Alvin I. Goldman (ed) - Readings in Philosophy and Cognitive Science -chap21.pdf - Stuff - asiaka88

This excerpt from

Readings in Philosophy and Cognitive Science.

Alvin I. Goldman, editor.

is provided in screen-viewable form for personal use only by members

of MIT CogNet.

Unauthorized use or dissemination of this information is expressly

forbidden.

If you have any questions about this material, please contact

cognetadmin@cognet.mit.edu.

Some

Elements

Conceptua

; l

Structure

Ray Jackendoff

Onto

logical Calms : Some Major Categories of Concepts

Let us consider the onto logical presuppositions of natural language -

what sorts of entities inhabit the world as consb"Uedand are capable of

being referred to by linguistic expressions .

One circumstanceunder which a speakerclearly construesthere to

be an entity in the world is when he refers to it by means of an

expression like ( 21.1 ) .

( 21.1 ) ( That [ pointing ] is a dog ) .

In ( 21.1 ) the use of the demonstrative pronoun is accompaniedby a

gesture that servesas an invitation to the hearerto locatethe entity in

his own visual field. If the hearer cannot identify an entity of the

appropriate sort , perhaps becausehe has his eyes shut , or the conversation

is taking place over the telephone , or the speaker is pointing to

something in a blurry photograph , the intended referentis unavailable

to the heare , and discoursecannot proceed . A demonstrative pronoun

( 21.2b ),

actions ( 21.2c ), events (21 .2d ), sounds ( 21.2e ) , manners ( 21.2 ) , amounts

( 21. 2g ), and numbers

( 21.2h ) .

( 21.2 ) a. Your hat is here [ pointin] and your coatis there [ pointin] .

b. He went thataway [ pointin] .

c. Can you do that [pointin ] ?

Can you do this [demonstrating

] ?

d. That [ pointin] had better not happen in my house !

From R. Jackendoff , Consciousness

and the computational mind (198 ) . Cambridge , MA : MIT

used in this fashion hasbeencalleda " pragmaticanapho " ; it takesits

reference from nonlinguistic contex.

So far this should be fairly unsurprising . The interestariseswhen we

observe , as pointed out by Hankamerand Sag ( 197 ) , that pragmatic

anaphora is possible not only to designatedobject , asin ( 21.1 ), but also

to entities best classifiedas places ( 21.2a ), 1Nlths or trajectories

Press. Reprinted by permission .

e. That

You

shuffle

[gesturing

cards

sounds

{ like

thus

way

}

[ demonstrating

f .

this

thiS

g . The fish that got away was that [ demonstratin ] long .

yay

h . Please bring back this many cookies [ holding up somenumber of

fingers ] .

The conditions on the interpretation of that in ( 21.1 ) also obtain with

the pragmatic anaphors in ( 21.2 ) . For instance , if the hearer is unable

to see or figure out what goings - on the speaker is pointing at in ( 21.2d ) ,

he will not fully understand the utterance - he will not have received

all the information he is intended to receive , and discourse cannot

properly continue .

If , as seems uncontroversial , the pragmatic anaphor in ( 21.1 ) refers

to a thing ( or physical object ) , those in ( 21.2 ) must also refer , but to

entities quite distinct from physical objects - namely , a place , a path ,

an action , an event , a sound , a manner , an amount , and a number ,

respectively . Thus , the world as construed must include such entities -

a variety rarely recognized in extant semantic theories .

Other grammatical constructions also support this range of entities .

One is the expression of identity and individuation with same and

different . Compare ( 21.3 ) , which express es identity and individuation of

physical objects , with ( 21.4a - f ) , which express identity and individuation

of other entity types . ( In some cases such sentences assert only

that two distinct individuals belong to a common type - for instance ,

Bill ate the samesandwichhe always eats , on the normal , nonregurgitation

interpretation . But even these cases presuppose the existence of distinct

individuals to be

{

}

categorize ,

. )

Bill picked

up thesame

things { ~

} Jack did.

( 21.3 )

[ Object

]

upsomething

different

than Jack did.

{ the

same

}

as }

{

place

( 2.4 ) a . Bill

ate

at .oJ

aU1l1erent

p ce

Iathan

Jack

di. [ Place

]

{ the

same

as }

way

b.Bill went

off .oJ

th Jack

di. [ Path

]

aU1l1erent

way

{ t ~::.ame

thinthin

. . g as

c.Bill

did

di. [ Action

} Jack

]

aUl

Ilerent

than

{ The

same

thing

{ as }

d . Adifferent

happened

today

happened

thing

}

than

. [ Event

yesterday

]

e .Bill

heard

{ t ~:: me

nois

.e as

} Jack

di. [ Sound

]

Jllerent

noise

than

482

Jackendoff

]

like Brahms.

this

f. Billcooks

meat ..I:"

{ thesame

t th

way an

( a ) }

hecooks

. [ Manner

eggs

]

andnumbers

are

identified

and

individuated

by different

Amounts

expressions

than

theother

entities

, but

thesemantic

parallelism

isclear

{ Bill

is as

tallas Jack

is. } [ Am t ]

g . Bill

is taller

than

Jack

is. oun

h.The

{ as many

notes

a }

there

were

marks

trumpeter

played t than

noes

onthe pag . [ Number

]

In order for these sentences to say what they do , there must be

entities of the requisite sort for the sentencesto talk about , and conceptual

structure must be capable of distinguishing among them . Accordingly

, we introduce into conceptual structure a set of onto

logicalcategory

] , [ PLAC ] , [ PATH ] , [ AcriON ] , [ EVEN ] ,

[ SOUN ], [ MANNE ], [ AMOUN ] , and [ NUMBE ] , as well as possible others

such as [ PRO

at least [OBJBCf

PERn ], [ SMEL ], and [ TIME ] . These can be thought of as

elements that serve as primitive " parts of speech " of conceptual structure

. Just as each syntactic constituent must be of a unique syntactic

category , so a conceptual unit must be of a unique onto logical category .

Each of these category features may be associated with either the

[ TOKE ] or the [ TYP ] feature. For instance , a perceived object will be

represented as an [ OBJBCf TOKE ] , and a category of objects as an [ OBJBCf

TYP ] . Similarly , a perceived event will be represented as an [ EVENT

TOKE ] , and a category of events as an [ EVENTTYP ] .

Now consider how the onto logical categories are expressed in language

. Traditional grammar implies that the correspondence between

syntactic categories and onto logical categories is fairly obvious : a noun

names a person , place , or thing; a verb names an action or state of

being;

] , which

seems always to be expressed by a noun . Otherwise , the situation is

more complex . The standard expression of [ EVEN ] is as a Sentence ; but

[ EVEN ] can also be expressed by a noun (earthquak ) . The standard

expression of [ PRo

PERn ] is as an adjective ( red , tall ) , but there are also

idiomatic noun phrases ( a gas , a bumme ) and prepositional phrases ( out

of luck ) that express [ PROPERTIES

] . And so on . This divergence shows

the potential complexity of the correspondence rules between syntactic

and conceptual structure . It is also important in showing that syntactic

structure cannot be based entirely on semantics , as is sometimes assumed

. ( See Jackendoff 1983 , chapter 4 , Grimshaw 1979 , and Jackendoff

1985for discussion . )

Let us look next at nonlinguistic connections to conceptual structure .

In order for the pragmatic anaphors in ( 21.2 ) to be interpreted , the

visual system must deliver information that corresponds to the visibly

483

SomeElementsof Conceptual Structure

aUllieren

way

features , including

and so on . Actually , the only simple case is [ OBJBCf

distinguishable onto logical categories in conceptual structure . For instance

, to distinguish ( 21.5a ) from ( 21.5b ) , the visual system must fill

in the pragmatic anaphors with objects in one caseand locations in the

other .

( 21.5 ) a. This is your coat , and that is your hat .

b. Here is your coat , and there is your hat .

To interpret ( 21.2e ), the auditory system must deliver information that

appears in conceptual structure as [ SOUND ] . In order to verify ( 21.4h ),

both the visual and the auditory systems must deliver information that

appears in conceptual structure as [ NUMBE ] .

Although perception of entities other than objects has not been prominent

in the literature , the work I have encountered ( for example , Mich -

otte 1954 on causation ; Jenkins et al. 1978 and Cutting 1981 on event -

perception ; remarks in Kohler 1929on temporal grouping; Piaget 1952

and Gelman and Gallistel1978 on amounts and numbers ) reveals characteristics

entirely parallel to the perception of physical objects , such as

the Gestalt properties of proximity , closure , " good form , " and the like .

There seems no bar in principle to the perceptual systems delivering

information about diverse onto logical categories , using mechanisms

similar to those for the perception of objects , if we think to look for it .

Besides giving evidence for an important class of primitives in conceptual

structure , this section reinforces earlier arguments on the priority

of Conceptual Semantics over Real Semantics. Even if we can refer

to this variety of entities , and even if truth -conditions must involve

them , we do not want to have to justify them as objective elements of

physical reality . For instance , the continuous flow of matter in the

physical world does not come neatly segmented into events , as language

seems to imply ; nor does it seem plausible that the Real World

contains manners segregated from the actions whose manners they are ;

nor does it contain numbers except in some curious Platonic sense. The

characteristics of these entities seem much less paradoxical if we regard

them in terms of how humans structure the world - what is real for us.

This in turn is determined by our capacity for mental representation ,

in particular , the properties of the onto logical categories available in

conceptual structure . Thus , the nature of the internal system of symbols

that support meaning must be a primary focus of semantic inquiry

Domains

We end our foray into conceptual structure with some further evidence

from language that bears on the organization of conceptual primitives .

( This material derives from the analysis in Gruber 1965and is developed

in greater detail in ] ackendoff 1976 , 1983 , chaps . 9 - 10. )

484

Jackendoff

Alvin I. Goldman (ed) - Readings in Philosophy and Cognitive Science -chap21.pdf

Plik z chomika:

Inne pliki z tego folderu:

Inne foldery tego chomika: