w i l l look at a
p o s s i b l e demon a s s o c i a t e d w i t h p i g g y banks
and a r g u e t h a t t h e d e d u c t i o n f t w o u l d a c c o u n t
f o r can b e b e t t e r h a n d l e d b y demon-demon
i n t e r a c t i o n b e t w e e n two o t h e r demons.
In
e f f e c t we w i l l be t r y i n g to d e t e r m i n e , on an
e x t r e m e l y s m a l l s c a l e , w h a t p e o p l e know.
341
the
(27) -
reason he Is g e t t i n g the PB Is to put It I n . "
( A c t u a l l y t h i s theorem Is t r u e of a wide
c l a s s of c o n t a i n e r s , but t h a t does not a f f e c t
the argument at hand,)
This demon w i l l
account f o r examples t i k e :
(34)
Janet put some money on the s i n k .
Mother s a i d , " I f you leave the money
t h e r e I t may f a l l I n the d r a i n . "
Janet put the money In a drawer.
(35)
Janet s a i d "I am going to put my
money away.
I will put it in my toy
box."
(40)
Ms. Jones gave Janet a dime.
Janet
went to get her PB.
"I want the
money to be In my PB," she t h o u g h t .
(36)
Janet helped Ms. Jones w i t h her
g r o c e r i e s . Ms. Jones gave Janet a
dime.
Jack came along and s a i d
" J a n e t , l e t ' s g o t o the p a r k . "
"OK," s a i d J a n e t .
"But I want to
put my money In a safe p l a c e , " Then
Janet went I n t o ' the house and put
the money In her room.
(41)
Janet got her PB and dropped some
money I n .
(42)
A f t e r Ms Jones gave Janet a dime,
Jack came by and asked Janet If she
wanted to go to the p a r k .
"OK,"
said Janet.
" I w i l l g o home f i r s t
and get my PB." Soon Janet came
back and s a i d "My money is In the
PB, l e t ' s g o ! "
These examples e x a c t l y m i r r o r (27) - ( 3 1 ) ,
except t h a t (32) - (3E) d o n ' t mention PB's.
N a t u r a l l y , In these examples the q u e s t i o n to
ask Is "Why d i d Janet put the money In the
d r a w e r ? " , " I n the house?", e t c .
Such examples tend to I n d i c a t e t h a t the
problem f a c i n g us Is wider t h a t j u s t PB's.
We w i l l name t h i s wider problem the " p u t
away" problem.
However It Is not the case
t h a t our problem w i t h PB's can be c o m p l e t e l y
reduced to the " p u t away" problem.
So w h i l e
In the non-piggy bank examples wo mention
t h a t Janet has or a c t u a l l y Intends to " p u t "
the money some p l a c e . In the PB examples a l l
we needed to say was t h a t Janet was going to
get the PB.
To put t h i s another way, our
knowledge of PC's allowed us to I n t e r p r e t
" g e t PB" as meaning t h a t Janet was g o i n g to
put money i n t o it. However our knowledge of
houses or shelves does not a l l o w us to make
s i m i l a r deductions In (32) - ( 3 6 ) .
Demo-Demon
Interacion.Now, If we
assume demon-demon i n t e r a c t i o n as discussed
In s e c t i o n 3 . 1 , PB-MONEY-If plus PUT-AWAY
w i l l I n t e r a c t t o s o l v e a l l the examples from
(27) to ( 3 1 ) . Let us see how t h i s w i l l
happen.
F i r s t note t h a t the r e s t r i c t i o n s w e
placed on demon-demon i n t e r a c t i o n s are met
here.
F i r s t both demons have the same
p a t t e r n , e . g . , "money is in PB." ( A c t u a l l y
the p a t t e r n f o r PUT-AWAY Is " < o b j e c t > is In
appropriate location)"
however <object>
w i l l be bound to the money at the time the
demon is a s s e r t e d , and a p p r o p r i a t e l o c a t i o n )
w i l l match PB when the demon is e x c i t e d . )
Secondly, we need both a m o t i v e and a r e s u l t
before we can "combine" demons.
In the case
at hand, PUT-AWAY Is a motive f o r having the
money In the PB, and " g e t PB" Is a r e s u l t of
i n t e n d i n g to put money In the PB.
The put-Away Demon.
I g n o r i n g piggy
banks f o r the moment, what would a s o l u t i o n
to (32) - (36) look l i k e ? We w i l l have some
demon, cal1ed the PUT-AWAY demon, which Is
activated by lines l i k e :
(37)
(38)
(39)
Saving Money.
F i n a l l y , note t h a t our
s o l u t i o n extends t o the f o l l o w i n g case:
(U3)
D o n ' t leave the money by the s i n k .
1 do not want to take my money to
the p a r k .
I w i l l put my money away.
Janet got a dime from Ms. Jones.
She s a i d "I am s a v i n g my money to
buy a b i c y c l e .
I am going home to
get my PB."
Here we know t h a t Janet Is g o i n g to put the
money In the PB because of the " s a v e "
statement.
However, we Immedlately note t h a t
we have cases 1i ke:
These l i n e s will put In a demon l o o k i n g f o r
" p u t away" and the demon w i l l a s s e r t t h a t the
reason f o r p u t t i n g the t h i n g away Is (37) (39).
U l t i m a t e l y we w i l l want a theory of
why people put t h i n g s away ( i . e . , what l i n e s
put In the " p u t away" demon), and how to
determine what c o n s t i t u t e s a "put away"
1 ocat t o n .
However, (32) - (36) c l e a r l y show
t h a t the problem Is d i s t i n c t from the
q u e s t i o n of what we know about PBs.
C4I4)
Janet got a dime from Ms. Jones.
Janet told her "I am s a v i n g my money
to buy a b i c y c l e .
I am going home
to put the money away.
(I am going
home to put the money In my
drawer.)"
N a t u r a l l y , (kk) I n d i c a t e s t h a t " s a v e " must
a c t i v a t e PUT-AWAY.
If t h i s Is the case, then
(i»3) Is accounted f o r in e x a c t l y the same
manner as a l l the I n i t i a l examples.
While
the reader may not be s u r p r i s e d at t h i s
r e s u l t . I am, s i n c e I n i t i a l l y I thought t h a t
the r e l a t i o n s h i p o f " s a v e " w i t h piggy banks
would need a separate PB demon.
The Plggv Bank Demon.
Whet we w i l l now
see is t h a t if we assume the PUT-Away demon,
a l l the examples In (27) - (31) f a l l out
e a s i l y , plus a few o t h e r s which we h a v e n ' t
even looked at y e t .
But f i r s t we need to
c o n s i d e r a new PB demon e n t i t l e d PB-MONEY-IN.
I t I s p a r a l l e l t o PB-NEED-MONEY, but w h i l e
the l a t t e r was f o r r e c o g n i z i n g t h a t money was
going to be taken out of he PB, PB-MONFY-IN
Is f o r r e c o g n i z i n g t h a t money Is g o i n g to be
put I n .
I t says " I f you see t h a t the person
wants some money to be In the PB then the
342
It
Conclusion
The two halves of t h i s paper stand In
c o n t r a s t to each o t h e r .
The p r e s e n t a t i o n of
the nodel ( s e c t i o n 2) Is general ( I n theory
c o v e r i n g a l l o f c h i l d r e n ' s s t o r i e s ) , but
vague and f u l l of c o v e r t appeals to the
reader's i n t u i t i o n .
Section 3 on the o t h e r
hand is narrow, o n l y t a l k i n g about small
p o r t i o n s of our knowledge of PCs, but tightly
reasoned ( h o p e f u l l y ) .
flow by themselves the conclusions of
s e c t i o n 3 are not t h a t I m p o r t a n t .
Of c o u r s e .
If we c o u l d p i n down one hundred f a c t s the
way we pinned down one in s e c t i o n 3.2 then vie
would have the beginnings of a theory of
knowledge.
Dut 1 d i d not w r i t e t h i s paper to
t e l l of one f a c t about PCs.
Rather I view
the paper as an I l l u s t r a t i o n of how one might
go about the task of c o n s t r u c t i n g a theory of
knowledge.
343