i

 

 

  

 

IIIIIIII III IIIIIIIIIIIIIIIIIII

  

   

31293 00797 1009

LIBRARY

Michigan State

s
I

  

 

  

 

University

 

 

This is to certify that the

dissertation entitled

CONS TRUC TIVI SM AND REFERENCE

presented by

David Brian Boersema

has been accepted towards fulfillment

of the requirements for

_Ph-D_ degree in Phi l O S Ophy

DateM

    ajor professor

Joseph

Hanna

I E

MSU i: an Aflinnarive Action/Equal Opportunity Institution

cry“ 7 17%,

0-12771

R n ,,,, Wf

 

    

 

MSU

  

LIBRARIES

n

  
  
 

V

RETURNING MATERIALS:

Place in book drop to

remove this checkout from

your record. ‘FINES will

be charged if book is

returned after the date

stamped below.

  

 

 

 

 

CONSTRUCTIVISM AND REFERENCE

by

David Brian Boersema

‘

A DISSERTATION

Submitted to

Michigan State University

in partial fulfillment of the requirements

5:

for the degree of

DOCTOR OF PHILOSOPHY

Department of Philosophy

1985

 

 

 

 

 

Copyright by

DAVID BRIAN BOERSEMA

1985

 
 

 

 

ABSTRACT

CONSTRUCTIVISM AND REFERENCE

by

David Brian Boersema

In this book I argue for a constructivist View of

reference.

By ‘constructivism’ I mean the view that we

socially construct reality (or, minimally, our understanding

of reality is only possible through social construction).

I

argue that the received theories of reference (the cluster

theory and the causal theory) are more alike than different,

and that they both share fundamental assumptions about

privacy, essentialism, truth and realism.

I argue against

these assumptions and

the

theories

of

reference

that are

based

on

them.

Finally,

I

offer

my alternative,

constructivist,

View

of

reference and

show

how

it

is

superior to the received theories.

  

 

 

 

For

J.F.H.

B.B.H.

S.A.H.

 

 
 

 

 

 

ACKNOWLEDGEMENTS

I want to

thank Barbara Abbott, Richard Hall, Ronald

Suter, and Randal Robinson for helpful questions, comments,

criticisms and suggestions.

I want especially to thank my

teacher and friend Joseph Hanna and acknowledge my debt to

him.

Joe’s thoughts, suggestions, examples and challenges

permeate this book.

Whatever merit this work has and

whatever philosophic skills I have are but shadows of him.

To Joe as a teacher, a scholar and - most importantly - as a

person,

I want to express my sincere gratitude.

With

apologies to Newton, even if I have been unable to see far,

I have, nonetheless, stood on the shoulders of a giant.

Finally, to Betina Henig and Sally Hiddinga, to whom I owe

so much, I want to acknowledge my deepest thanks.

  

 

 

TABLE OF CONTENTS

INTRODUCTION.

..... ......... ........ .......... ........

PART ONE.

THE RECEIVED THEORIES ...... ... ..........

CHAPTER ONE.

THE CLUSTER THEORY ..... . .............

12

CHAPTER TWO.

THE CAUSAL THEORY ........... .... .....

50

PART TWO.

CRITIQUE OF THE RECEIVED THEORIES .......

79

CHAPTER THREE.

CONVERGENCE OF THE RECEIVED THEORIES.

85

CHAPTER FOUR.

PRIVATE REFERENCE....................

96

CHAPTER FIVE.

INDIVIDUATION AND SIMILARITY.... .....

121

CHAPTER SIX.

ESSENTIALISM ..... ...... ..............

169

CHAPTER SEVEN.

CORRESPONDENCE AND REALISM ...........

201

PART THREE.

A CONSTRUCTIVIST ALTERNATIVE... .........

239

CHAPTER EIGHT.

COHERENCE AND VERIFICATIONISM ........

246

CHAPTER NINE.

COHERENCE AND REFERENCE ..............

272

CONCLUSION.

................................. .......

311

BIBLIOGRAPHY.

.. ....... . ..............................

314

  

 

 

INTRODUCTION

I am sitting with a philosopher in the garden; he says again

and again ‘I know that that’s a tree’, pointing to

a

tree

that is near us.

Someone else arrives and hears this, and I

tell him: ‘This fellow isn’t insane.

We are only doing

philosophy.’

(Ludwig Wittgenstein,

n Certainty)

  

 

 

 

INTRODUCTION

The main task in this book is

to propose and defend a

constructivist (i.e. arealist) position with respect to

epistemology and metaphysics and a coherence theory of

truth. (I don’t like the term ‘anti-realist’ because it

connotes to

me

a sense of

a position that is basically a

reaction, or perhaps response, to a prior position.

In the

same sense that ‘pro-choice’ carries a connotation different

than that of ‘anti-abortion’, I want a term that is the

analogue

to

‘pro-choice’

where

‘anti-realist’

is

the

analogue to ‘anti-abortionist’.

‘Constructivism’, I think,

"
?
I

fills

the bill.)

What I mean by ‘constructivism’ will, I

hope,

emerge

throughout the

book.

For

now,

a simple

statement of

it

is that it

is

the view that reality is

socially constructed.

The vehicle which I use to propose and defend my view

is that of the issue of reference.

My purpose for this is

two—fold.

First,

I

see

the issues dealt with and the

positions propogated under the rubric of reference as more

readily understandable and manageable if the epistemological

and metaphysical bases underlying them are exposed.

That

is,

one

of

my points is

to point out that questions about

reference rest on other, primarily epistemological,

qUestions which assume and are committed to conceptions of

 
 

 

 

truth and (a)realism.

Second,

I want to show the

significance of questions about truth and (a)realism by

showing how other issues (such as reference) rest on these

questions.

That is,

I want to say that dealing with these

questions is prior to and relevant to other philosophical

concerns.

The book is laid out as follows:

Part One:

The Received Theories

Chapter

1: The cluster theory

Chapter

2: The causal theory

Part Two:

Critique of the Received Theories

Chapter

3: Convergence of the received theories

Chapter

4: Private reference

Chapter

5: Individuation and similarity

Chapter

6: Essentialism

Chapter

7: Correspondence and realism

Part Three: A Constructivist Alternative

Chapter

8: Coherence and verificationism

Chapter

9: Coherence and reference

13'

In Part One I lay out the current debate between the

two

dominant

views

of

reference

as

presented

in

the

literature,

the cluster theory and

the causal theory.

In

Chapter 1

I first present the Searlean version of

the

Cluster theory,

summarize the causalist criticisms of

this

theory and finally respond - for

the cluster theory -

to

those criticisms.

The format in Chapter 2

is

the same.

I

fir-st

present

the

theory,

primarily

the

Kripkean and

 
 

 

Devittian versions, present the clusterist criticisms, and

finally respond - for

the causal theory -

to those

criticisms.

The overall point of Part One is

to ‘set the

stage’ in

the sense of presenting what I take the current

debate about reference to be.

In Part Two I suggest that

the

two received theories

are really two sides of the same theory.

I argue that they

share

many

fundamental

underlying assumptions

and

commitments, enough so as to see their dispute as a ‘family

squabble’.

In this part of the book I spell out what those

underlying

assumptions

and

commitments

are

and

criticize

them.

In Chapter 3

I argue that even on a ‘surface’ level

the two received theories are more alike than different and

suggest that when pushed they begin to converge into a

single theory.

They both, for example, are committed to the

speaker’s intention to refer as an essential element in

(successful) reference occurring.

In Chapter 4

I argue that

one

of

the fundamental reasons that

the received theories

converge

is

that

they

are

both

committed

to

private

reference (i.e. reference in a private language).

I first

lay out what I

take private reference to be,

then suggest

how I see both Searle and Kripke as being committed to this,

and finally explain how

this commitment is

a result of

a

particular prior conception of individuation and similarity.

I flesh out what I mean by this more thoroughly in Chapter

5-

Here I present the issues and problems concerning the

Principle of individuation and the similarity relation.

I

 

 

 

 

 

explain what these issues are and argue that they are

actually two sides of the same problem, that of determining

what there is.

I argue that individuation and similarity

are both language- and interest-dependent.

I then suggest

that both Searle and Kripke are committed to particular

conceptions of individuation and similarity, and which are

mistaken because

they

rest

on

the assumption

that

individuation and similarity are not language- and interest-

dependent.

In

Chapter

6

I argue

that

this

mistaken

conception of individuation and similarity is exhibited by

and presupposed by a commitment to essentialism by both

Searle and

Kripke.

I

first discuss

the

issue

of

essentialism,

examining

the

distinction

between

property

essences

and

individual

essences,

and

argue

that

essentialism -

the sort of essentialism which rests on

conceptions of individuation and similarity which are

not

language-

and

interest-dependent

- is

wrong.

I

then

explain how Searle and Kripke are committed to this sense of

essentialism and claim

that

to

the extent that

they

are,

their positions are weakened.

Finally, I remark on a

conception of essentialism which I suggest is acceptable (in

the

sense

of

resting on

the

‘right’ conception of

individuation and similarity).

This ‘non-realist’

essentialism,

while perhaps plausible,

is,

I argue,

quite

limited in scope.

In Chapter 7

I suggest that the received

theories of reference are committed to a correspondence

theory of

truth and

to realism.

I argue that

the

 

 

 

 

 

conceptions

of

individuation and

similarity as

not being

language— and interest-dependent rest on correspondence and

realism.

In fact, I identify realism with this particular

conception of individuation and similarity.

I argue that

the correspondence theory of truth entails this conception

of

individuation and similarity.

(In

the course of

this

examination, I consider Tarski’s semantic theory and argue

that

if

it

is committed to correspondence and realism it

begs the relevant questions here and if

it is not committed

to

correspondence and

realism,

but

is

theory-neutral,

it

fails to do the job that we want from a theory of truth.)

I

then argue that Searle and Kripke

are committed to

a

correspondence theory and to realism.

I conclude that the

two

received

theories

of

reference

share

fundamental

assumptions and commitments and

the particular assumptions

and commitments which they have are mistaken, such that

their theories of reference are also mistaken.

In Part Three I suggest an alternative view of

reference, a view which is based on the assumption and

commitment that individuation and similarity are language-

and interest—dependent.

This is borne out in the commitment

to a coherence theory of truth.

In Chapter 8

I argue for a

coherence theory.

I consider criticisms of coherence and

respond to those criticisms.

In Chapter 9

I suggest a view

of reference based on a coherence theory of truth (with its

underlying commitment to

a particular conception of

individuation and similarity).

I first examine what I take

  

 

 

 

to

be a Wittgensteinian view of reference, which I suggest

is

a version of

a coherence-based view of reference.

(I

also argue that Kripke is mistaken in his portrayal of

Wittgenstein as

holding a Searlean cluster

theory

of

reference.)

Finally I expand on this view by showing how it

overcomes the cases which are troublesome for the received

theories of reference and by giving examples which display

the superiority of a Wittgensteinian view of reference over

the received theories.

 

 

 

 

  

THE RECEIVED THEORIES OF REFERENCE

"When I use

a word," Humpty Dumpty said

in

a rather

scornful tone,

"it means just what I choose it

to mean —

neither more nor less."

"The question is,"

said Alice, "whether you can make

 

words mean so many different things."

(Lewis Carroll, Through The Looking Glass)

 

 

 

 

  

INTRODUCTION

In Naming and Necessity Kripke rejects the Russellian

 

theory of proper names as neither an adequate nor a correct

treatment of ordinary (proper) names.1

Kripke argues (with

many others) that the Russellian view fails to account for

the significance of the fact that different descriptions may

be (and are) used in place of a name to designate an object.

So one person might think of Aristotle as

the teacher of

\

Alexander’, another as

‘the most famous student of Plato’,

yet another as

‘the author of

the Metaphysics’, and so on.

(Even a single speaker might use these various descriptions

at different times when referring to Aristotle.)

No one of

these

descriptions

could

be

the

meaning

of

the

name

‘Aristotle’ or else the meaning of

the name would be in

constant flux.

Additionally, the notion of proper names as

disguised or shorthand definite descriptions is faulty, for

if

‘Aristotle’ means

‘the

teacher of Alexander’,

then the

statement ‘Aristotle was the teacher of Alexander’ would be

a

tautology - something it

is

not.

(Indeed, not only is

this statement not a tautology, but we could very well

discover that it is false.)

So, says Kripke, being the

teacher of Alexander cannot be part of

[the sense of]

the

name ‘Aristotle’.

Rripke then goes on to say that the most common way out

 
 

 

 

 

of this difficulty with such a view of names is

to say that

no particular description may

be substituted for

a name;

rather what is needed is

a family,

or cluster,

of

descriptions.

A good example of this, says Kripke, is found

in

Wittgenstein’s

Philosophical

Investigations.2

Kripke

quotes the following part of paragraph 79 as introducing the

idea of family resemblances:

Consider this example.

If one says ‘Moses did not

exist’, this may mean various things.

It may

mean:

the Israelites did not have a single leader

when they withdrew from Egypt-or: their leader was

not called Moses-or: there cannot have been anyone

who accomplished all that the Bible relates of

Moses-...But when I make a statement about Moses,-

am I always ready to substitute some one of these

 

descriptions for ‘Moses’?

I shall perhaps say: by

‘Moses’ I understand the man who did what the

Bible relates of Moses, or at any rate, a good

deal of

it.

But how much?

Have I decided how

much must be proved false for me to give up my

proposition as false?

Has the name ‘Moses’ got a

fixed and unequivocal use for me

in all possible

cases?

Kripke then states:

According to this view, and a locus classicus of it is

Searle’s article on proper names

Searle (1958)],

the

referent of

a name

is determined not

by

a single

description, but by some cluster or family.

Whatever

in some sense satisfies enough or most of the family is

the referent of the name. (1980, p.31).

In this first part I want to present the cluster theory

of reference (at least Searle’s version of

it)

in chapter

one, as well as present and evaluate Kripke’s criticisms of

the cluster theory.

In chapter two,

I want to present the

causal theory of reference (represented primarily by Kripke,

Devitt and

Donnellan),

and

present

and

evaluate

Searle’s

criticisms of the causal theory.

   

10

 

  

NOTE

lRussell’s theory of proper names is taken to be a

paradigm case of

the ‘descriptivist theory of names’.

For

Russell,

ordinary proper names are disguised definite

descriptions.

That

it,

ordinary proper names

can

be

replaced by descriptions (whichs the speaker associates with

the name).

For example, Russell (1912) states: "Common

words, even proper names, are usually really descriptions.

That is to say, the thought in the mind of a person using a

proper name correctly can generally only be expressed

explicitly if we replace the proper name by a description."

(1912, p.54).

Russell’s view is generally taken to be such

that for reference to occur/succeed when using a proper

name, the description which actually underlies the name must

be true of the object to which reference is made.

This

results in the problems noted below by both Searle and

Kripke.

(An important feature of Russell’s theory of proper

names

is

that

he

discusses

it within

the

context of

discussing different kinds

of knowledge,

knowledge by

acquaintance (e.g. immediate sensual knowledge) and

knowledge

by

description

(e.g.

propositional

knowledge).

For Russell, the latter is reducible to the former.)

21 believe Kripke is mistaken in ascribing the cluster

theory

to Wittgenstein,

but

that

issue

is

beyond

the

immediate concerns here.

This point will be taken up later.

11

 
 

  

 

CHAPTER ONE

THE CLUSTER THEORY

Searle

recognized

the difficulties

facing

the

Russellian

theory

of

names

as

shorthand

definite

descriptions, and amended it by claiming that a name refers

to an object in virtue not of a single description, but

rather of

a cluster,

or disjunctive set,

of descriptions.

(In this essay I will use ‘disjunctive set of descriptions’

to indicate the logical sum,

or disjunction, of

those

descriptions associated with a name.)

Says Searle:

Suppose

we

ask

the

users

of

the

name

"Aristotle" to state what they regard as certain

essential and established facts about him.

Their

answers would constitute a set of identifying

descriptions, and I wish to argue that though no

single one of

them is analytically true of

Aristotle, their disjunction is.

Put it this way:

suppose we have independent means of identifying

an object, what then are the conditions under

which I could say

of

the object,

"This

is

Aristotle?"

I wish to claim that the conditions,

the descriptive power of the statement, is that a

sufficient but so far unspecified number of these

statements

(or descriptions) are

true

of

the

object.

In short, if none of

the identifying

descriptions believed to be true of some object by

the users of the name of that object proved to be

true

of

some

independently located object,

then

the object could not be identical with the bearer

of

the name.

It is

a necessary condition for an

object to

be Aristotle that it satisfy at least

some of these descriptions. (1969, p.l69).

So, associated with a name ‘N’ is a disjunctive set of

descriptions (or descriptive predicates),

the satisfaction

 

 
 

 

 

 

by an object of some of which is necessary for the object to

be

the referent of

‘N’.

Clearly,

the disjunctive set of

descriptions which is associated with a name can vary from

speaker to speaker and from occasion to occasion;

as

new

beliefs are accepted about an object, new elements may be

added to

the

set of descriptions, and as

old beliefs are

rejected, some elements may be deleted from the set of

descriptions.

It

is

not

clear how many of

these

descriptions must be true of an object for a name to refer

to

that object, and it is no oversight on Searle’s part in

failing to specify such a sufficient number.

However, as

Searle says,

at least one of

the descriptions must be true

of an object in order for the name to refer to the object.

That is, it couldn’t be possible that all of the elements in

the set of descriptions associated with a name turn out

false and yet reference successfully occur.

The context within which Searle makes the above claims

is

that of

reference as

a speech act

(i.e.

an action

performed by

a speaker by

the

use of rule-governed

language).

The reason he concentrates his remarks on

reference

(and

philosophy

of

language

in

general)

within

this context is that

’all linguistic communication involves

linguistic acts.

The unit of linguistic communication

is...the production or

issuance of

the

symbol or word or

sentence in

the performance of

the speech act." (1969,

p.16).1

In his most recent work Searle places his views of

Speech acts

(and philosophy of language in general) within

13

 
 

 

 

the context of intentionality and philosophy of mind.

Philosophy of

language,

he

says,

is

a branch of

the

philosophy of mind.

"The capacity of

speech acts

to

represent objects and states of affairs in the world is an

extension of the more biologically fundamental capacities of

the mind (or brain) to relate the organism to the world by

way

of

such

mental

states

as

belief

and

desire,

and

especially through action and perception." (1983, p.vii).

Restricting his analysis to singular definite referring

expressions

(i.e.

proper names,

definite descriptions and

pronouns),

Searle claims

that

these

referring expressions

pick out or idenitfy one object, or ‘particular’, apart from

other objects or ‘particulars’ and

then go

on

to

say

something about that object or ‘particular’.

In discussing

the

success

of

a

referring

expression,

he

makes

the

distinction between a fully consumated reference and a

successful reference.

A fully consumated reference is one

in which an object is identified unambiguously for the

hearer; a successful reference (in the sense that we could

not accuse the speaker of having failed to refer) is one in

which

an

object

could

be,

on

demand,

identified

unambiguously for the hearer.

A question the theory of

reference must answer then becomes: What conditions are

necessary for

the utterance of

an expression to

be

sufficient to identify for the bearer an object intended by

the speaker?

The theory of reference which Searle proposes in order

1.

(Z/’

I

1/

 

 

 

 

  

to answer this question is capsulized in seven ‘rules

of

reference’

(listed below).

These rules presuppose two

‘axioms of reference’ and a ‘principle of identification.’

The axioms are:

A1.

The axiom of existence: There must exist one

and only one object to which the speaker’s

utterance of the expression applies.

A2.

The axiom

of

identification:

The hearer

must be given sufficient means to identify the

object from the speaker’s utterance of

the

expression.

(1969, p.82).

The principle of identification is:

P1.

A necessary condition for the successful

performance of

a definite reference in

the

utterance of an expression is that either the

expression must be an identifying description or

the speaker must be able to produce an identifying

description on demand. (1969, p.88).

Given these three conditions, we can new state Searle’s

seven

rules

of

reference

and

consider an

example

t0

illustrate the theory.

Searle states:

Given that S utters an expression R

in

the

presence of

H

in

a context C then in the literal

utterance of R,

S successfully and non-defectively

performs

the

speech act

of

singular identifying

reference if and only if

the following conditions

1-7 obtain:

1.

Normal input and output conditions obtain.2

2.

The utterance of

k

occurs as part of

the

utterance of some sentence (or similar stretch of

discourse) T.

3.

The

utterance

of

T

is

the

(purported)

performance of an illocutionary act.

4.

There exists some object X such that either R

contains an identifying description of

X

or

S

is

able

to

supplement

R

with

an

identifying

description of X.4

5.

S

intends that the utterance of

R will pick

13

 

 

 

  

out or identify X to H.

6.

S

intends

that

the utterance of

R will

identify X to H by means of H’s recognition of S’s

intention

to

identify X

and

he

intends

this

recognition to

be achieved by means of H’s

knowledge of

the rules governing R and his

awareness of C.

7.

The semantical rules governing R are such that

it

is correctly uttered in T in C

if and only if

conditions 1-6 obtain.

These rules can be exemplified with the following

sentence:

(I)

Venus is hidden from view by thick cloud cover.

Taking ‘Venus’ as

R,

it

is evident that the rules are

satisfied.

‘Venus’ occurs as part of

the utterance of

T,

where the utterance of T performs as illocutionary act (say,

that of informing H).

Venus exists and an identifying

description (e.g. the planet at such-and-such a place in the

Sky

at such-and-such a

time) could be offered if needed.

Additionally,

it

is intended

that ‘Venus’ picks out Venus

(and H knows this).

On

the other hand,

if not all of

the rules are

satisfied, then reference has not occurred.

For example, if

S uttered ‘Venus’

simply as

part of

a rhyming game

(e.g.

sounds that rhyme with ‘wean us’), then rule 5 (and perhaps

rule 2) would not be fulfilled and reference would not have

taken place.

Having laid out these rules of reference as a proposed

theory of reference, Searle turns directly to the problem of

definite descriptions and

proper names.

As noted above,

Searle amends Russell’s

view

by claiming

that

a singular

16

 
 

 

  

referring expression refers to an object in virtue of not a

single description, but rather of a cluster of descriptions

associated with a name.

Again, Searle claims: It

is

a

necessary condition for an object to

be Aristotle that it

”satisfy at least some of these descriptions.6

However,

Searle’s theory is more complex than that.

There are

problems, he says, with the view that names have no senses.

If names have no senses, then, as Frege pointed out, there

would be no cognitive difference between a=a and a=b.

In

addition,

negative existential statements

(e.g.

‘Cerberus

does not exist’) are meaningful (and,

in this case, true),

but they couldn’t be if names have no senses.

On

the

other hand,

says Searle,

strong arguments

militate against the view

that names

do have senses.

If

names have senses, at least in the form of being shorthand

descriptions, then "descriptions should be available as

definitional equivalents for proper names"

(1969,

p.166),

but

they are not.

In addition, if names have senses, then

if we substitute descriptions for names, then the following

(non-intuitive)

result would ensue:

some

non—analytically

true statements about an object using the name as subject

would be analytic (e.g.

‘Aristotle was

the

teacher of

Alexander’).

(This,

as noted above, is one of

the

criticisms levelled against Russell.)

Also, the meaning of

the name, and perhaps the identity of the object, would

ch ange every time there was a change at all in the object,

and the

name would

have different meanings

for different

17

  

speakers.

Surely this is not the case.

With strong arguments available both for and against

the hypothesis that names have senses, Searle reinterprets

the question ‘Do proper names have senses?’ as having two

forms (which he labels ‘weaker’ and ‘stronger’).

The weaker

form is:

‘Are any statements where the subject is

a proper

name and the predicate a descriptive expression analytic?’.

The stronger form is: ‘Are any statements where the subject

is

a proper name

and

the predicate an identifying

description analytic?’.

In answering these questions,

Searle states:

My answer, then, to the question, "Do proper names

have senses?" - if this asks whether or not proper

names

are

used

to

describe

or

specify

characteristics of objects -

is

"No".

But if

it

asks whether or not proper names are logically

connected with characteristics

of

the object

to

which they refer,

the answer is "Yes, in a loose

sort of way". (1969, p.170).

This "loose sort of way" is the necessity that for an object

to be X (e.g. Aristotle), it must satisfy the logical sum of

the properties attributed

to

X

(i.e.

at least one

description of the cluster must be true).

Searle summarizes

his position:

What I have said is

a sort of compromise

between Mill and Frege.

Mill was right in

thinking

that proper names do

not entail any

description,

that

they

do not have definitions,

but Frege was correct in assuming that any

singular term must have a mode of presentation and

hence, in

a way,

a sense.

His mistake was in

taking

the

identifying description which we

can

substitute for

the name as

a definition.

(1969,

p.170).

18

 
 

 

 

Kripke suggests that

the cluster theory contains

the

following six theses:

(1)

To every name or designating expression ‘X’,

there corresponds a cluster of properties, namely

the family of those properties 0 such that A [the

speaker (or hearer?)] believes ‘OX’.

(2)

One of the properties,or some conjointly, are

believed by

A

to pick out some individual

uniquely.

(3)

If most, or a weighted most, of

the d’s are

satisfied by one unique object Y,

then Y

is

the

referent of ‘X’.

(4)

If the vote yields no unique object, ‘X’ does

not refer.

(5)

The statement, ‘If X exists, then X has most

of the 0’s’ is known 3 priori by the speaker.

(6)

The statement, ‘If X exists, then X has most

of

the 0’s’ expresses a necessary truth (in the

idiolect of the speaker).

(1980, p.71).7

Kripke then presents a detailed critical analysis of these

theses.

Thesis (1):

This, Kripke tells

us,

is

a definition.

The import and legitimacy of this definition is

to be borne

out by

theses

(2)-(6) and so

he does not offer a critical

analysis of

this

thesis.

Rather,

an analysis

of

the

subsequent theses will,

if

they are shown to

be incorrect,

yield

the incorrectness

(or irrelevance) of

thesis

(1)

as

well.

Thesis

(2):

Rripke

offers

two counter-examples

to

demonstrate that thesis (2)

is incorrect.

First, he gives a

case to show that the thesis fails to be satisfied.

If we

consider the name ‘Feynman’, we note

that many people who

19

 

 

 

 

know very little about Feynman are nonetheless able to refer

to Feynman when using the name ‘Feynman’.

When asked about

Feynman, a person might say: well, he’s a physicist or

something.

The person may not

think

that

this picks out

anyone uniquely (and in this case, probably wouldn’t think

that it does).

Yet, says Kripke, it seems that such a

person is still using the name ‘Feynman’ as

a name for

Feynman.

Second, to show that the thesis is simply false,

Kripke offers the following case.

We can uniquely pick out

Einstein as

‘the

man

who discovered the

theory of

relativity’.

However, many people can only say of

the

theory of relativity that it is ‘Einstein’s theory.’

We are

led,

then,

"into

the most straightforward sort of vicious

circle."

(1980,

p.82).

The problem here

is

that

one

property is believed to pick out Einstein uniquely, but only

at

the cost of circularity, for in

this case the property

whicn picks out Einstein contains reference to Einstein

within it.

Thesis (3):

Kripke asks, "Suppose that most of the ¢’s

are in fact satisfied by a unique object.

Is that object

necessarily the referent of

‘X’ for A?" (1980, p.83).

His

answer is:

no.

To support this he presents the following

example.

Suppose someone says

that Gddel is

the man who

proved

the incompleteness of arithmetic.

Suppose further

that Gddel was not in fact the author of this theorem, but

rather a man named ‘Schmidt’ was.

0n the cluster view, says

Kripke, when the ‘ordinary’ speaker uses

the name ‘deel’,

20

 
 

 

 

 

satisfying the

description ‘the

discovered the incompleteness of arithmetic’."
So when the speaker talks of Gddel,

to Schmidt.8

p.85).

that nothing satisfies most,

number, of

the 0’s.

For example,

‘the inventor of the atomic
However, possibly no one really deserves to be
called the inventor of the device (or, at least, no single
Yet, even if ‘the inventor of the
were our only belief about Einstein, we would

.

referring to Einstein by ‘Einstein’.

So, this

thesis, too, is incorrect.

Thesis (5):

About this, Kripke says that it is simply

false.

Even if theses (3) and (4) happen to

be true,

this

hardly constitutes a priori knowledge that they are

true.

We

certainly believe that Einstein was

the man who

discovered the theory of relativity, but this belief is

21

 

 

¥

 

 

  

hardly 3 priori.9

Thesis

(6):

This

thesis, according to Rripke,

"need

not be a thesis of the theory if someone doesn’t think that

the cluster is part of

the meaning of the name." (1980,

p.65).10

This

thesis,

along

with

thesis

(5),

seems

primarily to say that a sufficiently reflective speaker

grasps this

theory of proper names.

Kripke’s attitude

toward this ‘necessity’ thesis is the same as toward the ‘a

prioricity’ thesis (5), namely, it

is obviously false.

He

states:

"It would seem

that

it

is

a contingent fact that

Aristotle ever did any of the things commonly attributed to

 

him today, any of

the great achievements

that we

so much

admire." (1980, p.75).

Having investigated each of

the theses (1)-(6) above,

Kripke concludes:

What I think the examples I’ve given show is not

simply

that there’s some technical error here or

some mistake there, but that the whole picture

given

by

this

theory

of

how

reference

is

determined

seems

to

be

wrong

from

the

fundamentals.

It seems to be wrong to think that

we

give ourselves

some properties which somehow

qualitatively uniquely pick out an object and

determine our reference in that manner. (1980,

pp.93-94).

Now, we need to ask: are Kripke’s criticisms legitimate

(i.e.

are

they fair criticisms of what Searle’s theory is

committed to)

and,

if

so,

are

they debilitating (i.e.

can

Searle’s

theory be defended or must it

be abandoned)?

In

answering

these questions each

of

the

theses

that Kripke

attributes

to

the cluster

theory will

be

investigated

in

turn.11

22

 
 

 

 

Thesis (1):

"To every name or designating expression

‘X’,

there corresponds a cluster of properties, namely the

family of

those properties 9 such that A believes

‘0X’."

Kripke regards this, as noted earlier, as a definition, the

legitimacy of which hinges on

(the fate of)

the other

theses.

Granting this assumption, we will turn to

the

remaining theses.

Thesis

(2):

"One

of

the

properties,

or

some

conjointly, are believed by A

to pick out some individual

uniquely."

The motivation for asserting this as a thesis of

the cluster theory is statements such as the following: "In

short, if none of

the identifying descriptions believed to

be true of some object by the users of the name of that

object proved to

be true of some independently located

object, then the object could not be identical with the

bearer of the name." (1969, p.169).

As

we saw, Kripke offers his Feynman example to show

that one can refer even though one does not believe that an

object has been uniquely picked out.

Such a counter-example

does indeed seem to violate the thesis (which does seem to

be implied by

the Searle quote above).

However, in other

places,

Searle seems

to allow for such cases.

Before

considering these cases,

though,

a preliminary distinction

that Searle makes between the primary and secondary aspects

of reference must be explicated.

Donnellan (1966) distinguishes between the referential

and

the attributive uses

of definite descriptions.

A

23

 
 

 

 

 

speaker who uses a definite description referentially uses

the description to enable

the bearer to pick out whom or

what the speaker is talking about and states something about

that person or

thing.

A speaker who uses a definite

description attributively uses the description to state

something about whomever or whatever is so-and-so.

For

example,

if

a speaker says

"Smith’s murderer is

insane,"

meaning that particular person over there, Jones,

the

speaker would be using the

term ‘Smith’s murderer’

referentially.

On

the other hand,

if

a speaker says

"Smith’s

murderer

is

insane,"

meaning not

any particular

person, but whoever it was who murdered Smith, the speaker

would be using the term ‘Smith’s murderer’ attributively.

Now,

in

arguing

that

the

referential-attributive

distinction is bogus, Searle distinguishes what he calls the

primary and secondary aspects of reference (or, under which

reference is made).12

Searle says:

Sometimes when one refers to an object one is

in

possession of

a whole lot of aspects under which

or in virtue of which one could have referred to

that object,

but one picks out one aspect under

which one refers

to

the object.

Usually the

aspect one picks out will be one that the speaker

supposes will enable

the hearer to pick out

the

same object.

In such cases...one means what one

says but one means something more as well.

these

cases

any

aspect will

do,

provided

In

it

enables the hearer to pick out the object.

(It

may even be something which both the hearer and

the speaker believe to be false of

the object...

(1979, p.144).

...provided

that

the

speaker’s

intentions

are

clear enough so that we can say that he really

knew what he meant,

then even though the aspect

expressed by

the expression he utters may not be

satisfied by

the object he

"has

in mind" or may

24

 
 

 

  

not be satisfied by anything, still there must be

some aspect (or collection of aspects) such that

if

nothing satisfies

it

(or

them)

the statement

cannot be true and if some one thing satisfies it

the statement will be true or false depending on

whether or not the thing that satisfies it has the

property ascribed to it. (1979, p.145).

The primary aspect under which reference

is

made

is

that aspect which, if not satisfied, would yield a statement

that cannot be

true.

The secondary aspect is any aspect

which the speaker expresses such that the speaker utters it

in

an attempt to secure reference to

the object which

satisfies the primary aspect, but which is not intended as

part of the truth conditions of the statement the speaker is

attempting to make.

For example,

the speaker,

looking at

someone

in

the

room,

says,

"Smith’s

murderer is

insane."

Now, the speaker and the hearer might agree that the speaker

has referred to, and made a true statement about,

that

particular person being looked at even though that person

(and perhaps everyone) fails

to satisfy the expression

‘Smith’s murderer’.

The speaker could, on demand, fall back

on another aspect,

say,

one expressed by ‘the person I am

looking at’.

If

it turns out that there is no person being

looked at, only a hologram perhaps,

then the speaker could

fall back on another aspect, say, one expressed by ‘the

person arrested by

the police and accused by

the District

Attorney as Smith’s murderer’.

If

it turns out that there

is

no such person,

then the speaker could fall back on

another aspect.

Eventually, however, an aspect must be

reached such that if no one satisfied it, then the statement

 
 

 

 

could not be true.

(And, though Searle does not say so

explicitly, we must assume that in such a case no one has

been referred to.)

The primary aspect of reference is this

‘last’ aspect, the aspect which either works or results in a

statement which is false.

The other aspects (e.g.

the one

expressed by ‘the person I am looking at’) are secondary.

As mentioned earlier, Searle believes that the

referential-attributive

distinction

is

bogus.

Having

introduced his primary-secondary aspect distinction, Searle

explains why.

According to Searle, all of Donnellan’s cases

are cases where the definite description is used to refer.

The difference in

the

cases

is

that

in

the so-called

referential cases

the reference is made under a secondary

aspect, and in

the so-called attributive cases

it

is made

under a primary aspect.

Furthermore, since every statement

containing a reference must have a primary aspect, then in

the so-called referential use

the speaker may

still

have

referred to something that satisfies the primary aspect even

though the expression uttered, which expresses a secondary

aspect, is not true of

that object and may not be true of

anything.

Having Searle’s primary~secondary aspect distinction in

hand,

we are now ready to return to Kripke’s criticism of

thesis

(2).

It seems

that with respect to secondary

aspects under which reference is made, Kripke’s thesis (2)

is not a thesis of the cluster theory; however, with respect

to primary aspects under which reference is made, Kripke’s

26

 
 

 

  

thesis (2)

— and his criticisms of

it - are on target, at

least for the cluster theory as Searle has fleshed it out.

One might be able to slightly amend Searle’s theory, though,

and handle the counter-examples which Kripke proposed.

Both

the Feynman

and

the Einstein-as-discoverer-of—relativity

cases work because the speaker has (apparently) no primary

aspect

under

which

to

refer

to

Feynman

and

Einstein

respectively.

However, if we allow that a speaker in such a

situation could appropriately appeal to another speaker or

source to supply other aspects by which to refer, then the

counter-examples would fail.

That is,

I may only know

Feynman as

‘a physicist or something’ and fully acknowledge

that

I

have

not uniquely picked out Feynman,

but believe

that reference has occurred successfully because I can add

something like,

"I don’t know anything more about Feynman,

but Keith does.

He can tell you all about Feynman."

I can

recognize that the set of properties which I associate with

‘Feynman’

do

not uniquely pick

out Feynman,

but also

recognize that someone else could amend the set such that

Feynman could be uniquely picked out.

(It would not be

necessary that Keith be

the person from whom I heard about

Feynman.

I might simply rely on

the fact that Keith is

a

physicist friend of mine who I know, or have good reason to

believe,

is familiar with other physicists.)

As noted

before, though, even if such an amendment to Searle’s theory

is legitimate, the theory is still committed to the primary

aspect under which reference

is

made as

having

to

be

27

 
 

 

  

~-'-_;—:p)n-;-_‘_‘-—_l-r

c:

.

,

satisfied in order for reference to occur.

While this would

entail that

thesis

(2)

is indeed a

thesis

of

the cluster

theory, Kripke’s particular counter-examples would no longer

be lethal to that thesis.

(Whether or not amended counter-

examples to the cluster theory -

in which this move to

appealing to other speakers to secure reference is blocked -

are possible or can work will be considered later.)

Thesis (3):

"If most, or a weighted most, of

the ¢’s

are satisfied by one unique object Y, then Y is the referent

of

‘X’."

This

thesis, Kripke tells us,

says

that the

speaker’s belief noted in thesis

(2)

- that

‘¢X’

-

is

correct.

That is, thesis (2)

is purely doxastic, it states

only that some property (or enough of them) is believed by

the speaker to uniquely pick out some object, whereas thesis

(3)

states that some property (or enough of

them)

in fact

does uniquely pick out some object.

The motivation for

13

making

this

a

thesis

of

the cluster theory

is statements

such as

those

noted above for

thesis

(2):

if none of

the

identifying descriptions believed to

be true of some object

by the users of the name of that object proved to be true of

some independently located object, then the object could not

be identical with the bearer of

the name.

This statement

clearly implies that if an object is identical with the

bearer of

a given name (i.e.

if

Y

is (identical with)

the

referent of

‘X’),

then at least one of

the identifying

descriptions believed to be true of

the object by the users

of

the

name

of

that

object

must

be

true

of

that

28

 

 

 

 

(independently located)

object

(i.e.

some

of

the fi’s

are

satisfied by Y).

As Searle says: "Since the speaker is

identifying an object to

the hearer,

there must,

in order

for this to be successful, exist an object which the speaker

is

attempting

to

identify,

and

the

utterance

of

the

expression by

the speaker must be sufficient to

identify

it." (1969, p.82).

Kripke’s counter-example to this thesis was deel being

identified as

the man who proved the incompleteness of

arithmetic, though, unbeknownst to the speaker, Schmidt was

actually

the

author

of

the

proof.

Since

the

(only)

descriptions associated with ‘deel’ are in fact satisfied

by Schmidt, under the cluster theory,

the speaker must be

referring to Schmidt by ‘deel’.

Searle’s (explicit) response to this proposed counter-

example is that depending upon our intention in a particular

context of using the name ‘deel’, the referent of ‘Gddel’

could "go in either direction." (1983, p.251).

Suppose, he

says, that Jones proclaims, “On line 17 of his proof, deel

makes what seems to me a fallacious inference."

If we query

Jones as

to who is meant by (his use of) ‘deel’, Jones

might

respond,

"I

mean

the

author

of

the

famous

incompleteness theorem."

If then informed that Schmidt was

the author, what would Jones say?

Says Searle:

It seems to me that he might well say that by

"Gddel"

he

just

means

the

author

of

the

incompleteness proof regardless of what he is,

in

fact, called.

Kripke concedes that there could be

such

uses.

They

involve what I

have called

secondary aspect uses of proper names.

(1983,

29

 
 

 

 

p.251).

0n the other hand, if Jones says, "Kurt deel lived in

Princeton" and we query Jones as to who is meant by ‘deel’,

Jones will likely be referring to deel and not Schmidt, and

will associate a different set of secondary aspect uses than

in

the first case (and as well a different primary aspect

use).

In any case, for Searle, it is not a singular, given

use of

a name

that determines reference, nor even the

particular description associated with that singular, given

use, but rather the underlying intentional content which is

attached to the name.

With the name ‘deel’, different

intentional

contents

and

primary

(and

secondary)

aspects

might be attached to the name for any given use of the name.

It

seems

to

me

that Searle would say

that

the reason

Kripke’s intuitions are so strong that when we use the name

‘Gddel’ we mean Gddel and not Schmidt is because in most

cases the intentional content attached to our use of a name

in fact allows us to pick out the ‘correct’ (i.e. intended)

object.

The fact is that we usually pick out the right

object when we use a name;

the primary aspect under which

reference is

made

(and usually

the secondary aspect) does

the job.

This should work for ‘deel’, too.

If

it doesn’t

(i.e.

if

the intentional content is ‘incorrect’, if the ¢’s

associated

with ‘deel’,

or

at

least

the

primary

associated

with ‘deel’,

turns

out

to

identify Schmidt),

then, for Searle, we have referred to Schmidt.

Now, it

seems to me that the motivation underlying Searle’s position

 

 

 

  

here is clear and very intuitive.

It is clear to see why

Searle would say that in

the cases above where all of

the

fl’s

(or

the primary d,

or

the intentional content)

associated with ‘Gddel’ turn out to identify Schmidt,

then

we obviously have referred to Schmidt.

On the other hand,

Kripke’s insistance that we refer to Gddel by our (every?)

use of ‘deel’ seems to be based on our belief that we pick

out the ‘correct’ object when we use names.

Once

again:

if

‘the author of

the incompleteness

theorem’ is the primary aspect under which reference is made

to deel, then,

a la Searle, we haven’t referred to Gddel,

but to Schmidt.

But why think that in such a case we have

referred to deel?

What is underlying Kripke’s claim that

even in

this situation we are in fact referring to Gddel?

It is not clear to me, unless it is the belief that usually

when we use

‘X’ we mean, and correctly pick out,

X rather

than Y.

However,

this hardly runs counter to Searle’s

theory; indeed, he agrees completely.

It

is noteworthy that in a footnote Kripke makes some

remarks

that sound rather conciliatory with regard to his

deel-Schmidt case and with regard to the cluster theory in

general.

The note reads:

The cluster-of-descriptions theory of naming

would make ‘Peano discovered the axioms for number

theory’

express

a

trivial

truth,

not

a

misconception,

and

similarly

for

other

misconceptions about the history of science.

Some

who have conceded such cases to me have argued

that there are other uses of the same proper names

satisfying the cluster theory.

For example, it is

argued,

if

we

say,

‘deel

proved

the

incompleteness of arithmetic,’ we are, of course,

31

 
 

 

  

referring to Gddel,

not to Schmidt.

But,

if we

say, ‘Gddel relied on a diagonal argument in this

step of

the proof,’ don’t we here, perhaps, refer

to whoever proved the theorem?

Similarly, if

 

someone asks, ‘What did Aristotle (or Shakespeare)

have in mind here?’,

isn’t he talking about the

author of

the passage in question, whoever he is?

By analogy to Donnellan’s usage for descriptions,

this might be called an ‘attributive’ use of

proper names.

If

this

is

so,

then assuming the

deel-Schmidt story,

the

sentence ‘deel proved

the incompleteness theorem’ is false,

but ‘deel

used a diagonal argument in

the proof’ is

(at

least in some contexts) true, and the reference of

the name ‘deel’ is ambiguous.

Since some

counter-examples

remain,

the

cluster-of-

descriptions

theory would still,

in general,

be

false, which was my main point in the text; but it

would be applicable in a wider class of cases than

I thought.

I think, however,

that no such

ambiguity need be postulated.

It is, perhaps,

true that sometimes when someone uses the name

‘deel’, his main interest is

in whoever proved

the

theorem,

and,

perhaps,

in some sense,

he

‘refers’ to him.

I do not think that this case is

different from the case

of Smith and Jones...

.

If

I mistake Jones for Smith, I may refer (in an

appropriate sense) to Jones when I say that Smith

is

raking the leaves;

nevertheless I

do not use

‘Smith’ ambiguously, as a name sometimes of Smith

and sometimes of Jones, but univocally as

a name

of Smith.

Similarly, if I erroneously think that

Aristotle wrote such-and-such passage,

I

may

perhaps use ‘Aristotle’ to refer to the actual

author of

the passage, even though there is no

ambiguity in my use of the name.

In both cases, I

will withdraw my

original

statement,

and

my

original use

of

the

name,

if apprised of

the

facts.

Recall that, in these lectures, ‘referent’

is used in the technical sense of the thing named

by a name (or uniquely satsifying a description),

and there should be no confusion. (1980, pp.85-86,

note 36).

Several points need to

be made here.

First, although

Kripke is obviously going to great lengths to put qualifiers

on his remarks (e.g. "perhaps," "in some sense,

"refer”

opposed to "refer"), he clearly concedes that not every use

the name ‘X’ results in X being referred to, and,

32

 
 

 

  

fact, a speaker might refer to Schmidt even when saying

‘Gddel’.

While Kripke admits that the cluster theory might

be applicable in a wider class of cases than he originally

thought,

he

states

that

other counter-examples

remain

to

nonetheless prove the theory false in general.

Part of the

purpose of

this chapter is

to suggest that none of the

counter-examples which Kripke has given do

the job,

and I

still see Kripke as bearing the onus to show that they do

the job.

Second, Kripke tries to overcome his concessions

to

the cluster theory by implying that the cluster theory

works (or might work) in these deel-Schmidt cases because

they imply that

the names are ambiguous.

However, Searle

never makes any claim to

that effect, nor is

it necessary

that he do so.

Searle claims that we have different primary

and secondary uses under which reference is made that result

in

the variation in reference.

For Searle, a name ‘X’ has

no meaning at all, so it certainly doesn’t have an ambiguous

meaning.

Rather, we intend to refer to

a given object by

using a given name and we associate different descriptions

with the name.

Because a given description might express a

given primary or secondary aspect under which reference is

made,

the same name ‘X’ might be used now to refer to X and

later to refer to Y.

Kripke’s charge of ambiguity here is

spurious.

Third, in dismissing even his concessions to the

cluster theory, Kripke emphasizes that we might refer to

Y

with ‘X’,

but,

of course, there’s no confusion of reference

here (since ‘referent’ is used in

the "technical sense" of

33

 
 

 

 

the thing named by a name), and happily we still refer to X

with ‘X’.

It is hard to believe that Kripke thinks that he

has explained anything by saying: well, ‘X’ refers to Y, but

‘X’ refers to X.

The unhelpful, non-explicatory, question-

begging nature of

this sort of claim needs no further

comment.

Perhaps a better candidate as a counter-example (because

it

is

a more ‘natural’ example than the deel case) is one

constructed by Donnellan (1972).

Consider a case, he says,

in which a young child is awakened during a party given by

his parents.

At this time the child encounters and speaks

with one of the party guests and learns the name of the

guest to

be ‘Tom’.

Later, reflecting on the event, the

child remarks,

Tom

is

a nice man’.

This is

the child’s

\

only description associated with ‘Tom’, but it does not

identify for the parents who Tom is, as they know many Toms.

It might even be that Tom is not a nice man, but the child

believes him to be so.

Now, suppose that another man was at

the party and this man, whose name is also

x

I

Tom , is nice.

However, the child did not meet this Tom.

It seems that if

thesis

(3)

is

true,

then the child referred to

the Tom he

did

not meet,

as

this man was

the unique object which

satisfied (a weighted) most - in this case all - of the ¢’s.

But

this

is

surely

incorrect.

Surely

the

child

was

referring to

the Tom he did meet,

even though he did not

satisfy the d’s and another man did.

Again, it seems that Searle’s response would be that as

34

 
 

 

 

 

discussed above: the aspect expressed by ‘is a nice man’ is

a secondary one in this case and could be replaced by the

aspect ‘is the man I met’.

Probably this aspect would serve

as

the primary aspect under which reference is made, but if

not,

then eventually one could be found which would - or

else reference would not have occurred.

And, again, appeal

could (theoretically) be made to other speakers to provide

the requisite expression of the primary aspect.

Thesis (4):

"If the vote yields no unique object, ‘X’

does

not referfl'

The motivation for asserting this

as

a

thesis of

the cluster theory is statements such as

the

following:

"There

must exist not more

than

one object to

which the speaker’s utterance of

the expression applies."

(1969, p.83).

As noted above, Kripke believes his previous

counter-examples show this thesis to be incorrect - just as

one can have false beliefs about X which are in fact true of

Y,

so one can have false beliefs about X which are in fact

true of

no one

or

of more than one.

So even if

‘the

inventor of the atomic bomb’ is the only description which a

speaker associates with ‘Einstein’, the speaker nonetheless

refers to Einstein by ‘Einstein’.

Now,

as we saw above Searle explicitly admits that we

sometimes use a description to identify an object even

though no unique object is picked out by that description.

The reason, for Searle,

that a description can work in

making reference

to an object even though the description

may be

true

of more

than one object or

true of

no one

is

35

 
 

 

 

 

that the description might express a secondary aspect under

which reference is

made.

The particular description may

work in identifying to

the bearer the object to which the

speaker is referring even though the description does not

uniquely pick out the intended object.

With respect to

secondary aspects under which reference is made,

then,

it

seems that thesis (4) is not a thesis of the cluster theory.

With respect to

the primary aspect under which

reference is made,

it seems

that if

no object is uniquely

picked out,

then ‘X’ indeed does not refer, and thesis (4)

is a thesis of the cluster theory.

But in such a case, why

think that reference has been made?

If ‘the inventor of the

atomic bomb’

is

in fact

the primary aspect under which a

speaker (supposedly) refers to Einstein, then it is not

clear why one should insist, with Kripke,

that the speaker

nonetheless

referred

to

Einstein.

As with

the deel

example, it seems that the motivation for claiming that the

speaker has still referred to Einstein is

that usually we

successfully refer to

the object to which we intend to

refer;

and again,

this

is hardly contrary to Searle’s

position.

So, as with theses (2) and (3), thesis (4) seems

to be a thesis of Searle’s theory only if

it is restricted

to

the

primary aspects under which reference is made

(in

which case, it is not at all obvious that it is false), and

even then only if

it is restricted such that reference can

be made only through the descriptions which the given

speaker associates with a name.

These restrictions are

30

 
 

 

 

 

certainly subject to challange.

Thesis (5):

"The statement, ‘If X exists, then X has

most of the w’s’ is known a priori by the speaker."

Kripke

says only that this is wrong - even if the above theses

happen to be true, a typical speaker hardly knows a priori

that they are.

Now, Searle says nothing explicitly to the effect that

the statement ‘If X exists, then X has most of

the @’s’ is

known a priori by the speaker.

An initial (minor) amendment

to this supposed thesis is in order.

The statement needs to

be changed to

‘If X exists, then X has some of (or the

 

‘primary’) fi’s’ to

be a thesis of Searle’s theory.

It

remains to

be seen if Searle’s theory is committed to this

amended thesis.

In investigating this, several points need

to

be made.

First, Kripke’s claim makes sense only if

Searle is proposing his

theory as giving the meaning of

a

name.

That is, only if ‘Aristotle’ means ’the most famous

student of Plato’ or ‘the teacher of Alexander’ or ..., can

this proposed thesis be

a thesis of

the theory.

However,

Searle does not see

the disjunctive set of descriptions as

giving the meaning of

a name,

but rather as

being an

identifying mode of presentation of a name.

The disjunctive

set of descriptions associated with a name is the means of

identifying an intended object, not the means of defining a

name.

(With respect to thesis (6), Kripke notes that if one

doesn’t think that the descriptions are part of the meaning

of

a name,

but only that they fix the reference,

then the

37

 
 

thesis need not be a thesis of the cluster theory.

The same

point,

it seems, would hold for

thesis

(5),

and it seems

that this could well be

the case for Searle, i.e.

the

descriptions only fix the reference.)

What if Searle’s theory were committed to the view that

the disjunctive set of descriptions associated with a name

did give

(in some

sense)

the meaning of

the name?

Would

this thesis then be a thesis of his theory, and, if so,

would this show a serious weakness in

the theory?

In

answering these questions, we need to get clearer on exactly

what Kripke is claiming.

It is clearly not a thesis of

Searle’s theory if the statement means that the speaker

knows a priori that some particular 9 is true of X.

Searle

definitely denies that this must be the case.

Searle does,

however, say that although no single description is

analytically true of

a given name,

the disjunctive set of

descriptions associated with the name is analytically true.

The doxastic corollary of

this would be

that

the speaker

knows

this analytic truth 3 priori.

This

reading of

Kripke’s claim,

then, seems not unlikely to be

a thesis of

Searle’s theory,

for it

is not unreasonable to suggest, as

Kripke does,

that a sufficiently reflective speaker knows

(or could come to know) this aspect of names and their uses.

However, even if this is

a thesis of Searle’s theory, it is

not obvious that this fact carries the unwelcome stigma that

Kripke attaches to

it.

The reason is that this reading of

Kripke’s claim seems to say nothing more controversial than

38

 
 

 

 

 

 

that the speaker knows ‘ahead of time’ (a priori, as it

were) that some description (but no particular one) is true

of X.

This is to say very little.

In addition, it seems to

be a statement not of our knowledge about the object or name

in question, but rather of our knowledge of the language and

how names are used.

That is, this reading seems to say

nothing more than what Searle readily asserts: if

I am to

refer by using

the name ‘Aristotle’,

then I must believe

that some

of

the descriptions which I associate with the

name are true of the intended object (and I know this fact

about names even before I use ‘Aristotle’).

Under this

reading of

thesis

(5), Searle is clearly committed to

it,

but

then the apparent ruinous significance of

this

commitment is lost.

There is another reason to suggest that while thesis

(5) may indeed be a thesis of Searle’s theory, that he would

be glad to accept it as such (and in fact this could be seen

by Kripke as not a point against Searle’s theory).

It

is

this:

if one believes that names have essences (or rather,

that objects named have essences), and if one believes that

at least one of

the descriptions in the disjunctive set of

descriptions

associated with

a

name

"picks

out"

this

essence, then one might be much more inclined to accept the

claim that the disjunctive set is analytically true of

the

object,

and,

as

a corollary,

that

the sufficiently

reflective speaker knows this analytic truth a priori.

Thesis (6):

"The statement,

‘If

X exists,

then X has

39

 
 

 

 

 

 

most of

the ¢’s’ expresses a necessary truth

(in

the

idiolect of

the speaker)."

As noted earlier, Kripke holds

that this ‘necessity’ thesis, like the ‘a prioricity’ thesis

(5),

is false.

He says:

"It would seem that it

is

a

contingent fact that Aristotle ever did any of

the

things

 

commonly attributed to him today, any_ of

the great

achievements that we so much admire." (1980, p.75).

Again,

making the amendment from ‘most of the fl’s’ to ‘some of

the

w’s,’ such a statement is true of Searle’s theory, but it is

not obvious

that

this

is

so objectionable or unintuitive.

If Kripke is saying that for the cluster theory it must be

the

case

that

if

X exists,

then some description must be

believed to

be

true of

X (and is

true of

X),

then indeed

Searle is committed to

this thesis (assuming Searle is

committed

to

names

having meanings).

But again,

as with

thesis (5),

this seems to be a commitment to

a fact about

language and

the

use

of names,

not a commitment to facts

about any object.

Furthermore, if one believes that names

have essences (or that objects named have essences) and if

one

believes

that

at

least one description of

the

disjunctive set of descriptions associated with a name

‘picks out’

this essence,

then this thesis may be not only

acceptable, but desirable.

Conclusion

What can be culled from this presentation and analysis

of Searle’s cluster theory of reference?

It seems that his

cluster theory is more sophisticated and defensible than it

40

 
 

 

 

 

 

 

has been given credit for being.

Searle’s primary—secondary

distinction staves off (or at least offers some promise at

staving off) many of Kripke’s objections.

In addition, this

distinction does not seem an ad hoc device; it has definite

intuitive appeal.

However,

the distinction does

not

immunize the theory from the criticisms that Kripke has

raised.

Searle’s notion of secondary aspects under which

reference is made does disarm much of Kripke’s attack, but

the necessity of

bringing

in

a primary aspect rearms

Kripke’s objections,

for

it still seems

that much of

the

basic theory which Kripke claims as

the cluster theory is

indeed the view that is held by Searle.

Of course, as

I

have tried to show, even to the extent that Searle holds a

version of

the theory that Kripke attributes to him,

it is

not obvious

that such a theory is

"wrong from the

fundamentals."

To those theses

to which Searle is

committed, it

is not so obvious that they are false or

implausible or even unintuitive.

Minimally, something other

than Kripke’s proposed counter-examples is needed

to

show

that Searle’s theory should be rejected.

The most evident response to these claims is that even

if Searle could successfully defend his

theory from these

particular counter-examples, others could be offered which

can do the trick.

All that is needed is one example where

no descriptions (or at least no

true descriptions) are

associated by

the speaker to a given name being used.

For

instance, the speaker might really know nothing at all about

41

 
 

 

  

deel.

So,

my attentive neighbor might remark to her

friend,

"My weird neighbor kept me

up last night yelling

something about ‘that damn turtle’ or ‘Gddel’ or something

like that.

I had to pound on the wall to get him to quiet

down."

Has my neighbor referred to Gddel?

The causal theory

devoid of descriptivist assumptions we will see later must

say: yes.

Searle, it seems, would say: no.

Why think that

reference has not occurred?

For Searle, the reason is that

 

rule 5 of his theory (and perhaps rule 6) has been violated.

[Rule 5:

S intends that the utterance of R will pick out or

identify X to H.

Rule 6:

S intends that the utterance of R

will identify X

to H by means of H’s recognition of S’s

intention to identify X and he intends this recognition to

be achieved by means of H’s knowledge of the rules governing

R

and

his awareness

of

C.]

Why

think reference has

 

occurred?

For the causal theory, as we will see later, the

reason is

that there

is

a causal chain linking ‘Gddel’ to

Gddel.

Apart from or prior to these theories, however, why

think that reference has or hasn’t occurred?

It is clear to

see why one would claim that reference hasn’t taken place:

the speaker didn’t intend to refer to deel; she didn’t even

know that ‘deel’ is

a name.

On the other hand,

it is not

clear why one would say that reference has taken place.

Perhaps another example would be fairer to

the causal

theory:

Suppose my obsession with Gddel has gotten to

the

point that I often blurt out my innermost thoughts

42

 
 

Gddel.

So,

my attentive neighbor might remark to her

friend,

"My weird neighbor kept me

up

last night yelling

something about ‘that damn turtle’ or ‘Gddel’ or something

like that.

I had to pound on the wall to get him to quiet

down."

Has my neighbor referred to deel?

The causal theory

devoid of descriptivist assumptions we will see later must

say: yes.

Searle, it seems, would say: no.

Why think that

reference has not occurred?

For Searle, the reason is that

 

rule 5 of his theory (and perhaps rule 6) has been violated.

a

[Rule 5:

S intends that the utterance of R will pick out or

identify X

to H.

Rule 6:

S intends that the utterance of R

will identify X

to H by means of H’s recognition of S’s

intention to identify X and he intends this recognition to

be achieved by means of H’s knowledge of the rules governing

R

and

his awareness

of

C.]

Why

think reference has

 

occurred?

For the causal theory, as we will see later, the

reason is

that

there is

a causal chain linking ‘Gddel’ to

Gddel.

Apart from or prior to these theories, however, why

think that reference has or hasn’t occurred?

It is clear to

see why one would claim that reference hasn’t taken place:

the speaker didn’t intend to refer to deel; she didn’t even

know that ‘deel’ is

a name.

On the other hand,

it is not

clear why one would say that reference has taken place.

Perhaps another example would be fairer to

the causal

theory:

Suppose my obsession with Gddel has gotten to

the

point that I often blurt out my innermost thoughts

 

 

 

 

concerning him.

On one such occasion, while standing at a

bus

stop,

I mutter, "Ha!

Gddel couldn’t hold a candle to

Cantor.

Now,

there was a man!"

Upon hearing this, a

bemused and intrigued stranger next to me asks, "Who is

deel?"

Has

the stranger referred to deel?

The causal

theory says:

yes,

because there is

a causal link from the

stranger’s utterance of ‘deel’ to Gddel.

Searle would most

likely say:

no, because it does not seem that the stranger

intended that

the utterance of ‘deel’ would pick out or

identify Gddel (i.e. Searle’s rule 5

is violated).14

The

theories, then, differ in the answer to whether the stranger

has referred or not to Gddel.

Which theory best matches our

pre-theoretic intuitions here?

It seems

to

me

that it

is

not clear, but I must admit that Searle’s theory comes

closer to my intuitions.

It seems to me that the very

question, ‘Who is Gddel?’ shows that the speaker has no

clear referent associated with the

name,

and

in fact

the

question can be paraphrased with no loss of meaning with a

different question,

‘Whom are you talking about?’ or ‘Whom

are you referring to when you say "G5del"?’

For the present

purposes,

I

think

the following

(minimally)

can

be

said:

there is no clear counter—example in which reference occurs

to Searle’s theory of

a name for which the speaker has no

descriptions at all associated with a name.

These examples illustrate a general point to which

Searle responds (properly, I

think)

in connection with the

variety of

proposed counter-examples

to

his

theory.

The

43

 

 

 

point is this:

the cases generally offered as inconsistent

with Searle’s theory are singular, private instances of

supposed reference.

That is,

these cases where reference

supposedly takes place are cases of singular utterances of

names, or singular statements in which the name occurs.

All

of

the examples

in which

the speaker has

at best

(apparently) one description associated with a name are of

this sort (e.g. ‘Feynman is

a physicist or something,’

’Einstein is the inventor of the atomic bomb’).

But it

seems

that

such

cases

(if

possible)

are

far

from

paradigmatic for reference.

We usually have an assortment

of descriptions associated with a given name, even if some

of

them are rather unilluminating in terms

of identifying

the object being referred to.

In those cases (if any) where

we have one or no descriptions associated with a name,

it

seems more plausible to either appeal to other speakers as a

source of reference or

to simply admit that reference has

not occurred.

It

is

in connection with the former

alternative

(appeal

to

other

speakers

as

a source of

reference)

that

I

say

that

the

usual counter-examples

to

Searle’s theory are cases of private instances of supposed

reference.

By that I mean the following: generally the

proposed counter-examples

to

Searle’s

theory

are

ones

in

which reference borrowing is not considered or allowed.

Yet,

as noted above in response to Kripke’s examples,

reference borrowing is a natural and common move.

To forbid

such a move would be

to make reference a ‘private’ matter,

44

 

 

 

 

i.e. possible in a private language.

This issue will become

significant in subsequent discussion, so I only mention it

here with

the

caveat

that

I

see

such a condition as

untenable.

While Kripke and others

have

said little or nothing

with respect to the private-public underpinnings of

reference (at least in the sense that I mean it - which will

be

discussed later),

they

have dealt explicitly with

the

issue of reference borrowing.

Kripke, for instance, briefly

discusses Strawson’s (1959) comments on reference borrowing

(cf. 1980, pp.90-92, 160-161) in connection with his remarks

on his non-circularity condition (C).

However,

the notion

of reference borrowing discussed there is that of appealing

to another speaker as

the historical source for

the given

speaker’s ability

to

refer by using a

name

(e.g.

‘By

‘Gddel’, I mean the man that Joe thinks proved the

incompleteness

theorem.’).

This

is

not

the

sense of

reference borrowing which I am considering here, and while

Kripke’s condition (C)

may well forbid cases of reference

borrowing in the Strawsonian sense, it is irrelevant to the

sense which I am using.

That sense of reference borrowing

is

this:

we

can

appeal

to

other

speakers

to

provide

requisite

information or descriptions

in

order

to secure

reference, not because they stand in the position of being

the historical source of my acquisition or acquaintance of a

name,

but

because

they

can

stand

in

the

position of

providing the needed identifying descriptions (which I lack)

  

 

 

in order to pick out the intended object.

I don’t

necessarily get my acquaintance with a name from these other

speakers, but I get identifying descriptions to help pick

out the intended object.

The other speakers who provide me

with these identifying descriptions could well be different

speakers than those from whom I first gained acquaintance of

a name.

So, for example, I could have first heard of Gddel

from Alice,

but appeal to Betty to provide me with an

identifying description so that I (and Carol) can pick out

deel.

Such appeal to other speakers is clearly not in

violation of Kripke’s condition (C)

(or Searle’s non—

circularity constraints).

Objections to

this sense of

reference borrowing, it seems to me, only make sense if one

presupposes a view of reference as a ‘private’ matter, and,

as mentioned above, this issue will be dealt with later.

Finally,

in connection with Searle’s theory of

reference,

I want

to remark that I have not tried to say

that Searle is right and Kripke is wrong.

Rather my point

so far has been to give a somewhat detailed - and hopefully

fair - treatment of

the

theory by presenting it and

then

considering criticisms of it from the perspective of someone

external to

the theory.

I now want to turn to

the causal

theory of reference and do the same.

Later I will suggest

how both the cluster and the causal theories can be seen to

be more similar than different, and having shown the shared

bases for that similarity, I will in

the final section

sketch out an alternative, third, theory of reference.

 

 

 

 

NOTES

1A few pages later Searle re-emphasizes this point of

reference as a speech act: "The term ‘referring expression’

is not meant to imply that expressions refer.

On the

contrary,

as previously emphasized,

reference is

a speech

act, and speech acts are performed by speakers in uttering

words, not by words.

To say that an expression refers

(predicates, asserts, etc.) in my terminology is either

senseless or is shorthand for saying that the expression is

used by speakers to refer (predicate, assert, etc.); this is

a shorthand I shall frequently employ." (1969, p.28).

sz "normal input and output" Searle simply means

to

cover such things as: both the speaker and bearer understand

the

language,

there are

no physical

impediments

(such as

laryngitis or deafness),

both are aware of what they are

doing, they are not acting or telling a joke, etc.

3An illocutionary act is

the performance of

an action

(such as referring, promising, requesting) by the utterance

of

a sentence.

Searle’s point here is

to rule out

gibberish.

4This condition is meant to capture the axioms and

principle mentioned above.

5This condition, says Searle, enables us to distinguish

referring to an object from other ways of calling attention

to it (such as hitting the bearer over the head with it).

6A few pages later he reiterates this:

it

is

a

necessary truth that Aristotle has

the logical sum

(inclusive disjunction)

of

the properties commonly

attributed to him.

7Kripke adds

a condition (C):

For

any successful

theory,

the account must not be circular.

The properties

which are used in the vote must not themselves involve the

notion of reference in such a way that it is ultimately

impossible to eliminate.

That is, among the properties

associated with the reference of

a

term ‘X’ cannot be

the

property ‘is called "X"’.

The motivation for this condition

is obvious - to prevent a theory which does not lead to any

independent determination of

a reference.

Searle makes the

same point: "...we are only justified in calling [an object]

‘Everest’

if

we

can give a reason for supposing it

to

be

identical with what we used to call ‘Everest’ and to give as

that reason that it is called ‘Everest’ would be circular."

(1969, p.167).

8While

the deel-Schmidt example might seem a bit

‘cooked up’, Kripke provides other, more natural, examples.

For instance,

it

is believed by some people

that Columbus

47

 

 

 

 

was

the first man

to realize that

the earth is round and

that he

was

the first European to land in

the western

hemisphere.

Prooably neither of these beliefs is true.

If

thesis

(3)

is

correct,

then when people use

the

term

‘Columbus’ they really refer to some Greek if they use the

roundness of the earth, or to some Norseman if they use the

‘discovery of America’.

But they don’t, says Kripke; they

refer to Columbus.

9Kripke never states in Naming

and Necessity just what

 

a

priori knowledge

is.

He argues against its conflation

With necessity.

He says that "the notion of a prioricity is

a concept of epistemology." (1980, p.34).

Beyond that,

however, he says only what it

is not

e.g.

it

is not

the

metaphysical concept of necessity).

Since he

is reticent

about meanings and argues against a prioricity as knowledge

that can be known prior to or independent of experience

(becadse such a definition includes the modal notion of

possibility, which is metaphysical rather than epistemic),

it is not clear to me what Kripke takes a prioricity to be.

10Kripke claims

that

there are

two ways in which the

cluster theory (or even the single description theory) can

be viewed.

One is

that of saying that the cluster (or

single description) actually gives the meaning of

the name

(e.g.

‘Aristotle’ means ‘the most famous student of Plato’

or

‘the teacher of Alexander’, etc.).

Another is

that of

saying that

the cluster (or single description) determines

the reference (e.g.

‘the teacher of Alexander’ is used

to

determine

to whom someone

is

referring when someone

says

‘Aristotle’).

Apparently thesis (6)

is

a thesis of the

cluster theory only if the descriptive referring expressions

are used to determine the reference of ‘X’.

llMichael McKinsey (1978)

has done a somewhat similar

critique and found Kripke’s arguments lacking credibility.

Other attempts have been made to defend the cluster theory,

e.g. Boer (1972) and Ingber (1979).

lZWhether or not Searle succeeds in showing Donnellan’s

distinction to

be bogus is not the main concern here.

It

should also be noted that Kripke (I977) criticizes

Donnellan’s distinction.

13It is not clear if Kripke thinks that for the cluster

theory the property (or set of properties) which the speaker

believes

uniquely

picks

out

the object

being

referred

to

must be

the same property (or set of properties) which in

fact uniquely picks out the object being referred to.

This

is a minor point, however.

/

_

.

l4Actually, for reasons seen later, I will clarify this

claim and assert that in the second case Searle thinks that

reference does occur.

 

For now,

though,

I want to portray

48

 

 

  

Searle as

a ‘pure’ cluster theorist,

such that having the

appropriate associated description is necessary and

sufficient for reference to occur.

 

 

 

 

CHAPTER TWO

THE CAUSAL THEORY

Having analyzed the cluster theory of reference,

and

finding it "wrong from the fundamentals,

Kripke proposes to

present a "better picture"

of

how reference

takes

place.

"In general," says Kripke, "our reference depends not just

on what we think ourselves, but on other people in the

community, the history of how the name reached one, and

things like

that.

It

is

by following such a history that

one gets to the reference." (1980, p.95).

This "better

picture," which Kripke says is "not a theory" (1980, p.96),

has nonetheless served as

the nascency and kernal of what

has come to be called the ‘causal theory of reference’.

Kripke gives a "rough statement" of such a theory:

A rough statement of

a theory might be

the

following: An initial ‘baptism’ takes place.

Here

the object may be named by ostension, or

the

reference

of

the

name

may

be

fixed

by

a

description.

When the name is ‘passed from link

to link’, the receiver of the name must, I think,

intend when he learns it

to use it with the same

referepce as the man from whom he heard it. (1980,

p.96).

The sense in which this is a causal theory of reference

is that the passage of a name from link to link is said to

secure a causal connection between the name of an object and

the object.

The initial baptismal act of naming the object

(by ostension, perhaps) establishes the causal connection in

50

  

 

  

the first place.

Later uses of

the name must be connected

to

the object in some sort of causal chain stretching back

to the original naming act.

As noted above, Kripke does not

explicitly propose a theory of reference.

However, others,

e.g. Devitt (1981), have attempted to forge a causal theory

of reference based on Kripke’s picture.

According to

Devitt:

The central idea of

a causal theory of names is

that our present uses of a name, say ‘Aristotle’,

designate the famous Greek philosopher Aristotle,

not in virtue of

the various things we (rightly)

 

believe true of him, but in virtue of a causal

network stretching back from our uses to the first

uses of the name to designate Aristotle.

It is in

this way that our present uses of a name "borrow

their reference" from earlier uses.

It

is

this

social mechanism that enables us all to designate

the same thing by a name. (1981, p.25).

This

simple statement of

the causal

theory indicates

what Devitt takes

to

be

the basic elements of

the theory.

Reference, or designation, is explained in terms of d-chains

(short for ‘designating chains’).

There are three types of

links in a d-chain: (l) groundings, which link the chain to

an object, (2) abilities to designate, and (3) communication

situations in which abilities are passed on or reinforced

(i.e. reference borrowing).

In a grounding, Devitt tells us, a person perceives an

object (preferably face to face), correctly believing it to

be

an object of

a certain very general category.

The

grounding consists in

the person coming to have ‘grounding

thoughts’ about that object as

a result of

the act of

perceiving the object.

A grounding thought about an object,

51

 

 

  

according to Devitt, "includes a mental representation" of

that object brought about by an act of perception.

The

thought is

one which a speaker would express using a

demonstrative from that language.

To use one of Devitt’s

examples:

"Consider the case of our late cat.

We acquired

her as a kitten.

My wife said, ‘Let’s call her "Nana" after

Zola’s courtesan.’

I agree.

Thus Nana was named." (1981,

p.26).

In this instance, Devitt perceived the cat and

correctly believed it

to be a cat.

Furthermore, it was

paramount that Devitt saw it as a member of the very general

category which contains cats.

(I must admit, I’m not sure

what

this category would be:

cat?

animal?

living animal?

domesticated animal? house pet? single-headed entity?)2

Having perceived it, Devitt was causally affected by it.

As

a result of seeing the object, and hence being causally

affected by

it, Devitt comes

to have ‘grounding thoughts’

about it,

i.e.

he comes

to have mental representations of

the

object

(and

of perceiving the object?).

For Devitt,

having the appropriate mental representations is a necessary

condition for having grounding thoughts.

The name, having been grounded, can now be used by

speakers to refer to

the object because a causal link has

been established between the initial use of the name and the

object, and a causal link can be established between later

uses of the name and the initial use.

For instance, having

named

the

cat

‘Nana’,

Devitt could then exercise his

new

ability (to refer to Nana) by saying, ‘Nana is hungry.’

He

52

 

 

  

could speak to his friends about Nana,

thus enabling them,

in turn, to refer to Nana.

These friends (and their uses of

‘Nana’) are causally linked to Nana by perceiving (hearing)

Devitt’s utterance, in which mention of Nana is made.

Since

to perceive is

to

be causally affected,

they are

now

causally affected (in a sense) by Nana.

These friends now

have

the ability to refer to Nana because gf this causal

 

linkage.

The nature of

the causal hook-up in

this chain

seems

to

be

based on identity beliefs,

says Devitt (i.e.

beliefs

we

would

express

using

identity

statements).

According to Devitt, non-initial uses of

a name become

grounded

in

objects

by

taking advantage

of grounding

thoughts.

This process of

taking advantage can be either

direct or indirect, but in both cases advantage is taken by

coming to hold identity beliefs.

Direct advantage of

grounding thoughts is taken by an identity belief containing

a demonstrative representation like those in the grounding

thoughts

(e.g.

‘that

cat

is

Nana’

or

by

definite

descriptions).

Indirect advantage of grounding thoughts is

taken when a speaker has hondemonstrative representations of

an object, i.e.

by an identity belief that includes an

already grounded hondemonstrative representation (e.g.

‘the

F’, knowing that Nana is the F).

In both cases of advantage

being taken,

it

is

the identity beliefs

that

lead

to

the

representational

thoughts and

these

identity beliefs

link

the

thoughts

associated

with

‘Nana’

to

the grounding

thoughts and thus

to Nana.

53

  

 

 

 

 

Closely connected with the notion of grounding and the

uses of a designating term is

the ability to designate.

Devitt claims that "we should see a person’s ability to

designate with a

term as,

basically,

his

having

thoughts

that include mental representations which are both of

the

object and associated with the

term"

(1981,

p.130).

The

ability to designate can be divided into the ability to

designate i1 thought and the ability to designate 12 Speech

(or writing and so forth), with the former being basic.

Says Devitt:

For a person to have an ability in thought to

designate an object, an ability associated with a

physical type,

is for him to have a set of

thoughts

including

tokens which are grounded

in

the object and which dispose him to use the type.

(1981, p.132).

...for a person to have an ability to designate an

object with a sound type

is for him to have

(a)

the power of speaking that type, and (b) a set of

thoughts

including

tokens which are grounded in

the object and which dispose him to speak the

type. (1981, p.131).

So, Devitt’s ability to designate his cat —

if

it

is

a

speech ability -

is

the ability to utter certain sound

tokens.

The set of thoughts associated with that type (viz.

sound) dispose him to speak it.

Likewise, mutatis mutandis,

for the ability to designate if it is a thought ability.

We have,

then,

the first

two

types

of links in

a

designating chain.

The

third

type of link is

that of

communication situations in which abilities to designate are

passed on or reinforced (i.e. reference borrowing).

In

reference borrowing,

says Devitt,

the act of perceiving a

I

54

 

designation of the object by the term plays the role for the

borrower that the earlier act of perceiving the object

played for a person present at a grounding; it grounds the

term in the object (1981, p.137).

So, when Devitt comments

to

his

friends

that

Nana

is

hungry,

‘Nana’

is

a

designational

term,

since

it

had

underlying

it

a

representation grounded in the object.

On the strength of

this utterance, Devitt tells us,

the friends come to have

certain thoughts and gain

the benefit of

that grounding.

The

friends

have mental representations which are

of

the

object in virtue of being causally linked to the object via

the act of borrowing.

Underlying these representations are

designating chains grounded in the object in virtue of

appropriate grounding thoughts and identity beliefs.

Criticisms of £33 Causal Theory

In response to

the causal theory of reference, Searle

has said little with respect to Kripke and almost nothing at

all with respect to Devitt.

He has, however, directed

criticisms against aspects of the causal theory (e.g. proper

names do not have senses) as far back as his 1958 essay and

against the causal theory and Kripke in his most recent

works.

Among his criticisms are:

(l)

The view that proper names have no senses is,

minimally, counter-intuitive.

(2)

Kripke’s

distinction between speaker

reference and semantic reference is bogus (as

is

Donnellan’s

distinction

between referential uses

and attributive uses of definite descriptions).

(3)

There are intuitive counter-examples showing

that the causal theory of reference fails to give

55

 

  

necessary conditions for successful reference and

fails to give sufficient conditions for successful

reference.

(4)

Kripke’s

(and Donnellan’s) position

is

really a descriptivist one.

These criticisms need to be looked at in detail.

Criticism (l):

The View that proper names have no

senses is, minimally, counter-intuitive.

In the context of

existential

statements

such

as

‘Princeton exists’

or

‘Cerberus does

not exist’ proper names cannot be

said

to

refer.

If

they

did,

the precondition of

such statements

having a truth value would guarantee its truth (in the case

of

affirmative existential statements)

or

its falsity

(in

the case of negative existential statements).

This is

because,

says

Searle,

every

existential

statement

states

that a certain predicate is instantiated.

An existential

statement does not refer to an object and state that it

exists,

rather it expresses a concept and states

that the

concept is instantiated.

If Cerberus does not exist and if

‘Cerberus’ has no sense,

then how can we make sense of

‘Cerbetus does

not exist’?

Surely,

at least intuitively,

the statement ‘Cerberus does not exist’ is true.

But this

can’t be if ‘Cerberus’ has no sense.

Another problem with the view that proper names have no

senses is that such a view cannot explain the cognitive

difference between identity statements of

the form a=b and

of

the form a=a.

That is,

if proper names have no senses,

but at best only referents,

then the assertion ‘Everest is

Chomolungma’ could convey no more information than the

56

 

 

  

assertion ‘Everest is Everest.’

However, this is counter-

intuitive, at least.

Criticism (2):

Kripke’s distinction between speaker

reference and semantic reference is bogus (as is Donnellan’s

distinction between referential uses and attributive uses of

definite descriptions).

According to Searle, Kripke (1977)

rejects

-

or

reinterprets

- Donnellan’s

referential-

attributive distinction as rather a distinction between

speaker reference and semantic reference.

For Kripke, if a

speaker has a designator in his idiolect, then certain

conventions of

his idiolect determine the referent in his

idiolect.

This is the semantic referent of the designator.

On the other hand,

the object which the speaker wishes to

talk about, on a given occasion, and which the speaker

believes fulfills the conditions for being the semantic

referent of

the designator is the speaker’s referent of

the

designator.

In Donnellan’s attributive cases, speaker

reference and

semantic reference coincide

because

the

speaker’s

intention

is

just

to

refer

to

the semantic

reference.

In the referential cases, however, the speaker’s

reference and the semantic reference would be different if

the

speaker

is

mistaken

in

believing

that

they

both

determine the same object.

For example,

two speakers see

Smith in

the distance and mistake him for Jones.

One

speaker asks, ‘What is Jones doing’?

The other speaker

answers, ‘Raking the leaves’.

Now, ‘Jones’ is

the name of

Jones,

not of Smith.

The semantic referent of ‘Jones’ is

57

 

 

 

 

Jones, but the speaker referent of ‘Jones’ in this case is

Smith.

Kripke’s account couldn’t quite be right as it stands,

says Searle,

because in

the ‘referehtial’ use

the speaker

need not even believe that the object referred to satisfies

the expression he

uses.

For example,

to avoid political

persecution one might refer to

the usurper of

a throne as

‘the King’ even though one does not believe the usurper to

be the king.

Where Kripke goes wrong is

in his attempt to

analyze both speaker’s reference and semantic reference in

terms of different kinds of intentions.

Kripke states that

in a given idiolect "the semantic reference of a designator

(without indexicals) is given by a general intention of the

speaker to refer to a certain object whenever the designator

is used.

The speaker’s referent is given by a specific

intention,

on

a given occasion,

to refer to

a certain

object." (1977, p.264; Kripke’s emphasis).

Searle responds

to this attempt:

This is where the account bogs down.

In the sense

in which I really do have both general and

specific

intentions

(e.g.

I

have

a

specific

intention to drive to Berkeley tomorrow, and a

general intention to drive on the right hand side

of

the rode, ceteris paribus, whenever I drive in

the United States)

I

have

no

such general

intentions about definite descriptions.

If my use

of definite descriptions required such general

intentions I would have to form an infinite number

of

them since there are an infinite number of

definite descriptions I am able to use and

understand in my language. (1979, p.156).

Rather

than

speaking

of

speaker’s

reference and

semantic

reference, Searle suggests that the real distinction to

be

58

 

 

  

noted with respect to reference is that between primary and

secondary aspects under which reference is made (which was

discussed in chapter 1 above).

Criticism (3):

There are

intuitive counter-examples

showing that the causal theory of reference fails to give

necessary conditions for successful reference and fails to

give

sufficient

conditions

for

successful

reference.3

Searle presents several counter-examples.

‘Madagascar’ was

originally the name of part of Africa.

Mistakenly, Marco

Polo used the name to refer to an island off

the coast of

Africa.

So, says Searle, the present uses of

the name

‘Madagascar’ (by which we now mean to refer to the island,

not the mainland) satisfy a causal connection that connects

it with the mainland, but that is not sufficient to enable

it to refer to

the mainland.

The use of names of abstract

entities (e.g.

‘two’) or of ‘future’ entities (e.g.

‘the K

car’ while still in the planning stage) are examples of

reference in which no causal connection is necessary between

an object and (the use of) a name.4

Criticism (4):

Kripke’s (and Donnellan’s) position is

really a descriptivist one.

Says Searle:

From

the

point

of

view

of

the descriptivist

theory, what the causal analysis amounts to is the

following:

the "causal chain of communication" is

simply a characterization of

the parasitic cases

..__.—_——_——-——____

seen from :3 external point of view.

Let m§_ffy

in

the chain of communication the speaker must

have the intention,

"when I utter ‘N’

I mean to

refer to the same object as the person from whom I

got

the name

‘N’".

The descriptivist says

that

one sort of

identifying description that one can

attach to a name ‘N’ is "the person referred to by

59

 

 

 

 

 

 

others in my linguistic community as

‘N’".

Both

sides agree that this is not enough by itself:

Kripke insists that the chain must terminate in an

.

initial

baptism;

the descriptivist allows

for

a

variety of ways in which it can terminate, of

which an initial baptism is one.

Where is the

difference?

As

far

as

the

issue between

descriptivism and

the causal theory is concerned

there is no difference: Kripke’s theory is just a

variant form of descriptivism. (1983, p.244).

Now, as with the earlier criticisms of Searle’s cluster

theory, we need to ask: are Searle’s criticisms of

the

causal

theory

legitimate,

and,

if

so,

are

they

debilitating?5

Criticism (1):

"The view that proper names

have no

senses

is, minimally, counter-intuitive."

As

stated,

the

first counter-intuitive result of denying senses to proper

names

is

to

make

(certain)

existential

statements

nonsensical

or paradoxically

false.

The

second counter-

intuitive result is

that the apparent cognitive difference

between a=a and a=b cannot be explained.

Donnellan (1974)

and Devitt (1981)

have responded to

the first counter-intuitive result.

Donnellan makes use of

an "admittedly not well-defined" notion of a ‘block’ to

sketch a means of dealing with the problem of negative

existentials.

For Donnellan, there is

a ‘block’ in the

history of

the

use of

a name when events preclude any

referent being identified.

An example of

a block would be

the following: a child who once believed in Santa Claus

comes to abandon this belief.

What has the child learned?

Says Donnellan:

60

 
 

 

 

 

Our account is

that he has learned that when in

the past he believed something,

for example,

"Santa

Claus

comes

tonight,"

and

would

have

thought himself when saying this

to be referring

to someone, the historical explanation of this

belief does not involve any individual who could

count as

the referent of "Santa Claus"; rather it

ends in a story given him by his parents, a story

told to him as factual. (1974, p.23).

The block,

then,

is

the introduction of

the name into the

child’s speech via

a fiction told

to him by his parents.

Another example

is

the

following:

the Homeric

poems,

suppose, were written not by

a single author, but by many

people.

Perhaps the poems were combined from a patchwork of

many writings over time.

Suppose further that someone later

attributed these poems to

a single author, Homer.

If

this

were the historical explanation of our saying, for example,

‘Homer wrote the Iliad’, then the block occurs at the point

at which the person attributed the poems to Homer.

Armed with the notion of a ‘block’, Donnellan proposes

the

following rule

for giving

truth conditions of

existential statements containing a name:

(R)

If

N

is

a proper name that has been used in

predicative statements with the intention to refer

to some individual, then ‘N does not exist’ is

true if and,only if the history of those uses ends

in a block.0

So we can assign truth values to negative existentials such

as

‘Cerberus does

not exist’ even though we deny

that

‘Cernerus’ has a sense.

Devitt responds

to

the Searlean objection of negative

existentials by proposing the following rule:

If

e1

is

a singular term,

then rel exists1 is m—

trueS if and only if there is an object a

to which

61

 
 

 

 

 

e1 m-refersS (where ‘trues’ means "true relative

to

the assignment of sequence 5," ‘referss’ means

"refers relative to the assignment of sequence 3,"

‘m-trues’ means "trues relative to a structure m

which maps terms onto the world," and ‘m-referss’

means "referss relative to a structure m which

maps terms onto the world").7

Negative existentials are handled by

the addition of

a

negation rule:

If

e1

is

a formula and e2

is

a negation symbol,

then re2(e1)1 is m-trues if and only if e1

is not

m-true .

s

As

with Donnellan’s

rule

(R),

these rules

do

not analyze

negative existentials in

the sense of giving a synonym or

trarlslation of

them,

but they do offer a :nethod of

determining the truth conditions of such statements.

Kripke more directly addresses the second counter-

intuitive result of names having no senses (viz. a=a and a=b

would then have no cognitive difference, but they do).

in

laving out his argument for why he thinks that names do not

4

n
nave

senses,

he

first attempts to clarify several notions

and distinctions which he sees as important.

Kripke argues

that,

contrary

to popular usage,

a prioricity

is

not

interchangeable with necessity.

Both a prioricity and

a

posterioricity are epistemic concepts, whereas both

necessity and possibility are metaphysical concepts.

To see

that

a

prioricity

is

not

interchangeable

with necessity,

consider the following example: Goldbach’s conjecture that

every even number greater than two is

the sum of

two prime

numbers is either true or false.

Furthermore, if

it is true

(false),

than

it

is

presumably

necessarily

true

(false).

62

 
 

 

 
 

However, none of

us

can be said

to

have any a priori

Knowledge about the answer to this question.

As Kripke puts

it,

we don’t know whether the conjecture is true or false,

so we certainly don’t know a priori whether it

is true or

false.

Having given some intuitive appeal for distinguishing

a prioricity and necessity, Kripke goes on to suggest that

some statements (such as Goldbach’s conjecture) are

necessary and a posteriori, while others

(such as

the

horsepower example below) are contingent and a priori.

In

giving examples he brings up discussion of possible worlds

and transworld identity.

Possible worlds are stipulated by

descriptive conditions we associate with them (and are, in a

sense, says Kripke, no more than counterfactual situations).

Transworld identity is

the identity of an object across

possible worlds.

So,

if we say that Nixon might have lost

the

l9o8 Presidential election, we will be asserting a

necessary or contingent statement depending upon what

singular term is used and what is meant to be referred to.

That is,

some referring terms

(for Kripke,

all names,

but

not all definite descriptions) are what Kripke calls rigid

designators,

i.e.

terms

that designate the same object in

every possible world in which that object exists.

‘Nixon’

is

a rigid designator because it refers to the same object

in all worlds in which that object exists, but ‘the winner

of

the

1968 Presidential election’ is

not a rigid

designator.

Clearly, Nixon c0uld not be anyone other than

 
 

 
 

Nixon, but the winner of

the 1968 Presidential election

could have

been

(and in some possible worlds,

is) someone

other than Nixon.

Now,

this notion of rigid designation and the earlier

distinction between a prioricity and necessity come together

in

the

following way.

Some statements,

such as

identity

statements involving names, are necessary and a posteriori,

while other statements,

such as ones involving (some)

definite descriptions, are

(or could be) contingent and a

priori.

For example, to say ‘Siddhartha is Gautama Buddha’

is

to make a statement that is known a posteriori.

That is,

we

come

to discover that ‘Siddhartha’ refers

to

the same

object as

does

‘Gautama Buddha’.

(Or,

‘Hesperus’ and

‘Phospherus’, we find out,

refer to

the same object, even

though we might believe them to refer to different objects.)

However,

if names are rigid designators, then the same

object is picked out in all possible worlds in which that

object exists.

So ‘Siddhartha’ picks out the same object in

all possible worlds and ‘Gautama Buddha’ picks out the same

object

in

all possible worlds.

Consequently,

if

the

identity is true, then (since the same object is picked out

in all possible worlds for each name) the identity is

true

in all possible worlds

(i.e.

the

identity

is necessarily

true).

The truth of

the statement ‘Siddhartna is Gautama

Buddha’ is (epistemically) a posteriori and (metaphysically)

necessary.

On the other hand, the statement ‘One horsepower equals

64

 
 

 

 

 

_I_

the power required by one horse to displace 150 pounds by

220 feet in one minute’ is contingent and a priori.

For

Kripke,

such a statement is contingent because it simply

fixes the reference of ‘horsepower’ (and given a different

horse in different possible worlds,

the actual value might

have been different than the value - 550 foot—pounds/second

- in this world).

The value in the actual world was picked

out by a particular accidental property (the distance a

certain mass could be displaced in a certain amount of time

by

a certain horse), and so

the statement is contingent.

Nonetheless, the statement is (now) known a priori.

This is

so because once the reference has been fixed, the definition

holds.

We don’t (need to) appeal to any evidence to know

the truth of the statement.

Returning,

then,

to

Searle’s

criticism

that

the

cognitive difference between a=a and a=b cannot be accounted

for unless names have senses, we see that Kripke disagrees.

Both identities, if true, are necessarily true - assuming we

are dealing with names

(i.e.

rigid designators).

On

the

other hand, while a=a is known a priori, a=b is known a

posteriori,

accounting

for

the

apparent

cognitive

difference.

However,

Kripke has not introduced senses in

accounting

for

this

cognitive

difference,

hence

Searle’s

criticism fails.

Criticism (2):

"Kripke’s distinction between speaker

reference and semantic reference is bogus."

Again, the

reason Searle makes

this claim is

that

he

sees

it

as

65

 

 

 
 

assuming that in referring we have both a general intention

to

refer

to

a certain object (whenever the designator is

used) and a specific intention to refer to a certain object

(on

a given occasion).

Searle finds

this assumption

questionable, because, he says, we have no such intentions

for definite descriptions.

Kripke anticipates this Searlean objection and responds

(to some extent, anyway) in a footnote (1977, p.273, note

22).

What he

is trying to show is that reference is not

subjective.

That is, it is true that we as speakers mean to

refer to a certain object when we use a referring term, but

obviously we cannot refer to whatever we want when we use a

referring term.

A referring term has a history and is

objective in

the sense that within a language a given

speaker’s

use

of

the

term must conform to

the other

speakers’s uses

of

that

term.

For Kripke,

the notion of

what words can mean and refer to is given by the conventions

of our language and what words mean and refer to on a given

occasion is determined by

these conventions together with

the

intentions

of

the

speaker and

various

contextual

features.

(This issue of objectivity versus subjectivity of

reference will be dealt with in some detail later, so I will

only say at this point that I think is Kripke is right.

It

is not clear to me, however, that Searle is lobbying for, or

is committed to, subjectivity of reference.

His objection

seems to be more that on a given occasion we do not have any

general intentions to refer in Kripke’s sense.)

66

 

 

 

——__——,

Kripke

continues

explicating

his

distinction

by

considering just the sort of cases that Searle mentions

(e.g. the ‘the King’ case).

Fearing the secret police,

people may well use

‘the King’ to refer to a usurper or,

humoring a lunatic,

call Smith ‘Napolean’.

Further,

such

cases shade into ironic or ‘inverted commas’ cases, such as

‘He’s

a "fine friend"’

(spoken ironically) or ‘"Napolean"

has gone to bed’ (spoken of Smith).

The point of all these

cases is that the more contextual phenomena one includes in

the notion of speaker’s reference (such as irony, linguistic

errors,

deliberate misuses

of

language),

the

further one

gets from any connection of

the notion with semantical

matters.

Kripke does not deny,

then,

that many of

these

Searlean features are present,

and important,

in ordinary

discourse, but he does deny that they play the role in

semantic analysis that Searle seems to think they do.

Additionally,

it

seems

to

me,

we could construct

intuitive counter-examples to Searle’s objection.

It seems

we often ‘mistakenly’ refer to someone (e.g. call someone by

the wrong name, or believe someone’s name is Smith when it

is

in fact Jones).

In such cases we are often corrected

(and even correct ourselves):

‘You mean Jones, don’t you?

Smith was

in Cleveland on

that day’.

Our response often

exhibits our acceptance of a speaker reference-semantic

reference distinction (e.g.

‘I’m sorry, I meant to say

"Jones",

not "Smith"’

or

‘I didn’t know her name was

"Jones"; I thought it was "Smith"’).

Such examples seem to

67

 

 

  

________7

me to indicate that to whom I believe I am referring and to

whom the (use of the) name refers is not always identical.

(I should add

that while I think Kripke is right on this

point, I don’t think he has said anything to refute Searle’s

‘infinity of descriptions’ objection.

The force of Kripke’s

point and

of Searle’s objection lies

on

their respective

notions of intention.

Kripke makes use of this notion, but,

I think, does not flesh it out.

A major difference between

Searle and Kripke, it seems to me, is that Searle places far

more emphasis

on

the role of intention in reference than

does Kripke.)

Criticism (3):

"There are intuitive counter-examples

showing that the causal theory of reference fails

to give

necessary conditions for successful reference and fails to

give sufficient conditions for successful reference."

The

case of reference to abstract or to ‘future’ objects is said

to demonstrate the failure to give necessary conditions for

reference and the case of Madagascar is said to demonstrate

the failure to give sufficient conditions for reference.

As noted before, Kripke (at least) does not see himself

as offering a

theory wherein necessary and sufficient

conditions are stipulated.

To criticize him for not having

done what he didn’t attempt to

do seems puzzling.

Of

course, a Searlean reply could be that not only has Kripke

not given necessary and sufficient conditions for reference,

but - as the examples above illustrate — he can’t, and this

is clearly a signal that the causal theory is weak.

68

 

  

  

—_——————————j

In

the addenda to Naming and Necessity Kripke

 

explicitly responds to the Madagascar case.

Says Kripke:

...a present intention to refer to a given entity

(or

to refer fictionally) overrides the original

intention to preserve reference in the historical

chain

of

transmission.

The

matter

deserves

extended discussion.

But

the phenomenon

is

perhaps explicable in terms of

the predominantly

social character of

the

use

of

proper names

emphasized

in

the

text:

we

use

names

to

communicate with other

speakers

in

a

common

language.

This character dictates ordinarily that

a speaker intend to use a name the same way as it

was

transmitted to

him;

but

in

the ‘Madagascar’

case

this social character dictates

that the

present intention to refer to an island overrides

the distant link to native usage. (1980, p.163).

Making this all clear, says Kripke, is a problem for further

work, but minimally a present intention to use a name for an

object and a present belief that the object is the only one

having a certain property must be distinguished.

Certainly,

however,

to

say

that

the problem has

not been thoroughly

worked out does not entail, contra Searle, that it can’t be.

With

respect

to

the

objection

that

reference

to

abstract entities illustrates the failure of the causal

theory to provide necessary conditions for reference, Kripke

(and others)

have

said practically nothing.

There are

a

variety of responses

that could be made.

One could deny

that the supposed names

of abstract entities are ‘really’

names.

For example, one could deny that

two’

is

a name.

\

Such a denial might take the form of claiming that all such

‘names’ are really disguised descriptions.

So,

one might

insist that

two’ can be translated into set~theoretic

\

notation and defined in such a way as

to eliminate the

69

 

 

  

\

‘name’.

Even if this were to work for numbers (and it’s not

obvious that it would),

it seems much less intuitive that

this move could succeed for reference to ‘future’ entities.

For example, a recent groom might exclaim, ‘The curtains in

Larry’s room have got

to go!’ where Larry is,

as yet,

an

unborn (unconceived?) son.

There is,

at

the

time of

the

utterance, no Larry in which ‘Larry’ is grounded.

Or,

in

the

initial

planning

stages

of

the

marketing and

manufacturing of

the K car, someone might say, ‘Good idea,

Noseworthy!

The K car could save this company!’

The K car

might be, at the time of the utterance, nothing more than an

idea in Noseworthy’s mind or the K car might be only in the

initial blueprint stage.

In any case, it is not at all

clear that ‘Larry’ or ‘the K car’ (where I take ‘the K car’

to function as a name, not as a definite description) can be

dismissed as disguised descriptions.

A causal theorist

might reSpond to such cases of supposed reference by taking

a

nominalistic

stance

with

respect

to

the

entity

in

question.

That is,

one could deny that such entities are

real and so treat reference to them in the same manner as to

other fictional entities,

say,

by appealing to Donnellan’s

notion of

a ‘block’.

So, just as supposed reference to

Santa Claus resulted in

a block, supposed reference to the

number two or to the K car might result in a block.

In each

case,

statements

containing

the

‘names’

can

still

be

assigned truth values even if

there are not really objects

being referred to.

(This tack could apply not only to

70

 

 

 

 

 

 

‘future’ entities, but to numbers as well, by insisting that

what we refer to are not numbers, but numerals (or

to

specific tokens).)

The cost of taking such a nominalistic

position, however, might be too high (because of other

difficulties) such that this sort of move would be

unsatisfactory.

Another response to Searle’s objection might be found

in Salmon (1981),

though his comments were not directed to

this issue.

In looking at the notion of rigid designation,

Salmon despairs that there are several distinct notions of

rigid designation that can be found in the literature.

The

most familiar notion is

that a designator is

rigid if

it

designates

the

same

thing with respect

to every possible

world in which that thing exists.

Salmon suggests that this

notion can be taken in

two ways.

One interpretation of

a

rigid designator Salmon dubs

a persistently rigid

designator.

A designator is persistently rigid if

it

designates

the

same

thing with respect

to every possible

world in which that thing exists and which designates

nothing with respect to possible worlds in which that thing

does not exist.

Another interpretation of

a rigid

designator Salmon dubs an obstinately rigid designator.

A

designator is obstinately rigid

if

it designates the same

thing with respect to every possible world, period.

Obstinate designators, Salmon tells us, "may be insensitive

to the question of whether their designate exist in a given

possible world;

they designate the same thing with respect

71

 

 

 

 
 

to every possible world, whether that thing exists or not."

(1981, p.34).

Further, Salmon contends that Kripke may very

well intend for his notion of rigid designation to be that

of obstinate rigid designation, and points to the following

note in Naming and Necessity as evidence:

 

Concerning rigidity: In many places, both in this

preface and

in

the

text of

this monograph, I

deliberately ignore delicate questions arising

from the possible nonexistence of an object. ...I

say that a proper name rigidly designates its

referent even when we speak of counterfactual

situations where that referent would not have

existed.

Thus the issues about nonexistence are

affected. (1980, p.21, note 21).

The

point

of

this,

with respect

to

the Searlean

objection of reference to abstract or ‘future’ entities, is

that

if

we

take

names

(of numbers,

say)

as obstinate

designators, then there is no problem for the causal theory.

Reference can be made whether or not the name refers to

anything which exists in the world,

if

the name designates

obstinately.

Since the actual world is obviously a possible

world, and if obstinate designators designate their objects

in all possible worlds whether or not the object exists in

those worlds,

then,

if names of numbers are obstinate

designators,

they designate their objects even if they

don’t exist.

For Salmon, names of numbers (or of numerals)

do designate obstinately (at least for the causal theory).

He states:

As

is most obvious in

the clearest case of

a

nondescriptional singular term - the free variable

whose denotation is fully and directly determined

by an assignment of values to variables - there is

no reason to suppose that the denotation of

a

nondescriptional term should change with respect

72

 

 

 

 

to different possible worlds.

...Proper names do

not denote by,

so to speak, rummaging through the

individuals in a possible world seeking anyone or

anything uniquely meeting certain conditions.

Hence,

there

remains

no

reason

to

expect

denotation to vary from one world to the next.

Instead, one may think of the denotation as being

fixed for all time and for all possibilities by an

initial dubbing ceremony, or an assignment of

values to variables, etc. (1981, p.33, note 35).

Since the actual world is a possible world, there is also no

reason to think that proper names denote in it by "rummaging

through

individuals...seeking anyone

or

anything

uniquely

meeting certain conditions."

If this is so, then reference

to abstract or ‘future’ entities should be

no problem for

the causal theory of reference.

Another response to Searle’s objection might be — and

this

seems

to

be Kripke’s response - that the initial

grounding of the causal chain may well be achieved by means

of a description which fixes the reference of a name.

Once

the reference of a name has been fixed by the use of a

description,

then future uses of

the name will link up

to

that baptismal event.

I say that this seems to be Kripke’s

response because of his comments like the following:

Neptune was hypothesized as

the planet which

caused such and such discrepancies in the orbits

of certain other planets.

If Leverrier indeed

gave

the name ‘Neptune’

to

the planet before it

was ever seen,

then he fixed the reference of

Neptune

by

means

of

the

description just

mentioned.

At that time he was unable to see the

planet even through a telescope. (1980, p.79, note

33).

A page later Kripke notes that the reference could be fixed

even if the description used turned out to be false, as when

‘Phospherus’ had its reference fixed by the description

\

the

73

 

 

  

morning star’ (which is false because Phospherus (=Venus) is

not a star at all).

By allowing that the reference of

a

name can be fixed by a description,

then,

there is no

problem for

the causal theory to refer to abstract or

‘future’ objects.

Since these were the sort of cases which

Searle proposed as counter-examples to the causal theory as

being

incapable

of providing necessary conditions for

successful reference, it seems that the criticism fails.8

Criticism (4):

"Kripke’s (and Donnellan’s) position is

really a descriptivist one."

Searle claims that the ‘causal

chain of communication’ is simply a characterization of

parasitic cases of reference seen from an external point of

view.

That is, for Searle, the causal (or historical) chain

plays no explanatory role in an account of reference (or, in

Kripke’s account of reference).

What does matter in an

explanatory role is the transfer of intentional content from

one speaker or speech act to another.

Reference is secured,

says

Searle,

under Kripke’s account in virtue of

the

intentional content in

the mind of

the speaker who uses a

particular

designating

expression.

After

an

initial

baptism, subsequent ‘intentional contents’ are parasitic on

the prior ones for achieving reference.

Of course, an

omniscient observer could give an historical (causal)

account of

the chain of reference, and hence describe

historically (causally) how a later use of

the expression

refers, what

secures reference is

the descriptive,

intentional content.

74

 

 

  

Now, even if Searle were correct that this is ‘really’

what is going on with Kripke’s account, it is a funny sort

of criticism.

That is, he seems to be saying that Kripke’s

account is fine as it stands, but that Kripke is mistaken in

thinking he has offered anything new, anything contrary to a

descriptivist theory.

Rather, he has wrongly identified or

overemphasized one aspect of

his theory as unique and

essential.9

However,

it does not seem that Kripke would

agree with Searle’s assessment of Kripke’s view.

While

admitting that intentionality plays a role in reference and

even that the descriptivist view might suffice for

explaining reference in limited contexts (such as an initial

baptism), Kripke clearly denies that the descriptivist

account is adequate.

His view is not descriptivist because

he insists that in general

our reference depends not just

on what we think ourselves, but on other people in the

community" (1980, p.95).

It is simply false that for Kripke

reference is secured in virtue of the intentional content in

the mind of

the speaker who uses a particular designating

expression.

Indeed, one of the points of making the speaker

reference-semantic

reference

distinction

is

to

show

that

reference is not dependent upon the intentional content in

the mind of

the speaker who uses a particular designating

expression.10

Conclusion

Just as

in chapter 1 above I was not trying to claim

that Searle is right and Kripke wrong when it comes to

75

 

  

 

explicating reference, so now I am not trying to claim that

Kripke is right and Searle wrong.

What I have tried to do

is

to fairly present both the cluster theory and the causal

theory, criticize each from the other’s perspective and

respond to

those criticisms from within the

two theories.

That is,

I have tried to capture (in an admittedly glossing

manner) the debate between the received theories of

reference as I see it and I have tried to let the authors of

those

theories

speak for

themselves

as

fully as

possible

(given what they have said in print).

I see the debate between the two theories as a family

squabble.

It seems to me that both theories are right in

some respects and wrong in others.

The important issues to

me,

though, are the shared assumptions and commitments

underlying the

two theories.

It strikes me

that both

theories are committed (in

a sense

I hope

to explain)

to

essentialism, realism and a correspondence theory of truth.

In

the remainder of

this work I want to address just what

these underlying positions are and to show how both the

cluster theory and the causal theory are committed to them.

I will then suggest a theory of reference (incorporating,

among other things, what I see as being right about the

cluster and causal theories) which does not share these

underlying commitments.

76

 

 

 
 
 

 

NOTES

1It is noteworthy that at the ‘initial baptism’ stage

the description theory, for Kripke, might be feasible.

He

states in a footnote: "...the primary applicability of

the

description theory is

to cases of initial baptism." (1980,

p.96, note 42).

Two things, however, must be emphasized.

First, the description used is not synonymous with the name

it introduces but rather fixes its reference.

Second, most

cases of initial baptism are not of

the sort which

originally inspired the description theory.

Usually a

baptizer is acquainted in some sense, according to Kripke,

with the object he names and is able to name it ostensively.

(In a later addendum, Kripke remarks that these brief

comments on initial baptism give a somewhat oversimplified

picture.

He stresses here that "there need not always be an

identifiable initial baptism" (1980, p.162), though cases of

such a lack of an identifiable initial baptism are rarer for

proper names than for other cases, such as natural kind

terms.)

2Devitt is somewhat vague here,

perhaps unavoidably.

He states: "...we insist only that the object be in the same

very general category as

it

is

taken to

be.

There is

an

element of arbitrariness in our determination of these

categories."

(1981,

p.63).

I

find

this

quite unhelpful.

Devitt claims that if we believe we are referring to a cat,

but the ‘cat’ is

in fact a mongoose, then we have referred

to

a mongoose, whereas

if

the

‘cat’

is

a shadow,

then we

have referred to nothing.

Devitt strikes me

as rather

cavalier about this issue, and it

is one that I see as

important.

I will return to it later in discussing the

underlying commitments of

the alternative

theories of

reference.

3Others, e.g.

Evans (1973) and Steinman (1982), have

made similar criticisms.

Kripke himself acknowledges that

in his ‘better picture’ one might never reach a set of

necessary and sufficient conditions.

In fact he states: "I

want to present a better picture without giving a set of

necessary and sufficient conditions for reference." (1980,

p.94).

However, other proponents of the causal theory, e.g.

Devitt,

seem

more enamored with establishing such

conditions.

4Searle does not need abstract or ‘future’ objects to

show that the necessity condition is faulty.

He gives

another example:

"I can,

for example, refer to M Street in

Washington [D. C. ]

simply because I know

that

there

is

in

that city an alphabetical sequence of

street names,

‘A’,

‘B’,

‘C’, etc.

I needn’t have any causal connection with M

Street to do that." (1983, p.239).

(Of course, we must also

assume that Searle knows the alphabet and that he knows that

the street names go

up

to,

at least, M Street, say, by

 

77

 

 

 

 

knowing that there is an N Street.)

5To my knowledge no one has responded in print directly

to

Searle,

though,

as noted below,

causal

theorists

have

considered some of the criticisms in some form or other.

6Donnellan considers several objections

to

rule

(R),

which are not essential to

the present point, which is

to

show how Donnellan could respond to Searle’s criticism.

Suffice it

to say that Donnellan considers himself to have

successfully defended rule

(R)

and

his

position from

the

objections that he himself raises.

7For Devitt, a sentence is true if and only if it is m—

true for every structure that corresponds to

it

(i.e.

for

every

structure

in

which

each

of

the

sentence’s

designational

terms partially designates

the object it m-

designates and applies to the objects it m-applies to).

8There are

many

cases of

reference being fixed by

a

description,

especially

for

‘theoretical’

entities.

For

example, only very recently have scientists at CERN claimed

to have ‘seen’ the ‘top’ quark, although it had been dubbed

as such, predicted to exist and sought for years.

9This statement should be tempered somewhat.

Searle

does see differences between the descriptivist view and the

causal view.

To

the extent that Kripke’s view works, says

Searle,

it works

because it

is descriptivist.

Among the

differences, Searle lists three that he sees as noteworthy:

(l) for the causalists the transfer of intentionality in the

reference chain is essential for proper names, for the

descriptivists it is incidental; (2) for the causalists the

only

intentionality which

secures

reference

is

that each

speaker intends to refer to the same object as the previous

speaker, for the descriptivists this is not the case, other

‘intentionalities’

are

important;

(3)

for

the

causalists

parasitic identifying descriptions are

paramount,

for

the

descriptivists they are not.

Whether or not Searle is

correct in his assessment here (and I don’t think he is)

is

not important for the present concerns.

10This point, just as some earlier ones related to

criticisms of Searle’s view,

involves the issue of private

reference versus public reference, which will be discussed

later.

78

 

 

   

 

CRITIQUE OF THE RECEIVED THEORIES

Language is

a kind of code dependent upon the life rhythms

of

the species which originated the language.

Unless you

learn those rhythms, the code remains mostly unintelligible.

(Frank Herbert, Whipping Star)

79

 

 
 

 

 

INTRODUCTION

In

this second part I want to expose and analyze

underlying commitments of

the received theories of

reference.

In particular I want to argue that both the

Searlean cluster theory and the Kripkean causal theory are

committed to - or, at least, presuppose - certain views with

respect to

the similarity relation and a principle of

individuation.

Commitments to certain views there are

exhibited in subsequent commitments to essentialism, realism

and a correspondence theory of truth.

However, before

turning to the task of laying out what I take these various

notions to

be and showing how the received theories are

committed to them,

I want to look more at the received

theories as

they have been presented in

the first two

chapters.

I want to argue that even prior to considering

any underlying commitments, there is reason to believe that

the two theories are not nearly so far apart as the debate

in

the literature would seem to indicate.

I also want to

make some initial observations and criticisms about the

received theories before going into

the more basic

observations and criticisms which I believe are lethal to

those theories.

First a review.

Minimally, much of the focus and force

of

the

criticisms

by

the causal theorists of

the cluster

80

 

 

 
 

theory of reference (and of

the descriptive theory in

general) is that it seems to allow a speaker’s intentions to

play too key a role in the business of reference.

As Kripke

says, "It seems to be wrong to think that we give ourselves

some properties which somehow qualitatively uniquely pick

out an object and determine our reference in that manner."

(1980, pp. 93-94).

Kripke’s speaker reference-semantic

reference distinction and his proposed counter-examples to

Searle’s position are designed to address this very point.

Reference is not the sort of thing, say the causalists, that

a speaker secures simply by uttering an expression and

intending to pick out some unique object.

Rather, reference

is

a public,

historical phenomenon.

Merely

intending

to

pick out X while saying ‘Y’ is not enough to refer to X

(especially if other speakers have used ‘Y’

to pick out Y

and ‘X’ to pick out X).

Whether we like it or not, ‘deel’

has a history which is intimately tied up with deel and not

with Schmidt.

On the other hand, Searle has claimed that the causal

theory is nothing more

than a descriptivist theory in

disguise.

The key feature of

the theory — that which

secures reference for the causal theory -

is

the intention

to refer to

the same object as that from whom the speaker

learned or acquired the name.

Granted, one could describe

an historical chain of the uses of a name, and so this is a

public phenomenon, nonetheless, a speaker’s ability to refer

to

an object by

(the use of)

a name depends upon the

81

 

 

 

 

speaker’s intention to in fact designate the same object by

that name as did the person from whom the speaker learned or

acquired the name1 (otherwise the speaker’s present use of

the name would not become part of the historical chain, but

would be part of some other historical chain stretching back

to a different object or to nothing at all).

On the surface,

then,

the two theories seem clearly at

odds.

The Kripkean counter-examples show - or are intended

to show — that the descriptions which a speaker associates

with a name may be totally inadequate for or even irrelevant

to picking out the actual referent of a name.

Indeed, it is

held that a speaker could legitimately use a name to refer

to an object and yet have no (true) descriptions associated

with that name which would uniquely pick out that object.

The Searlean counter-examples show - or are intended to show

- that an historical chain linking a speaker’s use of a name

and an object is inadequate as an explanation or theory of

successful reference.

It is held that there are legitimate

cases of successful reference (such as to abstract objects)

for which historical

chains

are

irrelevant

(and

even

nonsensical).

For Searle it is

the case that at least one

member of

the

disjunctive

set

of

descriptions associated

with the use of a name with an object must be true of

the

object for reference to successfully occur; for Kripke this

is not the case.

Despite

the apparent fundamental

differences

between

these two theories, I want to claim that they are much

82

 

 
 

  

closer in agreement than they at first appear.

Much of the

dispute seems to

be centered (at least implicitly) on

the

issue of public versus private reference.

That is, as noted

earlier,

Kripke has accused Searle of,

in effect,

seeing

reference as a private affair for any given speaker.

I

think that given much of what Searle has said Kripke is

right,

but I want also to accuse Kripke of the same thing.

In spite of Kripke’s insistence that reference depends "on

other people in the community, the history of how the name

reached one, and things like that" (1980, p.95), the causal

theory is just as compatible with, and in the manner in

which it has

been presented by

its adherents, just as

committed to private reference as

is

the cluster theory.

Before directly supporting this claim, however,

I want to

present some initial observations and criticisms of the

received theories, which I hope will give a better

background for a discussion of public and private reference.

83

 
 

 

  

_

NOTE

10:,

in

the case of baptisms, a speaker’s ability to

refer to an object by a name depends upon the speaker’s

intention to designate that object by (the use of) the name

(e.g. for Devitt, by having the mental representation of the

object).

84

 

 
 

CHAPTER THREE

CONVERGENCE OF THE RECEIVED THEORIES

Are

the cluster and causal

theories

as divergent as

they are usually portrayed as being?

I think not.

As

I

will argue later, they share fundamental underlying

commitments,

but even

on

a

more

‘surface’

level,

the

theories do not seem so far apart.

What has been said to be

unique to the cluster theory is that speakers have a set of

identifying

descriptions

associated

with a

name,

the

disjunction of those descriptions being analytically true of

the object referred to

by

the name.

However,

in several

places Searle remarks that, as far as he is concerned "the

issue is not really about analyzing proper names in words at

all." (1983, p.233).

Some cases, for example reference for

children (who often learn names before other expressions),

are such that the only ‘identifying description’ a speaker

might have

that he associates with the name is simply the

ability to recognize the object.

A young child, saying

‘Mama’ might well have

no descriptions at

all associated

with the name, and indeed ‘Mama’ might have no significance

to the child except when Mama is in sight and can be pointed

to.

For Searle, "pointing is precisely an example that fits

[the descriptivist’s]

thesis,

since pointing succeeds only

in virtue of

the intentions of the pointer." (1983, p.233).

85

 

 

 

 

That is,

for Searle, what is fundamental for reference is

that the speaker intends to pick out an object uniquely by

the use of a name, and, apparently, as long as the intention

is there,

the ‘description’ used to pick out an object may

not have to be verbal.

For Searle, ostension seems to count

as a means of picking out an object.

This sounds very much

like Kripke’s view that descriptions can be used to fix the

reference of a name and that, a la Devitt, in the presence

of an object we can name an object by having the appropriate

‘mental representations’ while uttering the name.

In fact,

ostension has

seemed

in some ways as much a paradigmatic

example of

the strength of

the causal theory as

of the

weakness of

the cluster theory.

Naming by ostension has

been seen as

a clear case in which a name is used but no

description (or disjunction of descriptions) associated with

the name is either necessary or sufficient for reference to

occur.

Now, we see Searle arguing that naming by ostension

is paradigmatic of the cluster (descriptivist) theory!

This

issue of ostension points to a focus of commonality for the

cluster and the causal theories: intentionality.

Two

of

Searle’s

seven

rules

of

reference

employ

intention in successful reference: Rule 5

[S intends that

the utterance of

R will pick out or identify X

to

H]

and

Rule 6 [S intends that the utterance of R will identify X to

H by means of H’s recognition of S’s intention to identify X

and he intends this recognition to be achieved by means of

H’s knowledge of the rules governing R and his awareness of

86

'

 r
l 

C].

A key element of Searle’s theory is that a speaker

intend to pick out an object by

(the use

of)

a name.

(Kripke’s objection is

that this

is

not sufficient to

guarantee successful reference,

because

the speaker might

intend to pick out one object by a name but in fact might

pick out a different object by that name.)

As just noted,

we see that for Searle this intention to refer to an object

by (the use of) a name does not entail that any verbal

descriptions must be employed by

the speaker; pointing is

fine.

So, cases where the speaker supposedly associates no

descriptions with the name

(as,

for example,

the cases in

chapter one) are no trouble for Searle because in those

cases

rules

5 and

6

of his

theory are violated (i.e.

the

intentional element of

the speech act of referring is

missing).

My neighbor, then, does not refer to deel simply

because she uttered the name ‘deel’.

The fact that

the

name ‘deel’ has an historical chain stretching back to an

initial use of

the name (to refer to deel) is irrelevant.

Furthermore,

just as

for Kripke,

for Searle,

intending to

refer to an object by (the use of) a name is not sufficient

to actually refer

to

that object.

If none of

the

 

descriptions in the disjunctive set of descriptions which a

speaker

associates

with

a

name

uniquely

pick

out

an

(otherwise

independently

located)

object,

then merely

intending to pick out

that object by

(the use of)

a name

does not secure reference.

I

do not and can not,

for

Searle,

refer to Aristotle by uttering ‘Aristotle’

simply

87

 

 
 

because I intend to refer to him.

I must have at least one

description (albeit not necessarily a verbal one) which is

true of Aristotle in order to pick him out (i.e. the primary

aspect under which reference is made must succeed).

In chapter one I presented two cases for which I said

that the cluster theory denies

(but

the causal theory

asserts) reference takes place.

The first case was my

neighbor complaining to her friend, "My weird neighbor kept

me up last night yelling something about ‘that damn turtle’

or ‘deel’ or something like that.

I had to pound on the

wall to get him to quiet down."

The second case was a

stranger asking me, “Who is deel?" after hearing me mention

deel’s name.

For both cases I claimed that under Searle’s

theory reference does not occur (and the reason why is that

in both cases the speaker did not intend to refer).

I also

claimed

that

under

the

causal

theory

‘devoid

of

descriptivist assumptions’ reference does occur.

I now want

clarify those claims.

First, under what I will call a

‘pure’ cluster theory of reference (i.e. a theory in which a

speaker’s intention to uniquely pick out an object by a name

is

a necessary and sufficient condition for reference

to

that object to occur), both my neighbor, in the first case,

and

the stranger,

in

the second case,

would not have

referred

(for

the

reasons

I

stated

in

chapter

one).

However,

Searle

is

not

a

‘pure’ cluster theorist in

that

sense.

While I

do think that he would claim that

in

the

first case, my neighbor did not refer to deel, I think in

 

 

 

 

the second case, he would allow (insist?) that the stranger

did refer to deel.

This is borne out by his claim that

"there are

uses

of

names where one

is almost totally

dependent on other people’s prior usage

to secure

reference."

(1983,

p.261).

In such a case,

the stranger

could secure reference through my use of the name ‘Gddel’ to

refer

to deel.

(This

move

by

Searle

brings up

two

important points.

First,

it makes his position sound more

and more like the causalist view.

Second, it points to the

issue of

the publicity of reference borrowing and reference

in general, an issue which will be explicitly addressed in

chapter four.)

Besides clarifying my initial strong claim about

the

cluster theory on

the

two cases

(viz.

my neighbor and the

stranger), I want to clarify my claims about the causal

theory with regard to these cases. Under what I will call a

‘pure’ causal theory of reference (i.e. a theory in which a

causal connection between an object and (the utterance of) a

name

is

both

necessary and sufficient for successful

reference),

both my neighbor,

in

the first case,

and

the

stranger, in the second case, would have referred (because

the historical, causal chain linking their use of

the name

‘Gddel’

with Gddel would have been in effect).

However,

neither Kripke nor Devitt are

‘pure’

causal

theorists

in

this sense, since both,

it seems, would deny (at least for

the

first

case)

that reference

has occurred.

Kripke

explicitly states

(in

his

"rough statement"

of

a

"better

89

 

 

 

 

picture" of reference): “When the name is ‘passed from link

to

link’,

the receiver of

the name must,

I think,

intend

when he learns it to use it with the same reference as the

man from whom he heard it."

(1980,

p.96).

Devitt is not

nearly so straightforward as Kripke about admitting

intention into the causal theory.

In fact, Devitt seems to

go to great lengths to avoid committing the causal theory to

intentions.

He states:

It

is tempting, and I think correct, to say that

when a person uses a name he (mostly) intends to

designate whatever was designated by

the person

from

whom

he

borrows

his

reference.

...Such

intentions seem as much in need of explanation as

designation

itself.

I

do

not

talk

of

these

intentions but capture the intuitions in another

way:

the tokens of the name a person produces are

tied

to

the

token from which he borrowed his

reference by causal links (of

the sort discussed)

between his mental representations of

the object.

(1981, p.138).

So, even though not actually using intentionalistic

language, Devitt nonetheless wants and needs to capture the

intuitions of such language in order to guarantee that

different speakers are in fact referring to the same object

I

when using a given name.

1

My point here is not to debate

whether talk of intentions is necessary in dealing with

reference, but

to

indicate

that advocates of

both

the

cluster theory and the causal theory invoke it

(or,

at

least,

invoke

the same intuitions that names are not used

helter skelter in referring).

Not only does Kripke seem to indicate that a necessary

condition for successful reference is the employment of the

‘proper’ intention (to refer to the same object by a name as

90

d
y
v
r
A
i
fi
r
V
:
;

_
,

.

l
l

)

 

:
I
"

 

 
 

the person from whom one acquires the name),

but he also

seems to see

the cluster theory as appropriate in certain

contexts, such as in the fixing of reference (say, during a

baptismal event).

In Kripke’s own example of the naming of

Neptune he acknowledges that there may be cases where the

cluster-of—descriptions view seems to make sense.

Both of

these considerations (the requirement for

Kripke of the ‘proper’ intention to refer and the admission

of the possibility of using a name to fix the reference a la

Searle) indicate that the causal theory and cluster theory

might be closer in agreement than initially thought.

But

there is

more.

As noted in chapter one, Kripke concedes

that there may be cases in which ‘X’ is used to refer to Y.

(For example, if someone asks,

‘What did Aristotle have in

mind here?’,

the speaker is referring (or,

for Kripke,

referring)

to whomever wrote the passage even if

it

turns

out not to have been Aristotle.)

With that concession,

Kripke remarks that the cluster theory would be "applicable

in a wider class of cases" than he originally thought (even

though he

still believes

it

to

be ultimately false).

Whether or not the cluster theory is false, the point here

is

that again Kripke seems to be softening his blanket

opposition to the cluster theory.

Another sign of

the convergence of the cluster and

causal

theories

is

Kripke’s

response

to

the Searlean

counter-examples to the causal theory involving reference to

abstract and

‘future’

objects.

Recall, Kripke’s apparent

 

 

 

 

 

way out of

the difficulty of referring to such objects for

the causal theory is

to allow that descriptions can fix the

reference of a name even if there is no causal hook-up to an

object.

Reference is secured, then, but not in a ‘purely’

causal way,

and

indeed Kripke’s response seems

to

fit

in

quite easily with Searle’s theory.

Devitt, too,

in a brief discussion of reference to

‘future’ objects sounds like Kripke (and like Searle).

Says

Devitt:

Consider,

for

example,

the

situation where

there exists a plan or blueprint for a ship or

building.

We could introduce a name for that

future object and it would seem to function like

an ordinary designational name.

This requires a

small modification in our theory:

the network for

a name may be grounded initially not in the object

itself but in some plan to produce the object

according

to certain specifications.

Later,

of

course, the network becomes grounded in the object

so produced.

Those who make

these claims about naming

future objects [viz.

that the causal theory can’t

account for reference to such objects] are struck

by the oddity of now naming, say, the first child

born in

the twenty-first century.

The oddity

arises, I suggest, because we could not have any

particular object in mind using the names

so

introduced.

This is striking because we mostly do

have an object in mind in using a name.

However,

sometimes we don’t... . (1981, pp.59-60).

This "small modification" of

the theory sounds more like a

major

concession

to

the

cluster

theory

than

a

minor

adjustment

in

the causal

theory.

Now, what is

the point of all of

this?

The point is

that even at

the level of what the cluster and causal

theories present explicitly as their positions, they are not

so far apart.

Both Searle and Kripke assert that intending

92

 

 

  

to refer to a given object by the use of a name (and indeed

intending to refer to

the same object as

the person from

whom one acquires

the name)

is

a necessary condition for

successful reference.

Both deny that such intention is

a

sufficient condition for successful

reference.

Both see

ostension as paradigmatic of a correct theory of reference.

Both see

the

use

of

reference fixing descriptions as

legitimate modes of referring.

Kripke’s insistence that we

"use names to communicate with other speakers in a common

language" (1980, p.163) sounds like Searle’s claim that we

"need to make repeated references to

the same object, even

when the object is not present, so we give the object a

name" (1983, p.231).

Both deny (as in the ‘Madagascar’

case)

that an historical link between an object and a name

is sufficient to secure reference.

Both deny

(as

in

the

reference to abstract objects) that a causal link between an

object and a name is necessary to secure reference.

With all of

these similarities between the

two

theories, we end up getting comments like the following:

Since the point of having proper names is just to

refer to objects, not to describe them,

it often

doesn’t really matter to

us much what descriptive

content is used to identify the object as long as

it identifies the right object, where the "right

object" is just the one that other people use the

name to refer to. (1983, p.250).

There may be cases...where some man really gives a

name by going into the privacy of his room and

saying that the referent is to be the unique thing

with certain identifying properties. (1980, p.94).

The

firsc,

‘anti-descriptivist’,

quote

is,

of

course,

Searle’s and

the second, ‘pro-descriptivist’, quote is,

of

93

 

 

  

course, Kripke’s.

But if these two theories actually share

many common features, what, then, is the difference between

them?

In the way it has been presented in the literature

(and primarily from the causalist perspective) the

difference is that the cluster theory is supposedly

committed to private reference and fails to account for the

historical, public nature of reference.

I now want to turn

to

this issue and I want to claim that even here the

two

theories agree; they both at bottom allow private reference.

To the extent that they do,

the distinction between the

theories is blurred even more and the received theories of

reference look more and more like a received theory of

reference.

94

NOTE

1Notice

the philosophical gymnastics

in which Devitt

engages to avoid talk of intentions.

Designation involves

three

types of links in

the designating-chain: (1)

groundings,

(2)

abilities

to

designate,

and

(3)

communication situations in which abilities are passed on or

reinforced.

The grounding consists in the person coming to

have "grounding thoughts" about that object as a result of

the act of perceiving the object.

"A grounding thought

about an object

includes

a mental representation of

that

object brought about by an act of perception.

The thought

is

one which a speaker of

a public language would express

using a demonstrative from

that language...

."

(1981,

p.133).

About the ability to designate, Devitt states that

his strategy is to tie an ability to an object and a term in

virtue of their role in bringing about the relevant mental

representations.

We have (at least) two abilities:

abilities in thought (which are basic) and abilities in

speech (or writing, etc.).

An ability to designate an

object in thought is to have thoughts including tokens which

are grounded in the object.

Now, if, as Devitt claims, the

notion of intention is

in as much need of explanation as is

the notion of designation, then I want to say likewise for

mental representation and his conception of thoughts.

While

he never gives an analysis of what a mental representation

is, he does say the following about thoughts:

A thought consists of an attitude to

a

sentence in

the language of thought.

I continue

my former practice of ignoring the fact that most

of

the thoughts we have we never entertain [?!];

also of using the word ‘thought’ to refer to the

sentence only...

.

Now a thought token will be

‘about our cat’ in the appropriate sense if it

includes a token -

a mental representation - that

designates

her.

That

token

in

thought will

designate her if she is

in fact the object in

which the token is grounded. (1981, p.131).

I,

for one, do not find this a great improvement over talk

of

intentions.

(Devitt’s prohibition against talk of

intentions is later relaxed, or forgotten, in his discussion

of fictional discourse.

He introduces an

‘F operator’ to

indicate

sentences

that

"can

be

true

even

though

the

contained singular

terms

are empty"

(1981,

p.130).

What

determines whether a given sentence token is or is not an F-

sentence (i.e. one with an F operator)?

Says Devitt: "That

is determined by what the speaker had

in mind,

meant,

or

intended." (1981, p.181).)

95

 

  

CHAPTER FOUR

PRIVATE REFERENCE

What I mean by ‘private reference’ will, I hope, become

clear in the ensuing discussion.

Initially, what is meant

here by

the notion of private reference is that a speaker

can refer to an object without any necessary conceptual or

linguistic link to other speakers, i.e.

that a speaker can

refer in a private language.

So,

a speaker can refer to X

(e.g.

Neptune)

simply by,

say, uttering a designating

expression (e.g. ‘the object which caused such and such

discrepancies in the orbits of certain planets’) and having

the intent to pick out X.

Before showing that both the

cluster theory of

reference and

the causal

theory of

reference accommodate (and,

in the way they are presented,

presuppose) private reference, more needs to be said about

the notion of a private language.

There is

a vast literature on the notion of a private

language1 and along with this literature has come various

conceptions of what a private language is

(or could be).

One conception is that a private language is one that, as a

matter of fact, only one person speaks, even though it could

be taught to others.

(Many of

us

as children have

‘invented’ our own private languages - which we quickly

96

 

 

 
 

taught to and conversed in with only our best friends.)

Another conception of a private language is one that no

other speaker could understand and that could not be taught;

in principle, only the speaker can understand this language.

A third conception of a private language is one that is (and

can be) understood only by the speaker because it

is

a

language used to talk about ‘private experiences’ (e.g.

the

speaker’s particular feelings, moods, etc.).

The first

conception of

a private language is,

I

think,

not

philosophically interesting, and is not the sense of a

private language which I claim is incorporated with the

cluster and causal theories of reference.

The second

conception of a private language also is,

I think, not

philosophically interesting because it seems to be simply a

glossing over of

the third conception.

I say this because

when one investigates why a language is

(would be)

in

principle unteachable, it turns out that it is because the

terms

in

the language (are meant to) refer to private

experiences.

This is the notion which Wittgenstein has when

he asks:

But could we also imagine a language in which

a person could write down or give vocal expression

to

his

inner experiences - his feelings,

moods,

and the rest - for his private use? - Well, can’t

we do so

in our ordinary language? - But that is

not what I mean.

The individual words of this

language are to refer to what can only be known to

the person speaking;

to his immediate private

sensations.

So another person cannot understand

the language. (1953, #243).

Now, I want to extend this notion of a private language

which refers to private experiences to include referring in

97

 

 

 
 

general.

That is,

I want to argue that the primary private

element in such a language isn’t so much the objects

referred to (e.g. feelings, moods, etc.), but is the act of

referring (and of naming).

It is the act of, or, perhaps,

the experience of,

referring (naming)

that is private,

in

the sense of being unique to the speaker.

What is private

with respect to my talk about my pain is not so much that it

is

talk about my pain, but that it is talk about my pain.

When Wittgenstein says:

"The individual words of

this

language are to refer to what can only be known to the

person speaking;

to his immediate private sensations," the

»emphasis (at least for my notion of privacy) is on ‘refer’

not on ‘sensations.’

(This will hopefully become clearer

below, as I criticize the notion of private reference.)

There does not seem to be any clear significant difference

between a speaker who "gives a name by going into the

privacy of

his room and saying that the referent is

to be

the unique thing with certain identifying properties" when

that unique thing is a headache or when that unique thing is

Neptune.

The

point

is

that

a

speaker,

outside of

a

linguistic community,

can refer to objects.

This

is

the

sense of private reference which I want to claim is not only

part of the cluster theory of reference, but is also part of

the causal theory of reference.2

I want to suggest two arguments showing that the Searle

and Kripke are both committed to private reference.

The

first argument involves the nature of

the examples and

98

 

  

counterexamples which they employ;

the second argument

involves

the inclusion of intentionality as

a (necessary)

component of reference for both of them.

The very examples and counterexamples which both Searle

and Kripke employ, I want to claim, show their commitment to

private reference.

Consider Searle’s discussion of

reference to Aristotle and his insistence that at least one

description in

the disjunctive set of descriptions

associated by the speaker must be true of an object for the

name to refer to

that object.

As Kripke correctly points

out, this view is clearly mistaken, for there are many cases

in which the descriptions in the disjunctive set all seem to

fail, yet our intuitions are that reference has successfully

occurred.

While Searle’s primary—secondary aspect

distinction may well work as a defense in some cases, I

think that it,

too, fails because the problem is simply

moved back one step.

Reference at the level of the primary

aspect can still suffer the Kripkesque problems.

The

important point here for my concerns is

that Searle’s

original position stresses the descriptions associated by a

single, given speaker and his counter-move to Kripke’s

objection is still at the level of reference made by a

single speaker without any apparent need to appeal to

the

language community.

While this is not a definitive argument

showing that Searle is committed to private reference,

it

is, I think, a fairly reliable indication.

On the other hand, for all his talk of the publicity of

99

reference (e.g. semantic reference), Kripke, I think, also

shows an affinity to private reference.

Consider his

Feynman example.

The example works against Searle only to

the extent that Searle is not allowed to appeal to other

speakers in the community for reference to successfully

occur, as

I mentioned in chapter one.

That is, Kripke

thinks

that a given speaker might still refer to Feynman

even though none of the descriptions associated by that

 

speaker with the name ‘Feynman’ are true of Feynman.

This

strikes me

as

a clear indication

that Kripke intuitively

sees reference as something that can be done by a single,

given speaker independent of

a language community.

While

his example works against Searle, it fails if we allow the

speaker to appeal to the language community for reference to

occur.

I agree with Kripke that one might refer even though

all

of

the descriptions which one associates with a name

might fail to pick out the intended object of reference

(e.g.

all

of

the descriptions which I associate with

‘Feynman’ might fail to pick out Feynman).

However, I

disagree with Kripke if he thinks that one might refer even

though all

of

the descriptions which

the entire

language

community associates with a name might fail to pick out the

intended object of reference.

Quite simply, we can’t all be

wrong and still have reference occur.

I will try to defend

this claim throughout the remaining chapters.

My main point

here

is

to suggest that both Searle and Kripke

offer

examples which presume the plausibility of private reference

100

 

 
 

(and this, I suggest, weakens both their views).

I now want

to turn to the second major argument that Searle and Kripke

are both committed to private reference, viz. the appeal to

intentionality.

Intentionality

The very inclusion of intentionality as a (necessary)

component of referring seems to be a commitment to privacy

in the above sense, as it is obviously something that a

speaker can do by "going into the privacy of his room."

That is,

the ability to stipulate that a name refers to an

object by - at least in a baptismal event - uttering the

name and intending to pick out a unique object seems to be

something that can be done by

a conceptually isolated

speaker.

(I will discuss, at the end of this chapter,

whether intentionality is necessarily private.

That is,

I

will consider the objection (to my claims above)

that the

inclusion of intentionality is not a commitment to privacy

because there is a sense in which intentionality can be

public.)

My immediate concern is

to show that both the

cluster theory and

the causal theory are committed to

intentionality as a necessary component to referring.

First

Searle:

The capacity of speech acts to represent objects

and states of affairs in the world is an extension

of the more biologically fundamental capacities of

the mind (or brain) to relate the organism to the

world by way of such mental states as belief and

desire,

and especially

through action and

perception. (1983, p.vii).

Searle seems to be saying here that reference (as a kind of

101

 

 

 

 

underlying mental (or brain) capacities.

This is just

another way of saying that for reference to occur (the

‘proper’)

intention must be present, and this seems borne

out in the seven rules of reference which he enunciates (see

chapter 1).

Furthermore,

by couching reference

in

these

terms it seems clear that reference can be and is a private

matter.

That is,

because it

is related to

the individual

organism’s mental (or brain) states,

there is no necessary

connection with it and other organisms (or,

in the case of

reference,

other speakers).

There is nothing in Searle’s

account of

reference

to prevent

speakers

from,

as

Kripke

says,

giving

"ourselves some properties which somehow

qualitatively uniquely pick out an object and determine our

reference

in

that

manner."

(1980,

p.94).

A

further

elaboration of

Searle’s

view

of

intentionality will

help

corroborate this claim.

[I want to say that to

the extent

that a theory allows private reference, it

is committed to

private reference.

I mean this

in

the sense that to

the

extent that a theory allows talk of (or quantification over)

particular (kinds of) entities, it

is committed to (the

existence of) those entities.]

According

to

Searle,

intentionality

(intending,

intentions)

is

one

form of Intentionality (with a capital

‘I’).

Other forms include believing, hoping, fearing,

loving.

The commonality here

is

that Intentionality is

a

property of (many) mental states and events by which these

mental states and events are directed at

(or about or

of)

102

 

 

 

 

mental states and events are directed at (or about or of)

objects and states of affairs in the world.

Every

Intentional state

(including intending) consists of

a

representative content in a certain psychological mode.

So,

if a speaker intends to refer to Aristotle by using the name

‘Aristotle’ in the statement, "Aristotle taught Alexander,"

then the representative content of the Intentional state is

the state of affairs of Aristotle having taught Alexander

and the psychological mode is intending (rather than,

say,

wishing).

Where

the

Intentional

content

(i.e.

representational content) is

a whole proposition and where

there is a direction of fit (i.e. the Intention must ‘match’

the world),

the Intentional content determines the

‘conditions of satisfaction’ (i.e.

those conditions which,

as determined by the Intentional content, must obtain if the

state is

to be satisfied).

In the example above, Aristotle

must have taught Alexander for the state to be satisfied.

With regard

to

reference,

for Searle,

there must be

some Intentional content associated with the name for

the

speaker for reference to occur.

Also, as mentioned in

chapter three, Searle allows the Intentional content to

be

non—linguistic.

For example,

one can refer by way of

"names, descriptions, indexicals, tags, labels, pictures, or

whatever." (1983, p.259).

The crucial point is that a

certain mental

(brain)

state

is directed toward the world

(so

that animals,

too,

have Intentional states).

The

important point for

the present discussion on private

103

 

 

 
 

reference is that Searle associates Intentional states

(including

intentions)

with

the mental

(brain)

states of

individual organisms.

This, coupled with the claim that the

Intentional content can be non-linguistic, seems clearly to

commit Searle to private reference.

For under his view, a

speaker (or,

for that matter, a non-speaker) can refer by

having certain conditions

of

satisfaction satisfied by

a

certain Intentional content.

The resulting Intentional

state is identifiable with a particular brain state of the

organism.

Where this commits Searle to private reference is

exhibited in

the following example (a

la Anscombe(1957)).

Suppose that a particular mental (brain) state could be

created

in

an aborigine

(or even,

say,

in

a

cat)

by

stimulating a certain area of the brain (say, by implanting

an electrode).

The particular resulting mental (brain)

state is the intention to go to the bank and withdraw money

from the person’s (organism’s) account.

Under Searle’s

view,

it seems

that the aborigine (cat) does have the

intention to withdraw money from the bank account because

the appropriate mental (brain) state is realized. (Or,

similarly, an intention to refer to Aristotle by ‘Aristotle’

could be induced.) Intentionality, and subsequently

reference, under Searle’s view,

then, allow private

reference.

(In fairness

to Searle,

he insists

that his View of

Intentionality does

not allow Intentional contents

to

determine

their

conditions

of

satisfaction

in

isolation:

104

 

 

 

 

Rather,

Intentional

contents

in

general

and

experiences in particular are internally related

in

a holistic way to other Intentional contents

(the

Network)

and

to

nonrepresentational

capacities (the Background).

They are internally

related in the sense that they could not have the

conditions of satisfaction that they do except in

relation to

the

rest of

the Network and

the

Background. (1983, p.66).

While at first glance this might seem to rescue Searle from

a commitment to private intentionality (and private

reference), on further examination it

is evident that the

commitment is

still

there.

The commitment to privacy is

still there because there is nothing in the introduction of

a Network or Background that demands publicity.

The Network

and Background can still be restricted to a single organism,

unrelated

to

any linguistic community or public

‘form of

life’.

For example, the Background only requires that there

be a non-Intentional capacity to identify the object in the

world and the Network only requires internal coherence, both

of which are consistent with the aborigine (cat) having an

entire Network of Intentional states induced by electrodes.

(The aborigine (cat) could be a brain in a vat!)

Nothing

about the requirement of

a Network or Background rules out

the possibility of an aborigine (cat) intending to withdraw

money from the bank account (or intending to propose in

marriage

or

intending

to

refer

to

Aristotle

by

‘Aristotle’).)

Now, not only is Searle committed to private reference,

but to

the extent that Kripke allows a cluster theory view

of reference fixing descriptions (as

in the Neptune case),

105

 

 

 

 

he, too, is committed to private reference.

In spite of his

insistence that reference depends not on just what we think

ourselves, "but on other people in

the community,

the

history of how the name reached one, and things like that"

(1980,

p.95),

he

does allow private reference at

the

baptismal point.

There is nothing in Kripke’s account of

reference to prevent speakers from intending to pick out an

object by (the use of) a name when a name is being

introduced (where this intending is

in

the sense just

discussed above).

The aborigine (cat) in the above example

could just as easily be

a speaker in Kripke’s view as

in

Searle’s.

This is especially true of Devitt’s conception of

the

causal theory of reference.

In fact

his discussion of

grounding thoughts (which result from the act of perceiving

an object and include mental representations of that object)

sounds very much like Searle’s portrayal of reference as an

ability based on mental (or brain) states.

Devitt

explicitly allows for private reference:

I

see no reason to deny that beings which do

not

speak public languages

could have grounding

thoughts.

If

they do have them,

then

they must

have appropriate mental representations.

Call all

these representations of

the object in grounding

beliefs "demonstrative representations."

The act

of perception leading to these representations

defines a mode of presenting the object.

It leads

to

an ability to designate made up

of

these

grounding thoughts.

Underlying the demonstrative

representations in those thoughts is a d-chain

grounded in the object in virtue of which they

designate the object. (1981, p.133).

Both the cluster and causal theories of reference,

106

 

  

then, are committed to the possibility of private reference

(in the sense of reference within a private language).

What

is

the significance of

this?

Two points are significant.

First, as I will argue now,

to the extent that they are so

committed, the theories are weakened.

Second, the issue of

private reference points to a deeper underlying commitment

of

the theories, a commitment to

a particular view of

the

similarity relation and of individuation.

(In the next

chapter I will develop this second point to try to show that

the cluster and causal theories share the same position on

these two fundamental concepts.)

In criticizing the notion of private reference I want

to focus on

two points:

(1)

the plausibility of

a private

language (within which private reference is possible), and

(2) the role of intention in private reference.3

Contra Private Intentionality

The issue of the plausibility and even possibility of a

private language is an old one and I will not engage in a

thorough analysis of

it.

I agree with Putnam (1981)

that

the notion of a private language is

a self-refuting one and

that Wittgenstein gave conclusive arguments against the

possibility of a private language.

What is Wittgenstein’s objection to the possibility of

a private language?

In a sentence, it is that the notion of

rule-governance makes no sense within a private language and

since all language must be rule-governed,

there can be

no

private language.

As

we saw in the quote earlier,

107

 

 

 

Wittgenstein is talking about the possibility of making

reference to one’s inner experiences, one’s private

sensations.

Now, Wittgenstein is not denying that we have

sensations or feelings4 or even that we can name them.

Rather, he is claiming that naming - even naming feelings -

presupposes a context, or ‘stage-setting’, which is possible

only within a public language: "‘To give a sensation a name’

means nothing unless

I know already in what sort of game

this name is

to

be used." (1968, pp.290-291).5

So, naming

presupposes a stage-setting which presupposes a public

language.

Why the presupposition of a public language?

Because naming (and referring)

stand in need of agreement

and outward criteria, or rule obedience.

(Wittgenstein, of

course,

speaks

of

this

in

terms

of language-games.)

The

‘game’ we play with ‘toothache’, for instance, entirely

depends upon there being a behavior (or behaviors) which we

call

the expression of

toothache.

If someone claimed to

have an unbearable toothache and yet exhibited none of the

behavior which we have learned to associate with having a

toothache,

we would wonder whether the person was telling

the truth or understood what having a toothache is.

We

would,

after a while,

wonder if

the person knew what

‘toothache’ means

(or

if

the person knew what ‘toothache’

refers

to).

Behind this insistence on public,

behavioral

criteria is

the demand for rule obedience in order for

understanding and communication to be possible:

And hence also ‘obeying a rule’ is a practice.

And to think one is obeying a rule is not to obey

108

 

 
 

a rule.

Hence it is not possible to obey a rule

‘privately’: otherwise thinking one was obeying a

rule would be

the same as obeying it.

(1953,

#202).

If one could in fact have a private language, one could not

know whether or not one was following a rule,

for

the

concept of rule involves publicity.

Without rules of

application,

one cannot have a language,

if

for no other

reason than simply because one would not know from one time

to another whether one was giving the same name to whatever

one was naming or whether one was referring to the same

thing from one

time

to another.6

(I will return to this

point of ‘giving the same name’ in the next chapter.

It is,

I think, of paramount importance.)

A private language,

then, is no language at all, and to the extent to which the

received theories of reference permit it (and in fact are

committed to it), those theories are weakened.

I said before that there were two points about private

reference which demand attention.

The first is that private

reference in the sense of reference in a private language is

impossible.

The second point involves the role of intention

in reference.

It is to this point that I now turn.

Putnam (1981) gives an ‘abbreviated version’ of

a

Wittgensteinian argument that there is

no necessary

connection between names and objects and that intentionality

does

not enable

reference to succeed or occur.

Suppose,

Putnam says,

that

in

its crawling across

the sand an ant

traces a particular pattern of lines which we depict as

a

recognizable caricature of Winston Churchill

(or

that

the

109

 

 

 

ant traces a particular pattern of lines which we depict as

a recognizable shape

resembling

the

name

‘Winston

Churchill’).

Would we say that the ant has referred to

Winston Churchill?

Putnam says: no.

What may seem to be

missing for reference

to have occurred is

the element of

intention.

However, this will not do:

But

to

have

the

intention

that

anything...should represent Churchill, I must have

been able

to

think about Churchill in

the first

place.

If lines in the sand, noises, etc., cannot

‘in themselves’ represent anything, then how is it

that thought forms can

‘in themselves’ represent

anything? (1981, p.2).

Just as physical representations

(e.g.

lines in

the sand)

have no necessary connection to objects, so mental

representations

also

have

no

necessary connection

to

objects.

(Words are included here: a discourse might seem

to be, say, a description of a tree, but if

it was produced

by monkeys

randomly hitting keys

on

a

typewriter for

millions of years, then the words would not refer.) Now, the

point that Putnam is making here is that intentionality does

not yield the ability

to refer,

but actually presupposses

the ability to refer.

To intend to withdraw money from a

bank account (or

to refer to Aristotle by ‘Aristotle’)

requires the ability to refer to those objects or actions or

capacities which are the objects of the intentions.

This is

because, as Putnam puts it,

to have the intention that

something happen (e.g. money being withdrawn) or be

represented

(e.g.

Aristotle

being

referred

to

by

‘Aristotle’),

one

must already have

the ability

to

think

110

 

 

about (i.e.

refer

to)

the event

(or object)

in

the first

place.

(Furthermore, the ability to intend itself depends

upon a public ‘form of life’, as I will argue at the end of

this chapter.

This bears on the question of whether there

is any sense of intentionality which is not private and

subsequently on the role of intention in reference.)

If Putnam is right (and I think that he is), this is a

telling criticism of

both

the cluster and

the causal

theories, since, as we saw above, they are both committed to

intention as an integral component of reference.

(This is

true also of Devitt’s account of

a causal theory, even

though he tries to avoid talk of intentions.

His talk of

mental representations avoids neither the commitment to

intentions nor Putnam’s criticism.)

Another criticism of

the role of intentionality in

reference comes from Elgin (1983).

She argues that the

notion of intentionality is conceptually intermingled with

that

of

interpretation and

that

this

requires publicity.

Says Elgin:

...intending to produce an effect by getting one’s

audience to recognize one’s intention to produce

that effect is not, or at least is not always,

required

for reference.

Someone talking in his

sleep or under the influence of anesthesia or, for

that matter,

someone who

is unconsciously

‘thinking out loud’ may have no such intention.

Whether to say that such a person refers depends

on how his words are interpreted.

If we can make

no sense of his utterances we put them down to

inchoate

ramblings.

If

their interpretation is

straightforward, we

take his utterances to be

sentences, and the terms in them to refer.

It is

the

availability of

a

reasonable

interpretation

rather than the intention with which they were

produced that is crucial in deciding whether his

 

111

 

 

  

—_————7

words refer. (1983, p.17).

Now, Elgin’s comments clearly are directed against

Searle’s theory (especially his rules 5 and 6) of reference

as a speech act which is based on intentionality.

She

reiterates Kripke’s speaker reference - semantic reference

distinction by repeating

that we

can distinguish between

what a speaker intends to refer to and what a speaker

actually refers

to.

However,

this same point can be

directed against the

causal theory as well,

since,

as

we

have seen,

it,

too,

incorporates intentionality.

Nor can

the

causal

theory

be

defended

by

claiming

that

this

criticism

is

appropriate

to

speaker

reference

(where

intention does play the role that Kripke assigns it),

but

that it is not appropriate to semantic reference.

This move

does

not help because,

as Putnam’s criticism points

out,

words have no more necessary connection to objects than do

physical or mental representations.

Yes, words (names) hook

up

to objects, but not by themselves.

What counts as a

name, and the fact that something functions as

a name,

is

not inherent in

the name

(nor

in

the object).

To

extrapolate Elgin’s point,

interpretation not only

determines (at least in part) what is conceptually

legitimate about intention (and

indeed what counts as

intention), but also interpretation determines (at least in

part) what is conceptually legitimate about naming (and

indeed what counts as naming).

It is true that naming and

referring is public,

but

the causal theory’s treatment of

 

  

semantic reference (and reference in general) does not

capture

the proper sense

of publicity to account for

successful reference.

I will try to defend this last claim

further in the next chapter.

For now, I will consider the

notion of intentionality in a public sense.

Afterwards, I

will conclude the discussion of private reference by saying

that it seems

to

me

that both the cluster and the causal

theories are committed to private reference (at least in the

way that they have been presented).

Furthermore, they are

both weakened by

that commitment.

Finally,

the issue of

private

reference

points

to

an even more fundamental

underlying commitment of

both

the cluster and causal

theories.

When Wittgenstein objected to a private language

and

to private reference because one would never know

whether one was using the same name to refer to an object

(or experience) or was referring to the same thing from one

time to another, he was pointing out the dependence of

reference on both a principle of individuation and on a

conception of sameness (or similarity).

I want to argue in

the next chapter that the cluster and causal theories share

a commitment to a particular view of individuation and the

similarity relation, a view which I want to criticize.

What I have been arguing in this chapter is that both

the cluster and the causal theories of reference depend upon

intentionality for successful reference to occur.

In

addition,

I

have

been

arguing

that

the

sense

of

intentionality which both theories allow (are committed to)

113

 

  

is private.

Given this allowance of (commitment to) private

intentionality, I have been arguing that both theories allow

(are committed to) private reference.

To

the extent that

this is

so,

the theories are weakened.

A question quickly

comes to mind: Must intentionality be private?

I will argue

now

that intentionality does

not

have

to

be private.

 

However, as I have been arguing, both the cluster and causal

theories

allow

(are

committed

to)

private

reference,

so,

given the theories as they are presented, the possibility of

public reference does not rescue them.

Furthermore, as the

Putnam and Elgin criticisms indicate, even the possibility

of public intentionality does

not

rescue

the cluster and

causal

theories

(or any

theory relying on intentionality)

from

serious

difficulties.

The

immediate

consideration,

though, is of the possibility of public intentionality.

I have been arguing that both Searle and Kripke (and

Devitt)

allow

(and

are

committed

to)

a

sense

of

intentionality which allows

(is committed to) private

reference,

because

the

intentionality

itself

is

private.

However,

is

there

a sense in which intentionality is

(or

could be) public?

I want to argue: yes.

In fact, I think

that

the conceptions

of intentionality which Wittgenstein

(1953,

1958,

1967)

and Anscombe

(1957)

have

laid out are

examples.

In

the Philosophical Investigations,

section 337,

Wittgenstein remarks that an intention "is embedded in

its

situation,

in human customs

and

institutions."

We could

114

never,

for example,

intend

to steal second base if

there

were no game of baseball.

What is crucial for Wittgenstein

is

that intentions, like other ‘mental’ events (or

experiences or states)

only arise in and are only

intelligible within ‘forms of life’.

These ‘forms of life’

are known to us only through shared behavior and practices.

As humans we engage in certain behaviors and practices which

enable us to make sense of and interact with the world.

(As

members of particular cultures or, say, religious groups or

social groups we engage in certain behaviors and practices

which enable us to make sense of and interact with the world

in other ways.)

Our ability to make sense of and interact

with the world requires our participation in some ‘form of

life’,

known

through

publically

observable

behaviors

and

practices.

As mentioned above, even ‘private’ experiences

such as pain stand in need of outward criteria and can make

sense only within the context of outward (behavioral)

criteria.

As Quine (1985) says, echoing Wittgenstein:

Someone observes my joyful or anxious expression,

or perhaps observes my gratifying or threatening

situation itself, or hears me tell about it.

She

then applies the word ‘joy’ or ‘anxiety’.

After

another such lesson or two I find myself applying

these words to some of my subsequent states in

cases where no outward signs are

to

be observed

beyond my report itself.

Without the outward

signs

to begin with, mentalistic terms could not

be learned at all. (1985, p.6).

Not only do we learn about intentions on the basis of

communal behavioral criteria, and

so epistemically

intentions are basically public, but we can only make sense

of what intentions are

(ontologically) in

a public sense.

115

 

  

In The

Blue Book Wittgenstein insists that intentions are

     

not particular states of mind or mental processes:

Suppose I had made a move in chess and

someone asked me "Did you intend to mate him?", I

answer "I did", and he now asks me "How could you

know you did,

as all you knew was what happened

 

within you when you made

the move?", I might

answer "Under these circumstances this was

intending to mate him." (1958, p.147).

Several points are worth noting here.

First, as noted

above,

intentions make sense only within public,

communal

behaviors and practices.

One can intend to make a move in

chess only if

there

is such a practice as playing chess.

Second,

in

the

context

laid

out

in

this

quote,

what

intention is is dictated by

the situation (e.g.

under

the

circumstances,

intending to mate his opponent in chess

is

nothing more

than

the

behaviors

and practices associated

with and which characterize that situation).

Does this mean

that intentions are identical with behaviors?

No,

it means

that intentions only make sense given shared behaviors and

practices.

While

intentions are

not behaviors,

they

are

possible only because there are (publically observable)

behaviors.

(Behaviors and practices are

the ‘natural

expression’

of intentions, for Wittgenstein, as can be

exemplified by looking at a cat when it stalks a bird.

We

say the cat has an intention only because of its behavior.

If the cat did not stalk the bird, it wouldn’t make sense to

say

that it intended to stalk (or catch or eat)

the bird.

Likewise, while a particular intention might not be

expressed behaviorally, the intention (and intention in

116

 

 

 

 

general) makes sense only because, as a rule, intentions are

learned and expressed behaviorally.)

Anscombe

(1957)

reiterates

and

expands

on

Wittgenstein’s remarks on intentionality.

How do we tell

someone’s intentions, she asks:

Well, if you want to say at least some true things

about a man’s intentions, you will have a strong

chance of success if you mention what he actually

did or is doing.

For whatever else he may intend,

or whatever may be his intentions in doing what he

does, the greater number of things which you would

say straight off a man did or was doing, will be

things he intends. (1957, p.8).

While recognizing that not all intentions are carried out

and that we often intend to do things without actually doing

them (and,

so, speak of intentions as being ‘inside us’ or

‘private’),

nonetheless these make sense and are possible

only because of larger public contexts in which we do learn

and express intentions behaviorally.

To say,

for example,

that only a particular person can know her own intentions

means nothing more

than

that

there comes a point where a

particular person can say,

"This

is

my intention," and no

one else can contribute anything to settle (or question) the

matter.

It does not mean, says Anscombe, that the person is

evincing a knowledge known only to her.

With Wittgenstein, then (and contra Searle and Kripke),

Anscombe maintains that having a particular ‘mental

representation’ or being in a particular ‘mental state’ is

not an adequate account of intentionality.

The aborigine

(cat)

with

the

brain implant does not intend

to withdraw

money from the bank because the appropriate public context —

117

 

 

 

including the historical context - is missing.

In a related

example, Anscombe states:

Let us suppose that the thought in [a man’s] mind

is ‘you silly little twit!’

Now here too,

it is

not enough that these words should occur to him.

He has to mean them.

This shews once more,

that

you cannot take any performance (even an interior

performance) as itself an act of intention; for if

you describe a performance,

the fact that

it has

taken place is not proof of intention; words for

example may occur in somebody’s mind without his

meaning them.

So intention is never a performance

in the mind, though in some matters a performance

in

the mind which is seriously meant may make a

difference to the correct account of the man’s

action - e.g.

in embracing someone.

But

the

matters in question are necessarily ones in which

outward acts are ‘significant’ in some way. (1957,

p.49).

Not only is Anscombe arguing against a private notion

of intentionality here, but she is also alluding to a point

made above by Elgin, viz.

the interrelation between

intention and interpretation.

Because intentionality makes

sense only within the context of ‘forms of life’, and

because we participate in ‘forms of life’ based on behaviors

and practices, we determine what intentionality is, and, in

the long run, what people’s intentions are.

It

is important to note that neither Wittgenstein nor

Anscombe is denying that we have intentions or denying the

meaningfulness of

talking about intentions,

so neither is

behavioristic in

a

‘strong’

sense.

However,

they are

insisting on a conception of intentionality which stands in

need

of

outward

behavioral criteria.

With regard

to

theories of reference (such as Searle’s and Kripke’s) which

require

intentionality,

they would rule

out such theories

118

 

 

 

which allow (are committed to) private intentionality.

As

noted above, Putnam’s and Elgin’s arguments against

intentionality

in general militate

against any

theory of

reference which requires intentionality - private or public

— as a necessary component.

With this in mind, we are now

ready to turn to the underlying elements of

the cluster and

causal theories,

and first to

the underlying principle of

individuation and similarity relation.

119

 

 

lSee especially Jones (1971).

NOTES

21 need to make an important distinction here.

In

response

to

the

following discussion of Searle’s

and

Kripke’s commitment to intentionality and private reference,

one might object that a commitment to intentionality, even

‘private’ intentionality, does not commit Searle or Kripke

to a commitment to private language.

The reason for this,

so

the objection goes,

is that speakers intend to refer to

public, not private, entities (i.e. entities which are ‘out

there’ independent of any speakers’s intentionality).

Being

committed to intentionality does not entail being committed

to private reference because the objects being referred to

are public.

I want to argue that this objection fails

because it rests on a particular conception of individuation

of

objects.

I will criticize

this conception in chapter

five

below.

For

now,

then,

I

want

to

critique

intentionality (or,

at least,

private intentionality).

After critiquing conceptions of individuation (and

similarity)

in

the next chapter,

I will return to

the

relation between intentionality,

private reference and

individuation.

At that point, after having made my argument

against

the

conception

of

individuation which

Searle

and

Kripke hold, I will flesh out my argument that the objection

nentioned here fails.

3Actually, my critique of the role of intention applies

not

simply

to private

reference,

but

to reference in

general.

4See Wittgenstein (1967, #487) and (1953, #244).

5See also Wittgenstein (1953,

#257, #290),

(1958,

p.69),(1967,#428)and<1969,#47zx

6Rhees (1954) makes the same point: "I know a headache

when I feel it, and I know I felt giddy yesterday afternoon,

because I know what giddiness is.

...But the identity - the

sameness - comes from the language... .But without language

I could not have told whether this feels the same, either;

if only because I could not have asked." (1954, pp.81,82).

120

 

 
 

_—__—#

L__ _._

CHAPTER FIVE

INDIVIDUATION AND SIMILARITY

As

we

have

seen,

reference,

at

least

in naming,

presupposes both a name and an object.

(of course,

this

does not mean that the object must exist; we have names for

fictional objects, ‘future’ objects, abstract objects, etc.)

We refer to

X by (the use of) ‘X’.

In considering the

plausibility of private reference, we saw that reference is

conceptually linked with interpretation and with public

criteria for successful reference.

Underlying these issues

is that of individuation and similarity.

At this point we

need to get clearer not only on this connection between

reference and individuation (and similarity),

but also on

just what individuation and similarity are.

Individuation and identity are

two sides of

the same

coin.

To ask what something is and to ask how something is

individuated are not only interrelated, but are essentially

to ask for the same information.

Identity might be said to

be

a metaphysical issue and individuation an epistemic

issue, but I do not think that characterizing them as such

differentiates them in any significant way.

Both are

questions of partitioning the world.

By saying,

‘This

is

what X is’, we are individuating X from Y; by saying, ‘This

is (an) X and that is

(a) Y’, we are saying what X is (e.g.

121

minimally something that

is not-Y).

The significance of

this with respect to reference is that reference only makes

sense within the context of an underlying conception of

individuation.

Our ability to refer to X by (the use of)

‘X’ rests on there being a prior understanding of what X is.

There must be (in some sense) an X for us to refer to X.

We

could not know whether we have referred to Glub by (the use

of)

‘Glub’

unless

there

were

a

thing which has

been

individuated as Glub.

Indeed we could not refer to Glub if

there were not an individuated Glub.

(Again, this does not

mean that Glub must exist;

Glub could be fictional or

‘future’

or abstract,

etc.)

For reference to occur,

individuation must already have occurred.

Similarity and sameness are also two sides of the same

coin.

To ask what makes this the same as that is the same

as

to ask what makes this similar to that.

In both cases

what is being asked for is criteria for identification.

To

say that X and Y have the same color is to say that they are

similar with respect to color.

With respect to ontology,

sameness is the limiting case of similarity:

to say that X

is the same as Y

is

to say that X and Y are similar in all

respects.

But notice that there seems to be no difference

between identity and sameness.

X’s identity with Y just is

X being the same as Y.

We are led, then, to the realization

that individuation and similarity are conceptually

interconnected (not an unwelcome result, as far as I am

concerned).

122

 

 

 

Now, what’s the point of all this?

As noted earlier, I

want to argue that the cluster and causal theories of

reference

are

committed

to

a

particular

view

of

individuation and the similarity relation.

I want to show

that they have a shared conception of these notions and it

is at this fundamental level that they are more similar than

different, that they are two versions of a received theory

of reference.

I

then want to argue against their shared

conception of individuation and similarity (and say more on

why I take these notions to be so fundamental, not only to

reference, but to many issues).

Individuation

In their discussions of reference, neither Searle nor

Kripke speaks explicitly about individuation.

Neither seems

to indicate that there is any question about or problem with

individuation.

That is, while both might question whether

this name picks out that individual, neither indicates any

question or problem with a principle of individuation.

Individuals are ‘out there’ so

to speak and reference is

a

matter of

‘the right’ hook-up of names to individuals.

Searle speaks of picking out "an independently located

object" as though there is no question or problem with

(conceptually)

locating such an

object.

The question of

what name picks out what individual is different than, and

rests on,

what individuals there are and how they are

individuated.

About the latter Searle says nothing.

About

the former he makes claims such as the following:

123

  

In

short,

if

none

of

the

identifying

descriptions believed to be true of some object by

the users of the name of that object proved to be

true

of

some independently located object,

then

that object could not be identical with the bearer

of the name. (1969, p.169).

Besides

statements

such as

this

one,

Searle’s

commitment to private reference indicates that he sees

individuation as

a straightforward matter,

and

the

‘fact’

about individuation is

that individuals are given in

the

world, independent of our concerns or interpretations.

For

Searle, there may be a question as to whether ‘Gddel’ refers

to deel or to Schmidt, but there is no question as to Gddel

and

Schmidt

being independently located

(or

at

least

locatable - in the sense of being capable of being located)

objects.

Kripke’s acceptance of the view that individuation is a

straightforward matter, essentially one of simply seeing or

coming to see what individuals there are ‘out there’, can be

seen in

the context of his discussion of possible worlds.

One of the issues connected with the postulation of possible

worlds is that of transworld identity.

That is, how do we

identify the same

thing across possible worlds?

What

properties must a thing have in different possible worlds in

order for

that

thing

to

be

the

same

thing

(e.g.

what

properties must a

thing have in different possible worlds

for that thing to be,

say, Nixon)?

In rejecting the idea

that questions about

transworld identification make

sense

only in terms of questions about component parts of an

object, Kripke says the following:

124

Theorists have often said that we identify objects

across possible worlds as objects resembling the

given one in the most important respects.

On the

contrary, Nixon, had he decided to act otherwise,

might have avoided politics like

the plague,

though privately harboring radical opinions.

Most

important, even when we can replace questions

 

about an object by questions about its parts, we

need not do

so.

We can refer to

the object and

 

ask what might have happened to it.

So, we do not

begin with worlds (which are supposed somehow to

be

real,

and whose qualities,

but

not whose

objects, are perceptible to us), and then ask

about

criteria

of

transworld

identification;

on

the contrary, we begin with the objects, which we

have, and can identify, in the actual world.

We

 

can then ask whether certain things Fight have

been true of the objects. (1980, p.53).

Kripke’s commitment is obvious here.

By saying that we

can "refer to the object and ask what might have happened to

it," and by saying that we begin with the objects "which we

have,

and

can identify," Kripke

is clearly implying

that

 

there

is

no problem of individuation in

this world.

His

later discussion of natural kinds reiterates this view.

Our

understanding of what gold ‘really’ is

(or what tigers

‘really’

are)

might change over time,

but

that

there are

natural kinds,

say,

elements

(or

species),

is not

in

question.

Gold is gold, and whatever it ‘really’ is,

it’s

that way independently of us.

So, says Kripke:

When I refer to heat, I refer not to an internal

sensation that someone may have, but to an

external phenomenon which we perceive through the

sense of feeling; it produces a characteristic

sensation which we call the sensation of heat.

Heat is the motion of molecules. (1980, p.129).

So, both Searle and Kripke are clearly committed to a

particular view of individuation, viz. one in which

individuals are given in

the world (i.e. independently

125

locatable and ‘really’ part of

the world independently of

any speakers of a language).

I want to argue now that this

‘realist’ conception of individuation is mistaken and

consequently both

the

cluster and causal

theories

of

reference are weakened by their commitment to it.

Goodman (1965) has argued - correctly I think - that

the

partitioning of

the

world

is

(at

least

in

part)

determined by our cognitive interests and is language-

dependent.2

Specifically, his argument is about legitimate

inductions based on lawlike generalizations, and hence about

the partitioning of the world in the sense of reference

classes.

I want to extrapolate from Goodman and suggest

that his argument applies to the partitioning of

the world

in

the sense of individuals as well.

That is,

the point

which he makes about reference classes (which I will review

below) can also be made about individuals, viz. which ones

there are is (at least in part) determined by our cognitive

interests and is language-dependent.

Goodman asks us

to consider a predicate ‘grue’, which

he defines in

the following way:

the predicate ‘grue’

"applies to all things examined before [time] t just in case

they are green but

to other things just in case they are

blue." (1965, p.74).

For example, we would say that an

emerald is grue iff

it

is examined before, say, January 1,

2000 and is green or it is not examined before that date and

is blue.

Up

to time t we have evidence that emeralds are

grue.

Assuming that all emeralds so far examined have been

126

green, then it follows that they have all been grue.

Before

5,

then,

an English speaker, knowing these facts about

emeralds, would make the inductive generalization that all

emeralds

are

green,

and

the English speaker’s prediction

that the next emerald examined (after i) will be green will

be based on this lawlike statement.

However, a ‘gruesome’

speaker, knowing these facts about emeralds, would make the

inductive generalization that all emeralds are grue, and the

gruesome speaker’s prediction that the next emerald examined

(after

t)

will be grue will be based on

this lawlike

 

statement.

The immediate point is

that the same evidence

(green emeralds examined before t) seems to confirm

equally

two different predictions, and,

in fact, contradictory

predictions (because the English prediction is saying that

the next emerald will be green, while the gruesome

prediction is,

in effect, saying that the next emerald will

be blue).

Says Goodman:

Thus although we are well aware which of the two

incompatible predictions

is genuinely confirmed,

they are equally well confirmed according to our

present definition [of confirmation].

Moreover,

it

is

clear

that

if

we

simply

choose

an

appropriate predicate, then on the basis of these

same

observations

we

shall

have

equal

confirmation,

by

our

definition,

for

any

prediction whatever about other emera ds

- or

indeed about anything else. (1965, p.74).

So,

says Goodman,

the original problem of justifying

induction (in a way, attempting to answer the question, ‘Why

does a positive instance of

a hypothesis give any grounds

for predicting further instances?’) has been displaced by a

new problem of defining confirmation (in a way, attempting

127

to answer the question,

‘What is

a positive instance of a

hypothesis?’)

and distinguishing confirmable and non-

confirmable hypotheses (in a way, attempting to answer the

question,

‘What hypotheses are confirmed by their positive

instances?’).

Past cases of green emeralds were used to

confirm

rival,

mutually

exclusive,

hypotheses.

Past

regularities of green emeralds seem to confirm the

prediction that

the next emerald would be green,

but

the

past regularities of grue emeralds doesn’t seem (to us) to

confirm the prediction that the next emerald would be grue.

However, the point of Goodman’s example is that saying that

valid predictions rely on past regularities without saying

which

past

regularities

is

completely

unhelpful.

"Regularities are where you find them, and you can find them

anywhere." (1965, p.82).

Now,

the

significance

of Goodman’s

argument

for

the

present concerns is

that if

he

is right that regularities

are where you find them and you can find them anywhere, then

the

conception of

regularities

(or

reference

classes)

as

being ‘out there’ in the world is mistaken.

We do not ‘bump

into’ regularities in the world, we make them up.

We find

what we’re looking for only after deciding what it is we’re

looking for.4

It

is

our cognitive interests that shape

what regularities (reference classes) there are, and, as the

grue example points out, what regularities there are depends

upon the language we speak.

A major issue,

of which Goodman’s grue example is an

128

instance,

in

the philosophy of science is whether or not

(and to what extent) facts are independent of

theory.

Goodman’s example is one example to argue that facts are not

theory-independent.

The debate about the theory-fact

distinction has usually focussed on the question of whether

or not observation is theory-laden.

It seems

to me that

while this particular question is significant with respect

to the theory-fact distinction, it is not crucial.

That is,

while I want to hold that observation is theory-laden (in

the sense that we observe individuals in the world and

individuation is,

I want to argue below, theory—dependent),

there is more to

the discussion about the theory-fact

distinction than simply questions of observation.

Even if

one

could

make

the

case

that observation

is

theory-

independent (so, for example, we all ‘see’ the same thing or

event),

there still remains the question of what this

observation means.

We

still must determine whether this

particular observation is

a relevant observation (that is,

if indeed we are observing what we think we are observing).

As Duhem (1962) argued, what is acceptable — and even what

is possible — as scientific fact (e.g. experimental results)

depends

on

theoretical

interpretation.

Without

some

theoretical justification for believing a given meter

reading as a relevant datum (rather than, say, an instrument

malfunction),

we would not consider

this

datum a genuine

‘fact’.

For example, without some theory of optics, we

would not

accept

the

claim

that

a

new

instrument,

a

129

‘fabriscope’, could allow the viewer to see through clothes

(much like the x-ray glasses that used to be advertized in

the back of comic books).

Indeed, much of the dispute about

Galileo’s use of

the telescope to support his claims about

heliocentricity centered on questions of the reliability of

this

new instrument.

The point here is

that even if

observation were theory-independent (say,

in

the sense of

sensations

that all observers would experience),

there

is

still

the

issue

of whether or not

these observations

constitute ‘facts’.

Questions about the theory-dependence

of partitionings is another version of this issue.

Consider

the following example

to

illustrate

the

interest- and language-dependence of world partitions (and

of individuals):

the (temporal) world is partitioned into,

say,

months

and seasons.

However,

it seems obvious

that

this partition is based on our interests.

For example,

seasonal divisions are (or were) based on farming or hunting

concerns.

Likewise, geographic divisions are determined by

our,

say, political interests.

Now,

it might be objected

that even if these examples do show partitioning to

be

interest- (and perhaps language-)

dependent,

they

are

artifactual and do not show that other, more ‘natural’,

partitions are interest- (and language-) dependent.

This

objection, I think,

is wrong.

Consider the partition made

in

the objection itself: artifactual versus natural.

The

objection seems to rest on the assumption that the very

division between what is artifactual and what is natural is,

130

so

to speak, natural; it is ‘out there’.

However, this is

far from obvious.

For example, through genetic engineering

new organisms can be (have been) created, but which are now

viable and reproductive.

Or,

there are many cases of

hybredized plants, which were ‘created’ by humans (and, so,

artifactual), but are ‘natural’, as they are reproductive.

These cases are meant to show that the partition into

artifactual versus natural is not a sharp one, so we cannot

point to what is ‘out there’ in nature to argue against the

interest-

(and

language-)

dependence of partitionings

(because what is ‘out there’ in nature, as opposed to what

is artifactual, is not clear).

There are other cases to illustrate the interest- (and

language-) dependence of partitionings.

Our taxonomic

divisions (kingdom, phyllum, etc.) are made on various

bases.

For example, divisions are made on the basis of

whether or not organisms are caryotic or not; divisions are

made on the basis of whether or not organisms are warm-

blooded; divisions are made on the basis of dentition;

divisions are

made

on

the basis

of pigmentation.

Even to

argue

that

these

features

(procaryotic versus eucaryotic,

warm-blooded

versus

cold-blooded)

are

really

\

out

there’

does not deny

that our taxonomic categories are interest-

(and language—) dependent.

We choose to consider, say,

dentition as a basis for partitioning rather than, say, eye

color.

We choose to classify an organism as a bird not

because

it

flies

(bats

are

not

birds,

but

they

fly;

131

   

 

 

 

ostriches are birds, but they don’t fly).

Instead, we

choose to classify an organism as

a bird on

the basis of

other criteria (say,

being feathered).

The point,

of

course, is that we choose to partition in certain ways and

based on certain criteria, and those partitionings both

reflect and are a function of our interests.

Another objection at

this point is

that

the examples

above have, at best, shown only that our partitionings are

theory-dependent,

not

interest-

(or

language-)

dependent.

It remains to be shown that theories are interest- (or

language-) dependent.

This issue, it seems to me, has been

at the heart of much (if not most) of

the debates in

the

philosophy of science for quite some

time

(at least since

Kuhn

(1962)).

There are,

of course,

numerous historical

cases which illustrate that theories have been proposed and

accepted on

the

basis

of particular

interests

(interests

both in a cognitive sense and in a moral, social, political

sense).

For example,

Gould

(1981)

has argued forcefully

that theories in intelligence testing have a long history of

being interest—based.

Harding

(1981)

has

offered

considerable

evidence

and

argumentation

that

various

antithetical approaches to social science research have

shared normative sexist assumptions.

Examples could be

extended ad nauseum to show that actual, historical cases

support the contention that theories are interest-dependent

(e.g. Lysenko, and Einstein’s remark that his physical

theories

followed from his epistemic and metaphysical

 

132

views).

Indeed,

the view that theories are not interest- (or

language-) dependent seems to me to rest on the assumption

that there

is

a sharp distinction between the context of

discovery and the context of justification.

That is,

one

might argue that it doesn’t matter whence our theories come,

what matters is whether or not they are good theories

(whatever ‘goodness’ means with respect to theories).

The

host of objections to and arguments against this distinction

are well-documented and don’t need to be repeated here.

I

do, however, want to make the following point: it seems

clear that theories are interest- (or language-) dependent

in

the

sense

that our cognitive

(or what McMullin (1983)

calls ‘epistemic’) values determine our theories.

That is,

features such as simplicity, generality,

testability, etc.

are features which we insist on in theories because we value

them

(in

a cognitive or epistemic sense).

To

the extent

that we demand reliability from theories, we demand features

such as testability, etc.

(Cognitive, or epistemic, values

are inherent in other aspects of science as well;

for

example, the demand for replicability of experiments.)

The

necessary connection between theories and cognitive values

seems uncontroversial.5

Given this point, it is not a great

leap to get to the point that ethical (or social, political,

etc.) values are inherent in theories.

Just as our theories

do not arise separate from our cognitive values, so the

argument goes,

our cognitive values do not arise separate

133

from

our ethical/social/political values.

Marxists

have

argued for

this claim for years.

Recently feminists have

developed strong arguments

that

the normative assumptions

and values

that underlie scientific methodologies are

predominantly based on male experience.

(Whether or not

cognitive values are a function of ethical/social/political

values is not crucial to my main point here, which is that

theories are interest- (or language-) dependent.

However,

if these arguments are sound,

then they add support to my

claim.)6

Now, even if the case can be made (as I have just tried

to do) that our partitionings are theory- and interest- (or

language-) dependent,

this

does

not say anything with

respect to individuation.

That is, even if Goodman is right

that reference classes are where we find

them and we can

find them anywhere, this says nothing about the individuals

which make up the reference classes.

They, it is generally

held, are ‘out there’.

Whether grass is green or grue, so

the claim goes, grass is ‘out there’ and individual blades

of grass are what they are.

I now want to argue that this

view is mistaken.

I want to claim that individuals are

where

you

find

them

and

you

can

find

them anywhere.

Individuation, I want to say,

is just as much theory- and

interest- (and language-) dependent as partitioning at the

level of reference classes (or predicates) is.

One argument for

the treatment of individuation in

a

‘non-realist’

way

can

be

extrapolated

from

Goodman’s

134

 

  

gruesome language and from Quine (1953).

In a context of

discussing non—referring singular terms, Quine suggests that

such terms can be translated into a description or even into

a predicate.

So, ‘Pegasus’ can be translated into ‘the

thing that pegasizes’ or simply ‘pegasizes’.

If Quine is

right, and if there is no reason to restrict his point to

non-referring singular terms (and I see no reason for such a

restriction), then this result coupled with Goodman’s point

of

the language-dependence of reference classes (which are

identified by predicates), seems to support the claim that

individuation, too, is language-dependent.7

Another, perhaps more intuitive, argument for a ‘non-

realist’

treatment of individuation comes from the

consideration of

the individuation of events.

We clearly

individuate one event from another.

8

For example, we speak

of

‘this coin toss’ or ‘that die roll’.

Many events have

even

been named,

‘the

battle

of Waterloo’,

‘the

1948

Presidential election’, and so on.

Now, it also seems clear

that the individuation of events is not merely a matter of

‘finding them out there in the world’.

We decide what

counts as

an event and hence what events there are.

This

can be seen (somewhat, at least) on an intuitive level with

the difficulty of saying what the parameters of a given

event are.

Exactly when was the 1948 Presidential election?

When did it begin?

On election day?

On the day that the

first candidate was nominated?

In asking, ‘When did it

begin?’ to what does the it refer?

(The answer cannot be:

135

 

 
 
 

 

to the 1948 Presidential election, because that is circular

and vacuous in this context.)

Or, when was the battle of

Waterloo?

Did it start when the first shot was fired?

When

Napolean advanced his troops past a certain point?

The

point of

these examples and questions is

that the

individuation of events is not a clearcut, straightforward

notion even at an intuitive level.

At a more technical level, Hanna (1969, 1978) has made

much the same point.

He distinguishes generic experiments

(or events)

from specific experiments

(or events).

By

a

generic experiment he means

"an empirical procedure which

can

be replicated by different experimenters at different

spatio-temporal locations - each particular replication

constituting a specific instance of the generic experiment."

(1978, p.533).

Scientific predictivity and explanation

refer, says Hanna, not to specific experiments, but to

generic ones.

So,

the scope of

the prediction of Jones’s

specific experiment conducted at this place and time under

these conditions is not limited to that specific experiment.

It covers every specific replication of

the generic

experiment.

The ‘problem’, though, is that it is not a

straightforward matter as to what counts as a replication of

that generic experiment.

Since predictions cover a

potential infinity of observations and predicting the

outcome of

a generic experiment involves

a commitment

to

corresponding projections for all replications of

the

experiment, it follows that scientific predictions can never

136

be established (or not) beyond question.

Says Hanna:

The relevance of any specific experiment, e.g., as

performed by Jones in laboratory X,

to prediction

(A) is to a great extent a theoretical issue.

The

outcome is relevant only if Jones has created the

appropriate stimulus situation: i.e., only if

Jones

has

replicated

the generic experiment.

But...this latter issue is not a straightforward

matter of

observation;

on

some

occasions

it

involves highly theoretical considerations. (1969,

p.312).

Consider a simple coin tossing experiment.

A coin is

tossed and

the outcome is

tails.

A prediction about the

outcome of this event only makes sense if the experiment can

be replicated.

The question, however, is: what is

the

generic experiment of which this is a specific instance?

We

need the generic experiment to be individuated before we can

say what this specific instance is an instance of, and the

individuation of

the generic experiment is what’s at issue

here.

I am arguing (as is Hanna) that the individuation of

the generic experiment is not ‘out there’.

We decide what

constitutes the generic experiment and that in turn

determines what counts as an instance of it.

To be able to

replicate the experiment, we first need to know what it is

we are attempting to replicate, and

that depends upon our

cognitive interests and concerns.

We decide what the

parameters of

the coin toss are (e.g. whether the toss

begins, say, when the coin loses contact with an other

object - like my hand — or when it reaches the apex of its

ascent or whatever).

If indeed replication is (at least in

part)

"highly

theoretical,"

then

this strengthens

the

argument that the individuation of events is interest-

137

 

 

 

dependent and language-dependent.

(This

point

of

replication as ‘non—realist’ is also relevant to the issue

of similarity, discussed below.)9

It seems, then, that the

individuation of events,

at least,

is language-dependent.

Is is also the case that the individuation of other kinds of

individuals is

language-dependent?

I want to argue: yes.

First,

following Quine’s

talk

of

translating

talk

of

individuals into talk of predicates, it seems that there is

no

reason

to

suggest

that

this

doesn’t apply

to

all

individuals (whether they are events or other kinds of

individuals).

While predicates such as ‘pegasizes’ or ‘is-

the-battle-of‘Waterloo’ sound a

bit unusual

(indeed,

more

gruesome than ‘grue’),

there is

no clear reason to forbid

them.

Furthermore, the distinction between objects and

actions

(i.e.

between event-individuals and

thing-

individuals) is not so sharp.

For example, storms seem

intuitively to

be events, not things, yet they are treated

linguistically on a par with things

(e.g.

both storms and

children can satisfy the sentence

x

is savage’).

Fists,

\

too,

seem to be actions (moving the hand into a particular

formation), yet ‘fist’ is

a noun, a grammatical feature

which we associate with things.

(What is

taken to

be an

event and what is taken to be a object varies among natural

languages.

See,

for

example,

Whorf

(1956),

on

such

differences between English and Hopi.)

This

points

to another argument

for

a

‘non-realist’

treatment of individuation.

Formally, individuals are

138

whatever can satisfy the values of individual variables

(e.g. whatever can replace the variable x in

‘x is red’).

Informally, individuals are whatever can be referred to by a

noun (e.g. whatever can be

the subject of

a sentence (in

English)).

As noted in the paragraph above, the distinction

between events and objects then becomes blurred.

Not only

can storms be referred to by nouns, but particularly severe

storms can be (and are) named (e.g. Hurricane Diana).

Also,

actions, such as running, become individuals when they are

referred to as gerunds (e.g. ‘Running is good for one’s

health’ or ‘Over-eating is unhealthy, but fun’.)

The

significance of

this

is

that if event-individuals are

individuated

based

(at

least

in

part)

on

our

theories/interests/language,

and

if

there

is

no

sharp

distinction between event-individuals and thing—individuals,

then there may be good reason to adopt the view that thing—

individuals

are

also

individuated

based

on

our

theories/interests/language.

With respect to individuals being whatever can replace

individual variables or whatever is referred to by a noun,

consider the following example.

‘July’, on these criteria,

is an individual.

For example, it is proper to say ‘July is

my favorite month’.

Clearly, however, July is not ‘out

there’.

What July

is

depends

on

our

theories/interests/

language.

July is nothing more than a set of temporal

coordinates.

But,

of

course,

we

determine

what

the

extension of

that

set

is.

Are

‘natural’

individuals any

139

different?

I

see

no reason to assume so.

‘Natural’

individuals are simply sets of space-time coordinates.

But,

then,

of

course,

we decide

the extensions of

those sets,

too.

We decide what set of space-time coordinates is this

bird (or,

for that matter,

is all birds).

We set the

parameters for what counts as

these

sets

of coordinates.

(An objection here is that I am confusing epistemic claims

with

metaphysical

claims.

That

is,

even

though

epistemically we decide

the parameters of

these sets,

nonetheless, metaphysically, these sets are independent of

us.

Of course, much of what I am arguing throughout this

entire book is that this distinction between epistemic and

metaphysical claims is bogus.

The entire work is directed

against this objection.)

I think, however,

that another, stronger, case can be

made

that individuation per

se

is language—dependent.

Perhaps the most intuitive notion of individuation (and of

individuals)

is

that

of

ostension.

The

use

of

demonstratives (e.g.

‘I mean this mess!

Clean it up now!’

 

or ‘Get a load of that guy!’) has been a paradigm case of

 

individuation.

Which mess?

This one.

Which guy?

That

one.

This mess and that guy are ‘out there’ if anything is.

The trouble here,

though, is that, as both Wittgenstein and

Quine have argued, even ostension is interest- and language-

dependent.

Wittgenstein offers several examples.

Suppose,

he says, that someone points to a pencil and says, ‘This is

tove’ or ‘This is called "tove".’

Does ostension work here?

140

 

 
 

 

No,

he

says,

because this could mean a number of

things:

this is a pencil, this is round, this is wood, this is one,

this is hard, etc.

Another example is ‘This is sosh.’

Wittgenstein states:

For example, from "This is sosh" you would not

understand the use of the word "sosh", though from

"This color is sosh" you would.

That is,

if a

person is to learn the meaning of a word from such

a definition he must already know what sort of

thing it stands for.

The word "color" already

fixes the use of "sosh".

The ostensive definition

is of use if you need to fill in only one blank.

(1979a, p.45).

Quine

(1960,

1969)

also criticizes the notion of

ostension as providing a clear means of individuation.

Suppose, he says, that a native points to an object and

says,

‘Gavagai.’

A non-native, on witnessing this, might

believe

that

the

native

has

pointed

to

a

rabbit and

‘gavagai’ can be translated as ‘rabbit’.

But things are not

so obvious for Quine.

It could be that ‘gavagai’ is more

nearly translatable as ‘rabbithood’ or perhaps ‘undetached

rabbit parts’ or even ‘rabbit stages’.

Says Quine:

The

only

difference

between

rabbits,

undetached

rabbit parts,

and

rabbit stages

is

in

their

individuation.

If you

take

the

total scattered

portion of

the spatiotemporal world that is made

up

of rabbits,

and

that which is made up

of

undetached rabbit parts, and that which is made up

of rabbit stages,

you come out with the same

scattered portion of

the world each of

the three

times.

The only difference is

in how you slice

it.

And how to slice it is what ostension or

simple

conditioning,

however

persistently

repeated, cannot teach. (1969, p.32).

Where does

this leave us?

It leaves us with the

conclusion

that

ostension,

the

paradigm of

‘realist’

individuation,

offers

no principle of

individuation after

141

all.

Ostension works only after the world has been sliced

up, and as Goodman and others have pointed out, how the

world is sliced up is language-dependent.

Returning to

the issue of reference, it

is precisely

this view of individuation which has been found to be

unwarranted,

viz.

the

‘realist’ view,

that underlies both

the cluster and the causal theories of reference.

When

Searle speaks of his principle of identification as "giving

me an ostensive presentation" of "some independently located

object" (1969, pp.168,

169), his commitment to a ‘realist’

View of ostension and individuation is clear.

Kripke

claims: "When you ask whether it is necessary or contingent

that Nixon won the election, you are asking the intuitive

question whether in some counterfactual situation, this man

   

would in fact have lost the election." (1980, p.41).

But

just who

this man is

is not a straightforward matter.

   

Indeed,

who this man is depends upon how we partition the

   

world, and Kripke, along with Searle, is mistaken in taking

that to be given.

10

(Devitt, likewise, in his discussion of

grounding as being dependent upon seeing an object as a

member of a very general category talks as though these

categories (or reference classes) are

out there.’

Of

course, much of what I am arguing in this chapter is

that

they are not.)

At this point I want to return to an question left

unanswered in

the last chapter.

In

the discussion of

private reference,

I claimed

that both Searle and Kripke

 

  

were committed to private intentionality and also to private

reference.

An objection which I mentioned, but did not

rebut, was

that a commitment to intentionality did not

entail a commitment to private reference because reference

was made to objects ‘out there’ in the world.

Even if

Searle and Kripke were committed to intentionality in a

private

sense

(of

being

brain

states

or

mental

representations), so the objection goes, there is still

publicity at the level of the objects to which speakers

intend to refer.

While I agree that objects are public, I

want to claim that they are not public in any ‘realist’,

theory-independent sense

(which is

the sense which Searle

and Kripke suppose).

As I have tried to argue in this

chapter,

the ‘realist’ sense of individuation to which

Searle and Kripke hold

(reflected,

for example,

in

their

statements about ostension) is mistaken.

Similarity

At the beginning of this chapter, I suggested that

individuation and similarity are interrelated.

I now want

to turn to the notion of similarity and claim that many of

the conclusions reached above concerning individuation are

applicable to similarity as well, viz. similarity is not a

straightforward matter,

with objects being similar

inherently;

rather

similarity

is

interest-

and

language-

dependent.ll

What does it mean to say that two objects are similar?

Intuitively, it means that they share some property or

143

 

 

 

 

properties in common (or perhaps that they have some of the

same properties).

To say that an object is more similar to

a second object than to a third means intuitively that the

first and second share more properties (or a weighted more)

in common than do

the first and third.

As

it stands, of

course, this will not do, because we need to know similar in

what respect(s).

Everything is similar to everything else

with respect,

say,

to existing or being capable of being

thought of.

On the other hand, nothing is similar to

anything else with respect, say, to being at a given spatio-

temporal location.

We say, then, that a hot, dry, still day

is similar to

a hot,

rainy, windy day with respect to

temperature, but not with respect to humidity.

But where

similarity

simplicater

is

incoherent,

does

similarity

relativized fare any better?

Not, I want to say,

if

it is

still assumed that similarity is independent of our language

and cognitive interests.

Consider the following ‘objects’:

e

c

E

The ‘object’ on the left is,

of course, similar to the

‘object’ in the middle in a presumably denumerable number of

respects

(e.g.

being printed on

this

page,

being English

letters, being a certain distance above _thi§_ivord).

However,

the

‘object’ on

the left

is also similar to

the

‘object’ on the right in a presumably denumerable number of

respects as well.

And the ‘object’ on the left is

dissimilar to

the

‘object’

in

the middle in

a presumably

denumerable number of

respects

(e.g.

being

less

than

one

144

 

 

 

 

mile to

the left of

it, being less than two miles to

the

left of it, etc.).

Likewise for the ‘object’ on the right.

These considerations seem to argue against similarity

simplicater, but what about similarity relativized?

That,

too, is

in trouble, for, as these ‘objects’ show, the very

notion of similarity per se rests on

the notion of

individuation.

One of the ways in which the left and middle

‘objects’ are similar to each other and dissimilar to

the

right ‘object’ is that they are both curved and it is not.

One of

the ways in which the left and right ‘objects’ are

similar to each other and dissimilar to the middle ‘object’

is that they are both instances of the English letter <e>.

However, for these claims to be made,

they rest on our

acknowledgment of

the existence of,

say,

English letters.

That is,

if our partitioning of the world does not include

English letters, then it would make no sense to say that two

objects are similar with respect to that or dissimilar with

respect to that.

What sense would it make to say that the

left and middle

‘objects’ are similar to each other with

respect to glub but dissimilar to

the right ‘object’ with

 

reSpect to glub?

If glubness is not a recognized reference

 

class, similarity (or dissimilarity) with respect to

it

is

meaningless.

And, of course, similarity - even relativized

to certain respects -

is incoherent if

it

is

taken to

be

similarity independent of

our theories,

interests and

languages.

12

Issues

of

sampling and replication I

take

to

be

145

 

 

 

 

subsumed within the issue of similarity and included in this

conclusion.

Goodman (1972,

1978) mentions sampling in the

course of discussing similarity.

Similarity, he says, does

not account for our predictive and inductive practices (one

of the points of the grue example).

He relays the following

incident:

Consider again an ordinary swatch of textile in a

tailor’s or upholsterer’s sample book. ...Mrs.

Mary Tricias studied such a sample book, made her

selection, and ordered from her favorite textile

shop enough material for her overstuffed chair and

sofa - insisting that it

be exactly like the

sample.

When the bundle came

she opened it

eagerly and was dismayed when several hundred 2" x

3"

pieces with zigzag edges exactly like

the

sample fluttered to

the floor.

When she called

the

shop,

protesting

loudly,

the

proprietor

replied, injured and weary, "But Mrs. Tricias, you

said the material must be exactly like the sample.

When it arrived from the factory yesterday, I kept

my assistants here half the night cutting it up to

match the sample."

The incident was nearly forgotten some months

later, when Mrs. Tricias, having sewed the pieces

together and covered her furniture, decided to

have a party.

She went to

the local bakery,

selected a chocolate cupcake from those on display

and ordered enough for fifty guests, to

be

delivered two weeks later.

Just as the guests

were beginning to arrive, a truck drove up with a

single huge cake.

The lady running the bake-shop

was utterly discouraged by

the complaint.

"But

Mrs. Tricias, you have no idea how much trouble we

went to.

My husband runs the textile shop and he

warned me that your order would have to be in one

piece." (1978, pp.63—64).

Replication, too,

as Hanna pointed out,

is related to

similarity.

To say that a given experiment is a replication

of another (or,

to say that one specific instance of

a

generic experiment is

a replication of that generic

experiment) is

to say that they are similar in relevant

respects.

But,

as

we saw, what counts as relevant is

146

determined by our cognitive interests and what generic

events there are is a reflection of our partitioning of the

world.

Where the issue of similarity most clearly touches base

with reference is with respect to the notion of natural

kinds.

Natural kinds are sets (or classes or categories -

groupings of some sort or other) of things that occur ‘in

nature’,

i.e.

in

the world,

independent of speakers of

a

language.

Examples abound: lions, tigers, oceans, cabbages,

etc.

The particular members of a natural kind are members

of

that kind because of some similarity (or similarities)

that they share with other members of that kind.

If one is

committed to essentialism, then to

be

a member of

a given

kind is

to have a particular property (or properties) that

is necessary and sufficient to attain membership

in

that

kind.

For instance, if one takes ‘having atomic number 79’

as

the defining property of something being gold, then all

and only those things having that property of having atomic

weight 79 are things that are gold.

Or if being feathered

is the necessary and sufficient condition for being a bird,

then all and only those things that are feathered are birds.

This property (or properties) which guarantees membership a

in kind is,

then, an essential property of

that kind.

(I

will say more on essences and essentialism in

the next

chapter.)

The

notion

of

a natural kind,

however,

does

not

necessarily depend upon the commitment to essential

147

properties or essences.

Another conception of a kind, or of

shared similarity, is what is called ‘family resemblances’.

This notion, originating with Wittgenstein (1953), suggests

that some kinds might not have essences.

The members of a

kind, say, might all be similar to one another, but not

necessarily because they all share any common properties.

For

instance

(and

this

is where

the

term

‘family

resemblance’ comes from),

the various members of

a family

might share certain features with one another, but not the

same feature.

A mother and son might have similarly shaped

noses, but the son and father might have similarly colored

eyes

(although very different shaped noses);

sister might

‘have mom’s hair’ and ‘have dad’s chin’, while only brother

and sister might have no (obvious) features in common.

They

all have enough features in common with at least some other

members

of

the

family

such

that

they

have a

‘family

resemblance’ with each other.

Wittgenstein’s famous example

to illustrate this notion is

that of games.

There is

no

essence,

he claims,

to what we consider games.

He says:

we

see

a complicated network of similarities overlapping

and criss-crossing: sometimes overall similarities,

sometimes similarities in detail." (1953, #66).

This notion

of family resemblance is not antithetical to that of natural

kinds.

One

can still be committed to

the notion of

(and

existence of) natural kinds, even if one denies that these

kinds have essences

(or,

that

the members of

a kind have

essential properties).

(The notion of family resemblance

148

 

can be taken as either an epistemic notion or a metaphysical

notion.

That is,

one might say that so far as we know a

given kind has no essence, but maybe we just haven’t found

it yet, or one might say that in fact a given kind has no

essence independently of our state of knowledge.)

Some problems of similarity have been noted above.

Others,

in connection with kinds,

have been explicated by

Quine (1969).

Treating similarity as a dyadic relation

(e.g. a is similar to b) won’t work, says Quine, because it

dismisses considerations of degrees.

For instance, if

colored

things

comprise a

kind,

then red

things will

constitute a set of things too narrow to count as a kind,

whereas if red things comprise a kind, then colored things

will constitute a set of things too broad to count as a

kind.

But there seems to be no reason to deny that kinds

can overlap (e.g. both mammals and tigers are kinds).

While

viewing similarity as

a triadic relation (e.g.

a

is more

similar to b than to c) escapes the problem just noted,

it

results in a different problem, viz. almost any two things

could count as members of some broad kind.

These problems of similarity are essentially problems

encountered by trying to define similarity in terms of

kinds.

What of the converse, trying to define kinds in

terms of similarity?

This, too, won’t work.

In arguing why

it won’t, Quine introduces two notions: paradigm and foil.

He states:

If without serious loss of accuracy we can assume

that there are one or more actual things (paradigm

149

cases) that nicely exemplify the desired norm, and

 

one or more actual things (foils) that deviate

just barely

too much to

be counted into

the

desired kind at all,

then our definition is easy:

the kind with paradigm a and foil b is the set of

 

all things to which a is more similar than a is to

b.

More generally, then, a set may be said to be

a kind if and only if there are a and b, known or

 

unknown,

such

that

the

set 1? the—kind with

paradigm a and foil 3. (1969, pp.119-120).

This definition, however, won’t work.

For example,

if we

take a particular shade of red as the paradigm for a kind,

we find that objects in that shade can come in all sorts of

shapes, weights,

sizes,

smells,

etc.

So mere degree of

overall similarity to any particular paradigm case (e.g.

a

certain apple) will afford little evidence of degree of

redness, since it will depend also on shape, weight, size,

smell, etc.

For Quine, the notions of kind and similarity

are substantially one notion.

There are other, related, problems with the notion of

natural kinds.

For example, most people countenencing

natural kinds would probably agree that birds are a natural

kind.

Not only are birds a kind, but they are natural; the

grouping of birds as birds reflects objective, external

similarities existing among birds and separate from our

languages.

(Science,

we believe,

allows us

to deal with

what we intuitively might

take

as borderline cases,

e.g.

penguins or ostriches or bats.)

Now, if we take birds as a

natural kind,

there seems no reason not to take,

say,

warblers as

a natural kind, also.

Indeed, one of the jobs

and virtues of ornithology is that it has shown in what

sense(s)

warblers

are

a

kind,

distinguishable

from other

150

kinds of birds.

Warblers are not hawks and neither warblers

nor hawks are ducks.

However, if we grant that warblers are a natural kind,

it seems that problems arise, problems of a slippery slope

sort.

If warblers are

a natural kind,

what of

Pine

Warblers?

0r Cape May Warblers?

0r Wilson’s Warblers?

Are

these,

too, natural kinds?

There is no obvious reason to

deny that they are, and in fact ornithologists classify them

all as separate species.

Ornithologists distinguish species

of warblers on the basis, at least at one level, of numerous

and various sorts of considerations: plumage color, shape,

size, geographic location, song, etc.

But if Pine Warblers

are a natural kind, different from the natural kind of Cape

May Warblers,

then,

as

I will now argue,

I see no escape

from the conclusion that this Pine Warbler is a natural kind

and that Pine Warbler is another - separate — natural kind.

For what is the basis for distinguishing Pine Warblers from

Cape May Warblers?

If

the distinguishing criteria are

phenomenal properties (or a property), then obviously every

Pine Warbler is distinct from every other Pine Warbler at a

specific enough level.

If

the distinguishing criteria are

genetic,

then

the same conclusion is reached.

If

the

distinguishing criteria are potential breeding and

reproduction (as Mayr (1942) suggested),

then, as Hull

(1965)

remarked,

we carry a commitment to

talk of

potentiality and besides it rules out things like mules as a

natural kind.

Actual breeding and reproduction might fare

151

 

better

than potential breeding and reproduction.

Whether

potential or actual, reproductive isolation might be a

(or

the) distinguishing criterion of a natural kind.

But if the

distinguishing

criterion

is

reproductive

isolation,

then

mules seem to stand as a counter-example (not only because

they are sterile,

and so not reproductively isolated, but

also because they are the offspring of horses and donkeys,

each of which is taken as a species (natural kind) separate

from the other (yet not reproductively isolated).)

If the

distinguishing criteria are ancestry or historical lineage,

then,

as Holsinger (1984) claims,

the determination of

natural kinds becomes somewhat arbitrary:

If

we

accept

this

definition

of

a

reproductive community and

this justification of

the "morphological yardstick", however, we are

left in a quandry.

In any "sexual" species there

will be

a large set of nested and overlapping

entities within a single interfertile unit

corresponding to different family geneologies.

Although we may be able to specify how far back we

are

to trace the ancestry for defining the

reproductive community by using the morphological

yardstick,

there

is

no obvious reason for

regarding that segment of geneology as

any more

important or fundamental than

the segment of

geneology shared

by

the members

of

a

single

population or group of populations or even the

segment of geneology shared by members of what

would usually be regarded as a genus.

Indeed,

depending

on

the

processes

that

we

were

considering, any one of these levels might be the

most important one. (1984, p.301).

There is no clear reason, then,

to rule out the slide

d°WQ

the slippery slope to

the conclusion that

individual entities

are kinds.

However,

this

conclusion would destroy the very notion of a kind, at

least to

the extent of making it superfluous.

There

152

 

seems no clear, non-arbitrary distinction between kinds

and individuals.13

Whether or not this slippery slope argument works,

it highlights

the problem of what kinds

are

and

Holsinger points to what I take to be an even more

fundamental trouble with talk of natural kinds

independent of our interests and language.

Echoing and

supplementing Holsinger’s claim, Elgin (1983) suggests

that like individuation and similarity (i.e.

what we

take

to

be individuals and what we

take

to

be

similarities), natural kinds (i.e. what we take to be

kinds) are dependent upon our cognitive interests and

our language:

The dependence of our systems of kinds on our

theories, and the dependence of these in turn on

our interests, values,

technology, and the like,

make questionable at best the thesis that our

predicates pick out real properties or natural

kinds whose existence, extension, and metaphysical

status are independent of any contribution of

ours.

And the claim that just these kinds or

properties are

required

to answer scientific

questions

or

provide

scientific explanations

supports that thesis only if backed by an account

of why these questions or forms of explanation

have priority — an account that does not in turn

appeal to

the practices or institutions of which

they are a part,

else all questions are begged.

(1983, p.35).

Natural Kinds and Reference

 

Now, what is

the significance of this view of natural

kinds

(and

of

similarity)

with

respect

to

issues

of

reference?

Searle has said nothing about natural kinds in

connection with reference.

However, Kripke has said a great

deal.

Within the context of discussing identity statements

153

 

 

 

 

(and his view that for rigid designators, if identity

statements are

true,

then

they are necessarily true),

he

summarizes his view on natural kinds (and natural kind

terms) in the following five claims:

First,

[Kripke’s]

argument

implicitly

concludes

that certain general terms,

those for natural

kinds,

have a greater kinship with proper names

than is generally realized.

Second,

[Kripke’s]

view asserts,

in

the case of

species terms as in that of proper names, that one

should bear in mind the contrast between the a

priori but perhaps contingent properties carried

with a

term, given by

the way its reference was

fixed, and the analytic (and hence necessary)

properties a term may carry, given by its meaning.

Third,

in

the case of natural kinds, certain

properties,

believed

to

be

at

least

roughly

characteristic of

the kind and believed to apply

to

the original sample, are used to place new

items, outside the original sample, in the kind.

Fourth,

scientific investigation generally

discovers characteristics of

gold which are far

better than the original set.

Fifth,

and independently of

the scientific

investigations

just mentioned,

the

‘original

sample’ gets augmented by the discovery of new

items. (1980, pp.134-139).

To get clearer on what Kripke is saying here and on what his

commitments are with respect to natural kinds (and, so, with

respect

to similarity)

each of

these five claims will be

inspected.

First, natural kind terms have a greater kinship with

proper names than is generally realized.

Kripke states that

this holds

"for certain for various species names" (1980,

p.134), including count terms (e.g. ‘tiger’) and mass terms

(e.g.

‘gold’).

It holds for terms designating natural

154

 

phenomena (e.g.

‘heat’) and corresponding adjectives (e.g.

‘hot’).

This claim is made clearer by and supported by the

arguments put forward for the other four claims, so at this

point I will only say that for Kripke this first claim is

true because he sees natural kind terms, just as proper

names, as rigid designators.

Identity statements involving

natural kind terms are, for Kripke, necessarily true if true

at all.

The reasons for this will be seen as the remaining

four claims are inspected.

Second, we must keep in mind the difference between the

properties by which reference is fixed for a natural kind

term and the properties which the

term carries

(which are

given by its meaning).

The reference fixing properties are

a priori

and

(perhaps)

contingent,

whereas

the

‘carried’

properties are analytic and (hence) necessary.

Just as

descriptions or ostension can establish a connection between

an object and a proper name

(e.g.

the Neptune case),

so,

too,

can descriptions or ostension establish a connection

between a kind and a

term (e.g.

‘gold’ designates the

substance instantiated by the items over there,

or at any

rate, by almost all of them).

To see why the reference

fixing properties

are

a

priori

(and

perhaps contingent),

while the ‘carried’ properties are necessary, consider the

following example.

Suppose gold has been defined as

a

yellow metal’ (‘gold’ means - perhaps among other things -

‘a yellow metal’).

Could we discover that in fact gold is

not yellow?

Perhaps because of some peculiar atmospheric

155

 

 

 

conditions or properties, we have always experienced an

optical illusion in seeing gold as yellow;

in fact, after

rectifying those conditions,

we come

to see

that gold is

really blue. Would we then say that there is no gold (since

gold is a yellow metal)?

Kripke says, no; rather, we would

say that gold is not really yellow.

That stuff, gold,

exists, but we were wrong to think it was yellow.

Our

designation of

iron pyrite,

which shares many of

the

phenomenal properties of gold, as fool’s gold (not as gold),

shows that we can change our descriptions which we associate

with a natural kind term,

yet say that we are nonetheless

talking about the same stuff

(e.g.

gold, even if

it turns

out to

be

blue).

So

the properties which are associated

with a natural kind when reference is fixed might well be

contingent properties (such as being yellow), but once they

have been used in the fixing of reference, they are a priori

properties

(because we

have stipulated,

in

the fixing of

reference, that this property is true of the object, or

kind).

On the other hand, since we could discover at a

later time that those reference fixing properties do not, in

fact, pick out the stuff we thought they did (e.g. gold

turns out to be blue, not yellow), they are not necessarily

true

of

the object,

or kind.

Rather,

the properties

‘carried’ by

the

term (i.e.

the properties that the stuff

‘really’ has) are the properties that are necessary.

Third,

certain roughly characteristic properties,

believed to apply to the original sample, are used to place

 

156

new items in the kind.

These properties, Kripke tells us,

need not hold a priori of the kind; we could at a later time

establish that some of the properties of the original sample

were peculiarities of that sample, and not generalizable to

the kind as a whole.

So, an object might not have all the

properties of the original sample, but nonetheless belong to

the kind

(e.g.

if

‘being capable of flight’ were

the

property which characterized

the natural kind

bird,

then

 

ostriches would not be birds,

but they are). On the other

hand,

an object might possess

the characteristics of

the

original sample by which the kind was established, yet not

belong to

the kind (e.g.

if ‘being capable of flight’ were

the property which characterized the natural kind bird, then

 

bats would be birds, but they aren’t).

It is an empirical

matter,

then, whether the properties originally associated

with the kind are in fact defining properties of the kind.

While

the joint sufficiency of

those properties might be

true,

they

are

not

necessarily

true.

Universal

applicability,

however,

is necessarily true,

if

true.

Kripke claims:

"‘Cats are animals’ has

turned out to be a

 

necessary truth.

Indeed,

of many such statements,

especially

those

subsuming one

species

under another,

we

know a priori that,

if

they are

true at all,

they are

necessarily true." (1980, p.138).

Fourth,

scientific

investigation

generally

discovers

characteristics of gold (or any kind) which are far better

than

the original set.

The standard of goodness here is

157

 

strict necessity, in the sense that identity statements

which express necessary truths are ‘better’ than those which

don’t.

Kripke gives the example of gold being identified

with the property ‘having atomic number 79’.

If it is true

that chemists have established that (pure) gold has atomic

number 79 (i.e. an object is (pure) gold iff it has atomic

number 79),

then this identity is stronger (‘better’) than

one which is based on the phenomenal characteristics of the

original sample (e.g. gold is

a yellow metal).

Kripke

states:

In general, science attempts, by investigating

basic structural traits,

to find the nature, and

thus

the essence (in the philosophical sense) of

the kind.

...The type of property used in science

seems to be associated with necessity, not with a

prioricity, or analyticity. (1980, p.138).

Fifth,

the ‘original sample’ gets augmented by the

discovery of new items.

Kripke doesn’t say further what he

means by ‘augmented’ here, but I take it that he means that

the extension of

the

term increases to include new or

different objects

than just

the original sample

(e.g.

we

find more chunks of gold or discover birds

that were not

discovered before - if for no other reason than that some of

these new birds were just born!).

Just as with proper names, species names

(or kind

names) can be passed from link to link in a chain and their

reference is determined by a causal (historical) chain, not

by

the use of any items.

(A slight difference between kind

names and proper names is that for some natural kind terms,

sensed phenomena, the way that the reference was fixed seems

153

‘overwhelmingly’ important to us, which, for Kripke, is not

the case with proper names.

The fact that we identify light

in a certain way, he says, "seems to us to be crucial, even

though it is not necessary“ (1980, p.139).

Now, what is

to be made of Kripke’s account of natural

kinds?

It

is,

I

think, basically mistaken, and mistaken

because it rests ultimately on ‘realist’ notions of

individuation and similarity.

As we saw above, both notions

are coherent only as interest- and language-dependent

notions.

With respect to kinds, what I have argued above is

that what ‘natural’ kinds

there are depends upon what

reference classes and similarities we recognize, and those

are interest- and language-dependent.

Perhaps the best way

to see Kripke’s commitment to a ‘realist’ view of kinds is

to look at his five claims about natural kinds.

His first claim, as we noted, was simply that natural

kind terms have a greater kinship with proper names than is

generally recognized.

This claim,

I said, would be borne

out with the other four claims, and so I turn to them.

His second claim was that we must keep in mind the

difference between the properties by which reference is

fixed for a natural kind term and the properties which the

term

‘carries’.

As

we

saw,

Kripke hypothesizes

that

the

reference of a natural kind term (just like a proper name)

can be/is fixed by ostension or association of certain

descriptions.

Both methods,

I have argued above,

are

mistaken if

taken in

a ‘realist’ way.

As

we

saw,

159

Wittgenstein and

Quine

showed ostension

to

be

not

a

straightforward matter at all, but interest- and language-

dependent.

The association of descriptions, we saw, depends

upon a conception of partitioning the world, and that was

interest- and language-dependent.14

Besides commitments to

talk of ostension and

reference fixing, Kripke’s talk of properties which a term

‘carries’ is,

if meant in a ‘realist’ way,

also mistaken.

That is, a term ‘carries’ properties only because terms are

part of public language, and while - because of being public

-

these

properties

are

indeed

‘carried’

by

the

term

independently of any given speaker, what properties a term

‘carries’ depends upon how we partition the world.

As

I

will argue in the next several chapters, Kripke’s view that

terms

‘carry’

properties

is committed

to

realism and

essentialism,

which in

turn are dependent upon particular

commitments

with

regard

to

individuation and

similarity.

Just as the latter are not independent of us,

so the former

also are not.

Third,

certain roughly characteristic properties,

believed to apply to the original sample, are used to place

new items

in

the kind.

There are several problems here.

First, Kripke uses the notion of an original sample as

though it were a straightforward matter.

However, as we saw

above, Goodman has shown that the notion of a sample depends

upon

the

notion

of

similarity,

and

that

is

not

straightforward.

(My criticism here is with respect to

160

 

Kripke’s apparent conception of a sample.

Below I will

criticize his apparent conception of an original sample.)

A

second problem with this claim of Kripke’s is his view of

the peculiarities of the original sample.

Those properties

roughly characteristic of a kind which are believed to apply

to

the original sample, Kripke tells us,

need not hold a

priori of the kind.

He says: "later empirical investigation

may establish that some of the properties did not belong to

the original sample, or that they were peculiarities of the

original sample, not to be generalized to the kind as a

whole." (1980, p.137).

The issue above was that what

properties belong to

the original sample (or

to anything,

for

that matter) depends upon our interests and language;

the issue here is

the other side of

the coin - what counts

as peculiarities of the original sample (or of anything, for

that matter) also depends upon our interests and language.

As noted earlier, objects are both similar and dissimilar to

other objects in

a denumerable number of ways.

So,

too,

there are a denumerable number of things which are peculiar

and not-peculiar to an object (or,

a denumerable number of

ways in which an object is both peculiar and not-peculiar).

Again, it depends upon our partitioning of the world.

Our

interests (or language or partitioning) might change in the

future,

so

that what is

taken as peculiar to

the original

sample (or to anything) changes, but that makes the notion

of peculiarity no less dependent.

A third problem with

Kripke’s claim is that I think he overstates his contention

161

 

 

 

that universal applicability, if true,

is necessarily true.

He remarks that ‘cats are animals’ is a necessary truth for

us.

While

I am somewhat sympathetic to his view here,

I

think that it

is far from obvious.

Consider the following

example.

I believe that all coal is hard, black and rough

until one day I read in the latest mineralogist journal that

Professor

X

has

uncovered

coal

with

the

exact

characteristics of chocolate pudding.

Coal is not always

hard, black and rough states Professor X, but some types are

in fact oozy, brown and velvety smooth.

What would we make

of such a claim?

I suspect, like Kripke, that we would be

more

inclined to believe that (fantastically) Professor X

had unearthed an underground vein of chocolate pudding than

that some coal is oozy, brown and smooth.

However, unlike

Kripke, I do not think that such an inclination is obvious.

The

thought of unearthing underground pudding is not much

less fantastic than the thought of some types of coal being

oozy.

(I am not suggesting that Kripke is wrong here, but

that

his confidence

that

he

is

right might

be

a bit

questionable.)

Fourth,

scientific

investigation

generally

discovers

characteristics of kinds which are far better than the

original set.

I will say more about this claim in the final

chapter, and in fact argue against it,

but for now I only

want to remark that Kripke’s assertion that the type of

property identity used

in science seems

to

be associated

with necessity not with a prioricity or analyticity is not

162

 

so

clear.

Kripke

takes

great care

in distinguishing

epistemic and metaphysical issues, between analyticity and

necessity.

However, I think that the distinction is not as

clear as Kripke indicates.

Using the coal example above,

for Kripke,

if

it turns out that ‘coal is hard, black and

rough’ is

a true identity statement, then it is necessarily

true.

We do not recognize this oozy stuff as coal because

it doesn’t have

the properties of coal.

It might not be

chocolate pudding, but, whatever it is,

it definitely isn’t

coal.

Now, I see little difference between saying something

like that (or,

‘cats are animals’ is necessarily true) and

saying that scientific identity statements are unfalsifiable

(an epistemic notion).

If ‘cats are animals’ expresses a

necessary truth,

then it seems to me that it also expresses

an unfalsifiable statement.

Where, then, is the difference

between the metaphysical and the epistemic with respect to

these sorts of identity statements?

It seems that all

necessary truths are unfalsifiable statements (something, I

take it, that we don’t want as part of science).

It may or

may not be

the case that all unfalsifiable statements are

necessary truths, but whether they are or not,

the

distinction between metaphysical and epistemic issues is not

so clear.

Fifth,

the ‘original sample’ gets augmented by the

discovery of new items.

In a footnote to this claim, Kripke

admits some hesitation to its formulation.

He states:

Obviously, there are artificialities in this

whole account.

For example, it may be hard to say

163

 

which items constitute the original sample.

Gold

may have been discovered independently by various

people at various times.

I do not feel that any

such

complications

will

radically

alter

the

picture. (1980, p.139, note 70).

Kripke’s

commitment to ostension and

a

‘realist’

conception of

similarity seems

to me obvious here.

Gold

could not be discovered, I want to say, by anyone anywhere

independently of our particular partitioning of

the world.

We partition the world into,

say,

metals and non-metals,

yellow and non-yellow, elements and non-elements, etc., and

in

that sense,

gold is not discovered at all,

but

stipulated.

An objection to this is that whatever we call

it (or don’t call it), this stuff was discovered (perhaps by

 

various people at various times).

But, of course, I want to

ask:

what stuff?

We are right back to ostension and

the

issue of similarity, and, as I have repeatedly said, Kripke

is mistaken to see these issues as interest- and language-

independent.

Summary

At

this

point

I would like

to quickly summarize the

main points I have tried to make so far in this

(and

the

preceding) chapters.

Both the cluster and

the causal

theories of reference, I have argued, share fundamental

commitments

to private

reference,

which in

turn

rests

on

particular commitments to and conceptions of individuation

and similarity.

Those conceptions,

I have urged,

are

mistaken, and so those commitments to those conceptions

weaken the theories.

To the extent that both theories share

164

certain commitments and conceptions, I see them as theories

which are more alike than different, and to the extent that

they hold the commitments and conceptions that I say they

do,

they are both mistaken and faulty.

I now want to show

that they are further committed to essentialism, realism and

a correspondence

theory of

truth,

and

to

the extent that

they hold certain conceptions about these commitments, they

are again mistaken and faulty.

In the final chapters I will

discuss alternative commitments and conceptions and an

alternative view of reference.

165

 

  

 

 

 

 

 

NICHIGQN STQTE UNIV. LIBRRRIES

Ill4|MINIill!MW“IllllllllllilllllilllilIllIllHI

31293007971009

 

 

A.

”3.21:...

.3...” 4.14.1”

{11.11.1111

. 1..I..11(1

..1

1.

111 11......1 .1

I...

...

t...

..-;

C3511...

....71.1%.17J1.Whl

11.1.1 AW...

fr

.

1.

C

. .

111 1.14.5.11111111111111.......1.
1....11rr1t1t111.94.711.11? (1411.11.11

.
r.1.....1..1.1111/1111_..:.u.11..
1.11.1 «11.73:..11111... 1.
.1r..0.rrrlr1...I7...rrzr11r.1.1.r.111.$.....1l2(.11
11.11.111.21
......1:1....11..11...111.1.1..1...n..1.m.1111J.1111.. 1
£111.13 11!?
:1111H11n..11111...1.u11.1..m111.

1 111.14..
f1];
.

(’31.

..

I

If

”Parch‘ulfvfifI

.11....1. .

.

1.11.11.41.31 11r.1.1.11.1r1

. 1.1. .1 .11....1...1..11m.

r

...

1311.11.31

....Eiu

.

1..

.

.

11.1....
......:Ir..trl.

1....
1

1

1

.

.

z
1......

$1.11.

1.1..

1441

J

..

.

.3

11.,

I

.

..qL.

.

.

1

.

.114.........

.

.

err...

.

J

4.1.5.1.“;5mm...
1 ..

..

.1 ._
1 ...»... . 211. .

.
1..

.1

.

1
. .11..

.

1.1:...
.1
..1
. 1111111.6.114.

,

11111.“...

5.2.... .

1.1.

. he “writ v2.11 tvhrtddin.
”17.11.31

.

1

1

..

1 1.13....

11.“. u............... .1.

.. .1. 0.11.} 11......1...11 111$ ......1..1.n.1....11..1.¢.fl11...1

.11.1.....1...._.1...m. ......w1..1....»1§.11.r11n111...111.11
1111... 11.11.1111...

4.1.0.1.... ....1111111..11¢1.1....M11.11.r21.11$
1114111111116. fifilwf11....

(1.1.111... I..... r ((1.1! Lng 1.1!.
. 11.1.1... .

11.4%....L11...11.u.u.111111.115511....

1.144111.1..1111411,..1..11.u11.1111,m1111..

1111.1
I?..I..Ir(:l(11
1F. J?r4}\.rrrf1...11.u.t1413at...u.111...r1...I.1

$111.84.... «91.413121.21.15.11

.11. . 115.11..
1 I

W111m111nmw111. 111...“...1 1311......11.11.....1mw11..15-5.11n.nm11._41......1m...111r..1.....1

1H14;

1.......1.....;. 1......1....nu.....1......1....1.n111111.511:...11..1........1..111.11¥
1111.1..11...1.......1111......1..1.1....1.

rrr1.14..rJ..n1TI!J1123. ((3...

11.....1J....11....
1.x..rlr111r I?

1.1...1r.....111..1.w.11.m

I IV 1 [It

1,
. I..!

1. 1 .

34.1.1111».323‘..I

13:114.“.

:1...11

Jun"?

:1... #11

3.11.1..5

...mM.

Bro“...

111....

uné

.3.

1.11:,

I1 (01..

1.. in

...1

11..

r ...
.

..111

....

.1

..

'1

1

I

1

.l

.

.

Q

.

4

.

.

..rdif...

.

.11 1.

1
(.11..1..1.1:

1
Uert11mehiae

.

.

1.

11»... 7.11.1
4.13”».

.3111d.r.u.1. .1. £1511!

. 1...

.

.1231:in

.11... 1 1.1111... 1 $11113.
Wfimfiflrwmflo:4?!15.11.:LW: 1 $111171”!

(WING... rlflufnmudf”1217f?

31.01.».

I111}?

..Ivuflttrfg

4

.

1

.

C.

.... 1.11.11.15.11...1...

1.11W1111.J1J..1.1.«n 111.43.11.31
I....11rlhwhultr.
1.1111111:

Lani.

3

.

1

.

l

I

1

X.

m»..mwfif
Fr: ”WWWdfvmtrié

.. 1v .

.

112111

(94.1.1315 ”fa—4.1?
141.1...

.4.

14.!

1%.?

I

13:. 1

..I 1.1.11fig... ...
hr ”#31: “13.1...w1...
. :11!
...W.Wtfl.. WM:..

1.1 .

.r

tymufirr. ...NfimeWL? “1.31.”...1....1i.1......ryryh11JWWH¢tIO1I1

.37.
:r. :

P.5W4k

.n..511

t

.r
.11...

(411‘ny

.1“.
..
. a
.1.

4.1fwflmr7.3.111

.13....1.1.1.,

I

amwhflnwx

31.1.1.3

...d

Hr

.

..

{M 1?..1”.l

351..
.

:11”.

.

.. 1

1 c

.H.1

.

..
.71 I
.

7.1.2.1.». I 1471131K<1rWLmH3r

..I(13.4.1.1...

.4ream-......1.. fl.

. 13$”.ng

«Irlflfiflm

ram»:
{1.111%}...11M.

...

,

.

....

1.

1.
1.1..

I..1

{.117

.11....

.1

.

1..

1...? 3311.11

.111...

54.4. 73.7!if1t3w1f1WMvéwr

gunfirfirflmrmflin.

.3”?

13:1.

mg...“

(man-31...?
£71
fill

1

5.11.51”!Lawn

fliiélznw A.

1 .1g

If.

7,4412]?

fi

g

1.

...... 5,1,1...

....
11.11... 1......1...M1M¥k§§..
1 ..
.21. -I 1! ..rr: II.
1.. ... .Mwuvfhfinfi

31.112.11.11.11

(11.

pxszezerév»

1.11... 1.1.11...

fflgfllflz

934541111115“

(flmfirt

.

,

.1.
v. 1(1(1.fi.r12.13111.(3113.

.firm”.

3.01.1.2: .

.

1m...

1.3.14“If I...

14..
15.1.2.

...4

.

.141».

1r

.

1

141.11....

”14.”..111

,

.If.3&«m11®11.%1w18

”WIN” a...
.

.

1

”my?.

111.1141.

.

fhowflwma.

.31......mfi. .

1......
L?$34er

.1..m.....w..11.m1(1.

53.1»... .1911... .

r I

1
.11
.11....(1:
11.9.1?{AWN
1131......)

...

..

“1.3.,

1.11.... 1.1.
.11.. ...:

4

...!

H

1.
1 .

.

741h1....d1H1..11?...1mrӣ

111.11%

....
ob

1

”'17

1:1z

11.143.13.131

.

....uu

...w.

7.11. 

flung/J.

E...I
S...3

M
3
1

:
:
1

;
3
.
3
1

”Ml/.71

9
3
1

.
‘
1

‘
3
5

.
,

-
L

7

‘
4
‘

‘

.

1

un{I@nmmjg‘Wnirmrnmn

W»

017

i

 

 

LIBRARY

Michigan State

 
  

University

 

RETURN BOX to remove this c

before due due.

heckout from your record.

PLACE IN

TO AVOID FINES return on or
  

DATE DUE

DATE DU

E

DATE DUE
   

    

 

An

c:\r:Irc\ddod\Brpm3—p.'

 
 

  

 

NOTES

lOr, another example: "When you ask whether it

is

necessary or contingent that Nixon won the election, you are

 

asking the intuitive question whether in some counterfactual

situation, this 332 would in fact have lost the election.

(1980, p.41).

2Elgin (1980) makes this same point with regard to the

notion of scientific progress.

3The

reason we

"are well aware which of

the

two

incompatible predictions

is genuinely confirmed," for

Goodman,

finally

boils

down

to

pragmatic

considerations.

English (at least in

this

case) works better than the

gruesome language.

4Goodman repeats this theme throughout his writings:

There are many different equally true descriptions

of the world, and their truth is the only standard

of their faithfulness.

And when we say of

them

that they all involve conventionalizations, we are

saying that no one of these different descriptions

is exclusively true, since the others are also

true.

None of them tells us the way the world is,

but each of them tells us a way the world is.

(1972, p.31).

Frames of reference,

though,

seem to belong

less

to what is described than to systems of

description ...

.If I ask about the world, you can

offer to tell me how it

is under one or more

frames of reference; but if I insist that you tell

me how it

is apart from all frames, what can you

say?

We

are confined to ways of describing

whatever is described.

Our universe, so to speak,

consists of

these ways rather than of a world or

of worlds. (1978, pp.2-3).

5Even

Nagel

(1961)

accepts

the

presence

of

‘characterizing value judgments’ (as opposed to ‘appraising

value judgments’) as part of science.

These ‘characterizing

value

judgments’

are

a

version of

what

I

am calling

‘cognitive values’.

6Another

argument

for

the

interest-

(and

language-)

dependence of theories is that theories are, in large part,

metaphorical (cf. Kuhn (1962, 1977, 1982), Hesse (1974,

1980,

1982) and Rothbart (1984)).

Quite simply, metaphors

reflect and display our interests and languages; these

metaphors are an integral part of our theories.

7Such a treatment of individuals

might seem strange

because our common sense notions of individuals as clearly

demarcated objects, each with a rather distinct spatio-

166

  

 

temporal continuity.

However, Quine (1976) and Goodman

(1972) have both argued that there is no good reason to

forbid the conception of individuals which consist of widely

scattered and very unlike parts.

(Quine, for example, see

nothing wrong with speaking of mass terms, such as sugar, as

individuals.)

Leonard and Goodman (1940) pursued this

conception of individuals to the point of attempting to

construct a calculus to accommodate it.

8There has been some debate whether events can be

treated as individuals.

Davidson (1970), I think, has

argued conclusively that they can.

9Cf.

Thomson

(1965)

for

further argumentation for

a

‘non-realist’ treatment of experimental replicability.

loin fact, Kripke’s presentation of

his Gddel-Schmidt

counter-example to what he calls thesis (3)

of the cluster

theory of reference begs the issue.

Kripke says:

"For

example,

it

is

argued,

if

we

say,

‘Gddel proved

the

incompleteness of arithmatic,’ we are, of course, referring

to deel, not to Schmidt." (1980, p.85, note 36).

This begs

the issue because by saying, "we are, of course, referring

to deel" Kripke seems

to think that this instance of

 

‘Gddel’ clearly refers

to an independently located object.

But it doesn’t.

Who is he referring to here?

It won’t help

to

say,

"Well,

to Gddel,

of course" because that is just

what is at issue.

Kripke seems to take it that we could, if

need be, rely on ostension to pick out deel.

llIndeed,

much

of

the

discussion above

about

partitioning could be interpreted as being more explicitly

related to similarity than to individuation.

For example,

the green/grue question is

a question of whether emeralds

are similar with respect to color.

Of course, one of

the

points of this chapter is that there is no clear distinction

between issues of individuation and issues of similarity.

12For

an

interesting and

detailed

discussion

of

similarity and its significance for philosophy of science,

see Barnes (1982).

13The issue of natural kinds has been one that has

generated a lot of attention recently within the philosophy

of

biology.

A large body of literature has flowered

concerning the notion of species versus individuals (or

species as

sets versus species as individuals).

See,

for

example, Hull (1978), Kitts and Kitts (1979), Dupre (1981),

Holsinger

(1984)

and Kitcher

(1984).

Similar conceptual

difficulties as those noted in the text above exist for

asexual and non-biological entities.

14It is interesting to note

that with regard to

this

point,

a footnote of Kripke’s sounds quite reminiscent of

167

 
 

 

 

his thesis (5)

of

the cluster theory (viz.

The statement,

‘If X exists, then X has most of the 0’s’ is known a priori

by

the speaker).

Kripke says in the text (with regard to

the baptismal reference fixing): "Several features of

this

baptism are worthy of note.

First, the identity in the

‘definition’ does not express a (completely) necessary

truth:

though each of these items is, indeed, essentially

(necessarily) gold,

gold might have existed even if

the

items did not." (1980, p.135).

His note 69 reads:

Assuming, of course, that they are all gold;

as I say below, some may be fool’s gold.

We know

in advance, a priori, that it is not the case that

the items are typically fool’s gold; and all those

items which are actually gold are,

of course,

essentially gold. (1980, p.135, note 69).

Now,

it

is not obvious to me

that there is any difference

between ‘If Aristotle exists, then Aristotle has most of the

¢’s’ and

‘If gold exists,

then gold has most of

the 0’s’

While Kripke explicitly states that we must distinguish

between the reference fixing properties of an object and the

(I take it) essential properties of an object, this reminder

does not prove to be so helpful.

For example, suppose one

of the reference fixing properties is an essential property,

and we know it (supposing that there are essential

properties and that we can in fact recognize them).

It

would

seem,

then,

that Kripke

is

committed

to

the

a

prioricity thesis.

A possible response to

this claim is

that even if one of

the reference fixing properties is an

essential property, that establishes a necessary truth, not

a priori knowledge.

One might say,

it might be

the case

that an essential reference fixing property is (necessarily)

true of an object, but it doesn’t follow that we know this a

priori.

However,

it

seems

to

me

that

if

we

fix

the

reference of

a

term a priori and if, when doing that,

we

know

that

one

of

the

reference

fixing properties

is

essential, then we know a priori that if that thing (stuff)

exists, then it has -

if not most of - one of the 0’s.

For

example, suppose in 1947 (before Israel became a sovereign

state,

but knowing

that

it soon would be),

someone said,

"Anyone born in Israel is an Israeli citizen."

It seems to

me that the speaker knows a priori that if someone is born

in Israel after time t (where time t is when Israel attains

statehood), then that person is an Israeli citizen (i.e.

if

X exists, then X has one of

the 0’s).

It might be argued

that being a citizen, or being an Israeli citizen, is not a

natural kind,

but,

of course,

that is part of what is

at

issue (viz. what a (natural) kind is).

Besides, the point

here is that Kripke sounds quite a bit like the cluster

theory of reference which he criticizes.

168

  

 

CHAPTER SIX

ESSENTIALISM

In

the previous section we

saw

that

the received

theories of reference are committed to certain conceptions

of individuation and similarity, conceptions which I argued

were mistaken.

An area of particular intercourse between

issues of reference and issues of individuation and

similarity is

that of

the notion of kinds, especially

natural kinds.

As we saw,

in the discussion of natural

kinds, the notion of essences (or essential properties) was

integral.

Among

Kripke’s claims

about

natural kinds

(natural kind terms) was his assertion that in general,

science attempts, by investigating basic structural traits,

to find the nature, and thus the essence of kinds.

In this

chapter

I

want

to

look

more

closely

at essences and

essentialism.

I

will

argue

that

conceptions

of

individuation and similarity carry with them conceptions of

essences and commitments to given views of individuation and

similarity carry with

them commitments

to given views

of

essences.

Furthermore, I will argue

that both Searle and

Kripke, by being committed to mistaken ‘realist’ conceptions

of individuation and similarity, are committed to a mistaken

‘realist’ conception of essences (i.e. to essentialism as it

is usually interpreted).

I69

  

 

One

of

the difficulties

in talking about essences or

essentialism is that philosophers are not always in

agreement about how they state the issue.

Depending on whom

reads, one gets a different statement of what essences

(essential properties) are or what essentialism is.

Brody

'"*

(1980), for example, speaks of essential properties as those

properties necessary for an object’s existence:

7

's

5

b

«
?

6

What is the basic idea behind the claim that

there is a distinction between the properties that

an object has essentially and the properties that

an object has accidentally?

It seems

to

be

the

following: on the one hand,

there are some

properties that an object must have; if the object

didn’t have them, it wouldn’t exist at all.

These

are the properties that an object has essentially.

On the other hand,

there are some properties that

an object has but that it might not have.

The

possession of

these properties

is

not necessary

for the object’s existence.

These are

the

properties that an object has accidentally. (1980,

p.84).

Plantinga (1974)

offers several definitions of

‘essence’ including the following:

E is an essence of [say] Socrates if and only if E

is essential to Socrates and everything distinct

from him has E (the complement of E) essentially.

(1974, p.70).

E is an essence if and only if there Is

a world W

in which there exists an object x that (1) has E

essentially, and (2)

is such that there is

no

world W* in which there exists an object distinct

from E that has E. (1974, p.72).

S

is an essence if and only if S

is a complete and

consistent set of world-indexed properties. (1974,

p.77).

For Flew

(1979)

the "notion of

a fixed and

timeless

possibility of existence"

(1979,

p.111)

is how he defines

‘essence’.

Subsequently, he offers several conceptions of

170

 

 

 

essentialism.

One conception of essentialism is: a position

holding that something can have an essential property in

virtue of a definition, or as described in a certain way.

Another conception of essentialism is: a view that maintains

that some objects - no matter how described - have essences;

that is,

they have essentially or necessarily, certain

properties, without which

they could not exist or be

the

things they are.

Hirsch (1971) characterizes essence in terms of spatio—

temporal continuity of an object.

He argues that among the

properties which are true of an object, some play a special

role in determining an object’s identity.

The terms which

name those special properties (he calls them E terms) are,

in a sense, necessary to the object’s identity.

All other

terms which name properties true of an object, but which are

not E terms, he calls A terms.

Be capsulizes his position

(with the insistence that this is only preliminary) as

follows:

A term is an E term if it cannot possibly be true

of an object temporarily; if

it

is true of an

object at all,

it

is necessarily true of

the

object throughout the object’s entire career.

(1971, p.34).

The point, of course,

of all these citations is

that

one

of

the difficulties in talking about essences and

essentialism is

that philosophers speak in a multitude of

ways when discussing these issues.

When I speak of essences

or essential properties I will mean the following:

to say

that an object1 has an essence (i.e. an essential property

171

 
 

 

or properties) is

to say that the object has some property

or properties that are necessary for that object to be that

object

(i.e.

for

that

object

to

be

what

it

is).

Essentialism I

take

to

be

the view

that objects have

essences or essential properties.

(I

take

the following

definition

of

essence

to

be

simply

a

different

characterization of

the definition I

have

just

given:

an

object has

an essence

(i.e.

an essential property or

properties) if

that property or properties is

true of

the

object in all possible worlds; or,

the object has a certain

property or properties

in all possible worlds.2

Another

characterization of this position I am trying to elucidate

is given by Slote (1975): a property 2 is essential to an

entity 3 if and only if 3 could not have failed to have R;

essentialism is

the view that various entities have certain

of their properties essentially (1975, p.1).)

As noted above, one objection to essentialism is

that

not all objects classified as the same kind necessarily have

a shared essence. (Keep in mind here that, for the moment, I

am accepting the distinction between individuals and kinds,

and,

for the moment, I am speaking about essences with

respect

to kinds.)

The notion of family resemblances is

meant to illustrate this point.

Some kinds might have an

essence; an object might need to have a particular property

in order to be a member of a certain set (e.g. for an object

to be

a bachelor it must be unmarried).

However, this

requirement of having an essential property might not hold

172

  

 

for all kinds

(e.g.

there might be no essence to games).

The notion of family resemblance, of course, is not so much

a direct criticism of the notion of essences as

it is a

criticism

of

the universality

of

the

doctrine

of

essentialism.

That is,

even if

the notion of family

resemblance is true, it does not show (nor is it intended to

show)

that

there are no essences or that essentialism is

wrong, rather it shows that not all objects have essences

(or that having an essence is not an essential property of

objects!).

While this is not so much a direct criticism of

essentialism,

it

is

a challenging ‘indirect’ criticism of

essentialism.

The challenging aspect of

it is that, at

least with respect to kinds,

the doctrine of essentialism

should not be assumed.

It might not at all be the case that

as

science progresses it discovers

the underlying traits,

and hence the essences, of (some) things.

There might

simply be no non-arbitrary underlying trait to games.

There is

a deeper challenge posed to

the doctrine of

essentialism by

the notion of family resemblance.

It

is

this:

the notion of family resemblance calls into question

the presumed sharp distinction between questions and issues

of ontology and epistemology.

Let me explain.

The doctrine

of essentialism presumes that if objects are considered to

be members of

the same kind (are called the same kind of

thing; are in the extension of the same predicate) then they

must share some property (or properties) in common.

As

science progresses the shared properties by which we group

173

 
 

objects into kinds (by which we call objects by a certain

term) will be

the underlying properties which we discover

these objects (and kinds) to have.

Ideally,

science will

result in

the discovery of essential properties.

But note

that this view presumes that the ontic question of whether

kinds have essences (and if

so what they are)

is separate

from the epistemic question of how we (come to) know this.

Likewise,

the semantic question of what we call essential

properties

is presumed to

be separate from

the epistemic

question of how we (come to) know this.

The notion of

family

resemblance,

however,

calls

this separation of

questions into question.

That is, if there is no compelling

reason to believe that all games share a particular property

(or properties) simply because they are called games,

then

there is no compelling reason to believe that kinds ‘really

do’ have essences or that there ‘really are’ things which we

call essential properties.

After all, what sense does

it

make to claim that games ‘really do’ have an essence even if

we can’t find it?3

While the notion of family resemblance is,

I think, a

strong challenge to

the doctrine of essentialism with

respect to kinds, it is not so clearly a strong challenge to

the doctrine of essentialism with respect to the identity of

objects.

The doctrine of essentialism pertains both to

kinds and to individuals.

Not only is

it claimed that for

an object to belong to a given kind, the object must have an

essential property (or properties) 33 that kind, but also it

174

 
 

 

is claimed that for an object to be what it is, rather than

 

being some

other object,

it must have some essential

property (or properties).

While the notion of family

resemblance challenges the doctrine of essentialism in the

former case,

it seems irrelevant to

the latter case.

It

seems to make little sense to speak of an object having a

family resemblance to itself.

(The only sense it might make

is

trivially -

it resembles itself

since it

is identical

with itself.

Indeed,

if

the doctrine of essentialism is

true at all, then, as Cartwright (1968) has said, it is true

in this sense,

since the attribute of being self—identical

is essential to everything.)

Now,

it seems

to me

that

the distinction between the

essence of kinds and the essence of individuals is not so

obvious

(in

one

sense),

nor

does

the doctrine

of

essentialism hold in either case.

At this point I want to

argue for this claim and then indicate how I see both Searle

and

Kripke

as

being committed

to

the

doctrine

of

essentialism.

Individual Essence versus Kind Essence

 

Intuitively

the

distinction

between

the

essences

of

kinds and the essences of individuals seems plausible.

This

is reinforced by the consideration that the notion of family

resemblance makes sense with respect to kinds but not with

respect to individuals.

However, other than with regard to

the applicability of

the notion of family resemblance, what

is

the basis of

the distinction?

It seems to be

this:

it

175

 
 

 

makes sense to speak of an object being what it

is

independently of an object being an instance of a kind.

True, Nixon is

a member of

a denumerable number of kinds

(e.g.

former

presidents,

Americans,

males),

but

independently of all of those, Nixon is Nixon.

In some

possible world, he

is not a member of

the kind ‘former

presidents’, but in all possible worlds Nixon is Nixon.

The

essential property (or properties) that Nixon has such that

he is Nixon in virtue of having that property (or those

properties)

is

different

than

the

essential property

(or

properties)

that

Nixon

has

such

that

he

is

a

former

president in virtue of having that property (or

those

properties).

For example, Nixon’s essence as Nixon might be

his particular genetic make—up, whereas Nixon’s essence as a

former president is his having been president at some time

in the past.

So

the distinction between the essence of

kinds and the essence of individuals is legitimate.

Now, I want to suggest that the intuitive legitimacy of

this distinction is mistaken, or at least not so clear.

We

cannot speak of, say, Nixon as Nixon outside of speaking of

Nixon as a member of some kind.

What sense does it make to

say that Nixon is Nixon regardless of whether he is a former

president or an American or

...?

Speaking of Nixon as

the

object with such-and—such a genetic make-up won’t help

because

that,

too,

places

Nixon within a particular kind

(e.g. biological organisms).4

My point here is that we can

never say who (or what) Nixon is without individuating

176

 

 

 

 

Nixon.

But the process of individuating Nixon is also to

place Nixon within kinds.

To say

that Nixon is

this

individual is

to say that Nixon is not that individual (or

anything else).

This

requires

some

principle

of

individuation.

After all,

we must use

some criterion to

individuate Nixon.

But as I argued in the previous chapter,

individuation is interest- and language-dependent.

This, I

take it,

is much of

the point and the force of Quine’s

opposition to essentialism:

Mathematicians may conceivably be

said

to

be

necessarily rational and not necessarily two—

legged;

and

cyclists necessarily

two-legged and

not necessarily rational.

But what of

an

individual who counts among his eccentricities

both mathematics and cycling?

Is

this concrete

individual necessarily rational and contingently

two-legged or vice versa?

Just insofar as we are

talking referentiaily of

the object,

with no

special bias

toward a background grouping of

mathematicians as against cyclists or vice versa,

there is no semblance of sense in rating some of

his attributes as necessary and others as

contingent.

Some of

his attributes count as

important and others as unimportant, yes; some as

enduring and some as fleeting; but none as

necessary or contingent. (1960, p.199).

The gist of Quine’s example, I believe, is that it makes no

sense to speak of essential (or accidental) properties

without the presumption of some ‘background grouping’.

While Quine does not address the issue directly, I take his

point as pertaining to both individuals and kinds.

It makes

no sense to say that a particular property is essential to

an object unless it is clear what kind of object the object

is.

In addition, it makes no sense to say that a particular

Property is essential to an object unless it

is clear what

177

 

 
 

 

object is being referred to and that the object is this

object,

not

that object.

This,

however,

presupposes

the

process of

individuation, which in

turn presupposes a

principle of individuation.

We cannot separate the issue of

Nixon as an individual from the issue of Nixon as the member

of kinds.

Such a separation,

though,

is vital to

the

distinction between the essence of kinds and the essence of

individuals.

At this point one might object that my claim is false

that we cannot separate the issue of Nixon as an individual

from the issue of Nixon as a member of kinds.

The objection

is that something must be some thing in order for it to be a

member of some kind.

To be a member of some kind is to be a

unember,

i.e.

to

be

an

(identifiable)

individual.

Individuation, then, precedes membership (and partitioning).

However, I

think that

this objection fails.

It fails

because, as

I argued in the previous chapter, individuation

(and

partitioning)

is

theory- and

interest-dependent.

Individuals

are

not

out

there’

waiting

to

be

given

\

membership

into

various

kinds.

Individuation

requires

a

principle of individuation, which rests on some conception

of similarity.

As we saw,

the paradigm case for supposed

objective individuation is ostension.

But,

as we saw,

ostension makes sense only when the objects referred to are

partitioned into kinds (e.g. ‘sosh’ makes sense, and objects

can be identifiable as being sosh, only when we know that,

say,

sosh is

a color).

While

it

is

true

that similarity

 
 

 

makes

sense

only when there are individuals which we can

call similar or dissimilar (and, so, individuation precedes

membership and partitioning), it

is also true

that

individuation makes sense only when there is a principle of

individuation (i.e. some similarity relation so that we can

determine that an object is what it

is

and not something

else).

Partitioning,

then,

precedes individuation.

I

do

not see this result as contradictory.

Rather, it seems to

me that it points to the interdependence between individuals

and kinds,

between a principle of individuation and a

similarity relation.

To ask which comes first conceptually

is,

to me,

to ask which comes first, the chicken or the egg.

With respect to the particular point at issue, I think that

the objection fails which says that because individuals

precede membership the distinction between essences of kinds

and essences of individuals is not blurred.

Earlier

I

claimed

that

the

notion

of

family

resemblances was a strong criticism of essences of kinds but

that

it

seemed

irrelevant to essences of

individuals.

I

then argued that the distinction between essences of kinds

and essences of individuals is blurred.

These

two claims

might seem to entail that the notion of family resemblances

is relevant, after all,

to the issue of

the essences of

individuals

(since we can’t sharply distinguish kinds

and

individualS).

However,

it

is

still not clear how

family

resemblances could be relevant to individuals.

The reason

that this is not clear, it seems to me,

is that it rests on

 

 

 

 

an underlying conception of

individuals as simple,

indivisible entities.

That

is,

family

resemblance makes

sense for comparing different entities, but it doesn’t make

sense when ‘comparing’ a single entity to itself; the notion

of comparison requires two or more entities being compared.

This underlying conception, though, is questionable for two

reasons.

First, if Ghiselin (1974) and Hull (1978) are

right, some individuals are not simple, indivisible entities

(e.g.

they

argue

that

species

can

be

considered

individuals).

I will return to their proposal at the end of

this chapter.

Second, as argued above, we cannot treat

individuals as simple,

indivisible entities outside of

the

context

of

treating

individuals as

members

of

kinds.

Individuals are not just ‘out there’ as simple, indivisible

entities.

I

take

talk of possible world individuals and

cross-world identity to be an illustration of this.

We

speak of

the

same

individual in different possible worlds

(or

in different counterfactual

situations)

and

speak of

identifying

this

individual

in

these various possible

worlds.

Criteria for cross-world identification might be

the presence of some essential property (a la Kripke), but

there

is nothing incoherent with suggesting family

resemblance as

a criterion for cross-world identification.

While Kripke is

a philosopher in this world, he might be a

plumber or an exotic dancer in other possible worlds,

and

there is nothing incoherent in claiming that we use family

resemblance criteria for identifying him in

these various

 
 

possible worlds.

While this does not show the necessity of

family resemblance criteria,

it

is

not intended

to.

The

notion of family resemblance was never intended to deny the

existence

or plausibility of essences;

rather

it

was

intended to deny the necessity of

the existence or

plausibility of essences.

Above I remarked that the separation of

the issue of,

say, Nixon as an individual from the issue of, say, Nixon as

the member of

kinds was vital

to

the distinction between

essences of individuals and essences of kinds.

In addition,

it seems to me that it is vital to the essentialism to which

both Searle and Kripke are committed.

I want, then, to turn

to

them.

I want to show

in what sense they are both

committed to essentialism and show how their commitment is

based on this mistaken distinction and ultimately on their

mistaken conceptions of individuation and similarity.

Reference and Essence

Searle does

not directly address

the

issue of

essentialism in his discussion of reference,

however,

his

theory of reference does, I want to say, commit him to

essentialism.

As noted in chapter one, Searle states that

it is

"a necessary truth that Aristotle has the logical sum

[inclusive disjunction]

of

the properties commonly

attributed to

him"

(1969,

p.173).

It

is

a necessary

condition for an object to be Aristotle that it satisfy at

least some of the descriptions which speakers associate with

the name ‘Aristotle’.

This sort of claim strikes me as

a

181

  

 

 

 

clear commitment to essentialism.

There is some description

(property) affiliated with Aristotle such that for an object

to be Aristotle, it must satisfy this description (have this

property).

Granted,

we might not know exactly what that

description (property)

is, nonetheless, within the

disjunctive set of descriptions (properties) associated with

‘Aristotle’, it

is necessary that for an object to

be

Aristotle it must satisfy that disjunctive set.5

Searle makes other essentialist claims.

For example,

with respect to

the

issue of whether or not proper names

have senses, Searle interprets the question ‘Do proper names

have senses?’ in

two forms,

a weaker and a stronger form.

The weaker interpretation of

the question ‘Do proper names

have senses?’ is: ‘Are any propositions where the subject is

a proper name

and

the predicate a descriptive expression

analytic?’

To answer this question, for Searle, we need to

be clear on the identity of the object referred to; we need

to know

that we

are

talking about the

same object on

different occasions.

We need to know that a particular name

 

names

the same mountain,

person,

river,

etc.

on different

occasions.

So, he says, this gives an affirmative answer to

the weaker version of

the question of whether or not any

propositions where the subject is

a proper name and the

predicate is

a descriptive expression are analytic.

Says

Searle:

Some general term is analytically tied to any

proper

name:

Everest

is

a

mountain,

the

Mississippi is a river, de Gaulle is a person.

Anything which was not a mountain could not be

182

 
 

 

Everest, etc.,

for to secure continuity of

reference we need a criterion of identity, and the

general term associated with the name provides the

criterion.

Even for those people who would want

to assert that de Gaulle could turn into a tree or

a horse and still be de Gaulle, there must be some

identity criterion.

De Gaulle could not turn into

anything whatever, e.g., a prime number, and still

remain de Gaulle, and to say this is

to say that

some term or range of terms is analytically tied

to the name "de Gaulle".

(1969, p.167).

Now,

it seems to me that Searle is explicitly stating

here that being a mountain is

a necessary condition (an

essential property) for being Everest, being a person is

a

necessary condition

(an essential property)

for

being de

Gaulle. etc.

There are several points here which are worth

noting.

First, Searle does not seem (at least in this

quotation) to acknowledge a difference in kind between

essence of kinds and essence of individuals.

What is

essential for an object to be de Gaulle is for the object to

be an instance of a certain kind, viz. a person.

Of course,

this particular quotation does not rule out that Searle

accepts a distinction between essences of kinds and essences

 

of individuals, but I have found nothing to indicate that he

notes such a distinction.

And, as seen in the above quote,

vnen considering what is essential to de Gaulle, Searle

Lirst suggests that it

is his personhood, not, say, his

genetic make-up or historical origin.

A second point worth noting with regard

to

Searle’s

position is that he insists that a criterion of identity is

necessary for continuity of reference.

That is, we need to

be assured that we are referring to

the same thing when we

 
 

 

refer on different occasions.

With this I agree, but where

I

think Searle goes wrong is

that he seems to think that

this is

a fairly straightforward matter; a thing’s sameness

is given to us.

Our ordinary language apparently provides

us with no difficulty here because it reflects the thing’s

given sameness.

Granting Searle his claim that Everest or

de Gaulle couldn’t turn into just anything, Searle makes no

comment that Everest’s mountainhood or

de Gaulle’s

personhood is problematic.

His lack of comment here, I want

to

say,

is

indicative of

his acceptance of

his

‘realist’

conception of individuation and similarity.

Searle is

right, I think,

in pointing out that we need to consider a

thing’s identity in terms of its sameness, but he

is wrong

in failing to question the ‘givenness’ of this identity and

sameness.

That is, Searle fails to note the interest- and

language—dependence of

issues of individuation and

similarity.

Consequently,

in discussing

reference

to

an

object,

he stresses a conception of

the object’s essence,

since that is ultimately how its identity will be assured

(and, so, proper reference will be assured).6

 

As opposed to Searle, Kripke speaks frequently of

essences and essentialism.

He seems to carefully avoid

making any explicit commitment to essentialism, and Salmon

(1981) has argued that the causal theory of reference is not

committed to any non-trivial version of essentialism, but, I

want to argue, he is committed to essentialism all the same.

As we saw earlier, Kripke suggests that as science

 
 

progresses

it

discovers

underlying

traits,

and

hence

essences, of natural kinds.

And, as we saw earlier, Kripke

ties natural kind terms to proper names by considering them

both as rigid designators.

It

is with respect to this

notion of rigid designation that I see Kripke’s commitment

to essentialism, particularly the dependence of this notion

on a ‘realist’ conception of individuation and similarity.

To this end, we must look at Kripke’s discussion of the

necessary conditions for the identity of an object.7

Kripke offers several examples which, he says, suggest

essentialist principles.

One example is that of trying to

imagine ‘this very woman’ as having different parents than

she did.

He asks:

How could a person originating from different

parents, from a totally different sperm and egg,

be

this very woman?

   

One can imagine,

given the

woman, that various things in her life could have

changed: that she should have become a pauper;

that her royal blood should have been unknown, and

so on.

...But what is harder to imagine is her

being born of different parents.

It seems to me

that anything coming from a different origin would

not be this object.

(1980, p.113).

Another example is that of the origin of ‘this table’.

Kripke asks:

Now could this table have been made from a

 

completely different block of wood, or even of

water cleverly hardened into ice - water taken

from

the Thames River?

We could conceivably

discover that, contrary to what we now think, this

table is indeed made from the river.

But let us

suppose that it

is not.

Then,

though we can

imagine making a table out of another block of

wood or even from ice, identical in appearance

with this one, and though we could have put it in

this very position in the room, it seems to me

that this is not to imagine this table as made of

 

 

wood or ice,

but rather it

is

to imagine another

185

 
 

table, resembling this one in all external

details, made of another block of wood, or even of

ice. (1980, pp.113-114).

In footnotes to these examples, Kripke remarks that two

principles are suggested.

First, an ‘origin’ principle is

suggested, viz.

If

a material object has its origin from a

certain hunk of matter, it could not have had its origin in

any other matter.

Second, a ‘substance’ principle is

suggested, viz.

an object’s substance, not its origin,

is

what is essential to

it.

While Kripke does not explicitly

endorse either principle, and while he claims that a full

discussion of

the

problem of

essential properties

of

particulars is impossible here,

he offers several (obtuse)

points for consideration.

The first point is

that

ordinarily when we ask whether something might have happened

to

a given object, we ask whether the universe could have

gone on as it actually did up to a certain time, but diverge

in

its

history

from

that point forward

"so

that

the

vicissitudes of that object would have been different from

that time forth." (1980, p.115, note 57).

Perhaps, he says,

this feature [the history of the object?] should be erected

into a general principle about essence.

The second point

that Kripke makes

is

that he

is

not suggesting that only

origin and substantial makeup are essential.

For example,

had

the very block of wood from which this table was made

been used instead to make a vase, then the table would never

have existed.

So, says Kripke, being a table seems to be an

eSsential property of

the table.

8

The third point is that

 

 
 

the question of whether an object essentially has a certain

property can be vague [E].

The fourth point is that Kripke

is not convinced by

the counterexamples to

the origin

principle that appear in ordinary parlance.

The exact

analysis of those counterexamples, he says, is difficult and

he cannot discuss them here.

Now, it seems to me that while Kripke is careful not to

explicitly claim that he holds to essences, he nonetheless

does.

If we take an essential property of an object to be a

property which it is necessary for an object to have to be

that

object,

then clearly

Kripke

is

committed

to

essentialism.

His remark that "anything coming from a

different origin would not be

this object" and

that

this

table must have come from the block of wood which it did,

seem, despite his hesitance, to indicate that a particular

historical origin is at least a necessary condition for this

object to

be

this object.

Likewise, Kripke’s belief

that

science discovers the essence of kinds (at least ideally) is

also an obvious commitment to essentialism.

The connection

between these cases,

and

the underlying source of

the

commitment

to essentialism,

is

in Kripke’s

conception of

rigid designation.

As noted in chapter one, for Kripke, a

rigid designator refers to

the same object in all possible

worlds in which that thing exists.

What I want to argue,

and what I have been indicating in both this chapter and the

previous chapter, is that Kripke takes ‘the same object’ as

given and unproblematic.

in

the previous chapter I argued

 

 
 

 

that this is not the case; individuation and similarity are

not given and unproblematic, but are interest- and language-

dependent.

I

now want to argue, and have been trying to

indicate in

this chapter,

that

it

is

at

this point where

Kripke’s commitment to essentialism comes in.

That is,

I

want to argue that Kripke’s notion of rigid designation

commits him to essentialism.

Now, it seems to me that this claim is rather obvious.

Devitt takes a similar view.

He spends several pages

commenting on

the doctrine of essentialism and suggesting

that Kripke has made such a doctrine "both intelligible and

plausible." (1981, p.209).

Devitt remarks that Kripke

suggests

that

"a

person

has

such

[non-trivial]

essential

properties as

coming from a certain sperm and egg,

being

human, being of a certain sex, and being of a certain race."

(1981, p.208).

It is only after mentioning the doctrine of

essentialism that Devitt talks about rigid designation.

He

does so,

I believe, because the latter notion rests on the

former one.

He asks,

in virtue of what is

a name a rigid

designator, and answers in

the following way:

Suppose

we

take

‘name’

here

to

refer

to

designational names only.

Such a name designates

whatever

is

causally

linked

to

it

in

the

appropriate way.

It can only be causally linked

to

an object in

the actual world.

So

it

designates an object in another possible world

simply in virtue of that object being the same as

   

the object to which it

is causally linked in the

actual world.

It

is this that makes the name

rigid. (1981, p.212).

(It should be noted that for Devitt the important point here

is that of the scope of the modal operator and the roles of

188

 

 
 

 

singular terms.

He believes that the Kripkean account is

suspect because it rests on

the metaphor of ‘possible

worlds’.

He rejects this metaphor and instead suggests that

the

differing roles

of

names

and descriptions can be

accounted for in the following way: names do not give rise

to ambiguities of scope because they are designational,

whereas descriptions do give

rise

to ambiguities of scope

because they are attributive (in the Donnellan sense of

designational and

attributive).

My

interest here

is

different than Devitt’s and his rejection of Kripke’s

possible world talk is irrelevant to my point.

Now,

I am

not trying to make Devitt sound as if he interprets Kripke

as

I do, but to indicate that even he sees Kripke’s notion

of rigid designation as resting on the prior conception of

£33 £333 object being referred to across possible worlds.

From

my

perspective,

Devitt’s

reliance

on

the

designational/attributive distinction does not get him (or

Kripke) away from

the essentialist commitment of

the very

notion of rigid designation.)

Kripke introduces the notion of rigid designation and

then

immediately

discusses

another,

related,

notion:

transworld identity.

He spends several pages arguing

against the objection that

rigid designation rests on

criteria of transworld identity.

He claims that it

is

because we can refer (rigidly) to, say, Nixon, and stipulate

that we are speaking of what might have happened to him that

transworld identifications are unproblematic (in those

189

 
 

 

cases).

I think Kripke is right here - at least to the

extent

that

rigid

designation precedes

transworld

identification and

that if actual world identification is

unproblematic (for rigid designators) then so,

too,

is

transworld identification.

Where I think Kripke is wrong is

in

thinking

that

actual

world

identification

is

unproblematic, as

I argued in the previous chapter.

It is

also clear,

though, from this

that rigid designation rests

on a commitment to essentialism.

Rigid designation rests on

the possibility of being able to refer to the same object on

different occasions and on a name referring to the same

object on

all occasions.

As

we

saw,

for Kripke,

for

something to be ‘this object’, and for us

to speak of ‘the

same object’, requires (at least for names) some essential

property

(or

properties),

such as

genetic make-up

or

historical origin.

For Kripke, Nixon is Nixon and ‘Nixon’

refers to Nixon regardless of

the myriad of properties

associated with him which could have

been otherwise;

but

what makes Nixon Nixon and assures

that ‘Nixon’ refers

to

Nixon is that there is (in principle) a clearly identifiable

object which is Nixon and nothing else is that object.9

There must be some property (or properties) common to Nixon

in all possible worlds such that ‘Nixon’ refers to that

object in all

of

those worlds.

That property (or

properties) is what is essential to Nixon.

What else could

guarantee that Nixon is referred to

by ‘Nixon’ in all

possible worlds?10

 

 
 

 

‘Legitimate’ Essence

I have tried to show that Kripke’s notion of rigid
designation rests on a prior conception of sameness (i.e. of
individuation) and on the identification of the same object
in virtue of some essential property (or properties).11
me review the context in which I have made this claim.
have argued (a la Quine) that the doctrine of essentialism
a particular conception of individuation and

rests on

Let

I

similarity.

both essences of kinds
essences of individuals rest on these prior conceptions.

Furthermore,

and

As

I argued in

the previous chapter ‘realist’ conceptions of

individuation

and

similarity

are

mistaken,

and,

consequently, views of essentialism which are based on those

‘realist’ conceptions are also mistaken.

I then argued that

both Searle and Kripke are committed to essentialism, and

indeed, given their ‘realist’ positions on individuation and

similarity,

they are committed to

a particular version of

essentialism.

(To me, they are doubly wrong!)

But this is

not all.

In the next chapter I want to argue that

essentialism rests on

the doctrines of realism and

a

correspondence theory of

truth.

I will suggest that both

dOCCrines are mistaken.

Since both the cluster and the

Causal

theories of reference are based on these doctrines

(Which I will show), they, too, are mistaken.

Before turning to those issues, however, I want to make

good on my promise earlier in this chapter to explain what I

See

as

correct about

Searle s

and

Kripke 3

views

191

 

essences.

While I hold that for the most part they are

mistaken,

I

do

think that

they touch on particular cases

which match my intuitions.

The particular cases

(and

the

general cases which these examples instantiate) are the

following.

For Searle, Aristotle can’t be just anything.

An object must have some property, say, personhood, in order

for it

to be Aristotle.

That is, Aristotle can’t be a

mountain or

a prime number even though there may be many

 

properties which we associate with Aristotle (many

descriptions which we associate with the name ‘Aristotle’)

which turn out to

be false.

For example, Aristotle might

not have been the teacher of Alexander, but for ‘Aristotle’

to

refer

to Aristotle there must be

some property

(description)

which an object must have

to

be Aristotle.

That property (description) might be personhood.

The Kripke example was that a particular person, to be

that person, must have come from the parents which she did.

As Kripke put it, this X331 woman could not have come from a

totally different sperm and egg.

Likewise, this table could

not have come from a completely different origin than it did

and still be this table.

Now, while I maintain that both Searle and Kripke are

mistaken for the reasons which I outlined above in this

chapter, I

do share their intuitions somewhat.

I will

explain the sense in which I think they are correct by

discussing the examples.

What I see as correct about Searle’s example (and view)

192

 
 

 

 

is

his insistence that there must be some property

(description) associated with an object (a name) in order

for it

to

be what it

is and not something else.

As

I

mentioned above,

I

think he

is

right

that we need a

criterion of identity to have continuity of reference.

However, I insist that this is true only in a public sense

(i.e. within a public language).

That is, Searle is wrong

to claim that a speaker must have some description

associated with a name be true of a object for that name to

refer to that object.

Kripke’s counter-examples have shown

that that view is troublesome.

What I want to claim is that

the language community (not necessarily any individual

speaker) must have some description associated with a name

in order for that name to

be

the name of

a particular

object.

For example, if none of the descriptions associated

with ‘Aristotle’ by the entire language community were true

of Aristotle,

then,

I

insist, reference has not occurred.

Kripke rightly points out that language (and reference) is

public.

Indeed,

I think that Kripke’s notion of semantic

reference is meant in large part to capture what I am

claiming here

(although he dissociates it completely from

talk of descriptions associated with a name).

This ties in

with Searle’s discussion of essences in the sense that there

must be some property (description) which the language

community - not an individual speaker - associates with an

object (name).

If

the entire language community were

willing to give up personhood as

a necessary property for

193

 

 

something to have in order for it to be Aristotle, then, I

want to argue, being a person is not essential to Aristotle.

(For example,

this could be the case for Homer.

It might

turn out that there was no person Homer.

If

the language

community were to come to this conclusion, then being a

person would not be essential to Homer.)

A crucial point

connected with

what

I

am claiming here

is

that what

properties turn out to be essential (if any) are determined

by the language community.

This is

to deny the ‘realist’,

metaphysical sense

of

essences and essentialism.

If

Aristotle has an essence, it is not because Aristotle has an

essence; rather it is because the language community agrees

that

some

property

is

necessary

for

something

to

be

Aristotle.

Searle is right,

then,

but only to the extent

that

essences

are

public,

linguistic

(and

ultimately

epistemic) entities.

This insistence on publicity also underlies my

assessment of Kripke’s examples (and View).

My intuitions

are, with Kripke’s, that this 1351 woman could not have come

from a

different sperm and

egg

or this table from a

different block of wood.

However, I maintain that this is

the

case only within the context of publically-determined

individuals.

That is,

given that language is public and

given that individuation is linguistically determined (i.e.

given that the language community determines who this 1351

woman is and what this table is), then, yes, this XEEX woman

could not have come from a different sperm and egg.

Again,

 

 

 

194

 

however, what is being abandoned is any commitment to a

‘realist’, metaphysical conception of essences.

A possible

(and, I think, plausible) objection to what I have said here

is

that how we,

as

the language community,

determine who

EELE XEEZ woman is is by appeal to her origin, i.e.

to what

particular sperm and egg she came from.

Consequently, so

the objection goes,

it makes no difference in the analysis

to

insist

on

a

publically,

linguistically-determined

individuation (and essence).

The objection fails, though,

because it rests on the assumption that what counts as

the

origin of EELE 1551 woman is given ‘out there’ in the world

(i.e. a ‘realist’ view of origins).

Such an assumption is

untenable for the reasons which I claimed in

the previous

chapter concerning individuation and similarity.

That is,

histories are no more ‘out there’

than are individuals or

kinds.

As

I argued, what counts as a particular event-

individual is not ‘out there’, but is determined by our

theories/interests/language.

For example,

we decide what

counts as Esis ssis £255 (or as

the

1948 Presidential

election, etc.).

If this is true, then we cannot appeal to

the history (origin) of an individual (e.g. £313 £351 woman)

or

a kind as

though the history is not determined by our

theories/interests/language.

Past

event-individuals

and

thing-individuals are just as theory- and interest- and

language—dependent as

are present or future event-

individuals

and

thing—individuals

(e.g.

the

1948

Presidential election).

Hence,

the objection fails to the

 

 

 

extent that it relies on a ‘realist’ conception of history

(origins).

Again, given that we, as the language community,

have determined what the origin of Eflli zssy woman is, then

we can appeal to her origin to individuate her.

I want to make one final point in connection with what

I have just said.

Earlier I commented that the distinction

between individuals and kinds (and essences of each) might

not be so sharp.

I mentioned that recent debate in the

philosophy of biology has included the consideration that

species can be treated as individuals.

This debate is

related

to

the present discussion of

the origin (and

essence) of individuals because in large part the argument

for treating species as individuals rests on looking at the

history of

given

species.

That

is,

an argument for

considering species as individuals runs along the following

lines.

Just

as

for

‘normal’

individuals,

for which

spatiotemporal continuity tends to be the criterion (or, at

least,

one criterion) for individuation,

so,

too,

for

species,

spatiotemporal

continuity

should

be

the

(a)

criterion of individuation.

Phenotypic features, even

phenotypic similarity, is irrelevant for individuation.

For

example, identical twins might be phenotypically identical,

but they are, nonetheless separate individuals.

On the

other hand,

an individual can change (drastically)

phenotypically yet remain

the

same individual (e.g.

a

caterpillar/butterfly).

What

does/should

matter

for

individuation is

historical origin and spatiotemporal

196

 

  

 

continuity.

As Hull puts it: "Organisms are not included in

the same species because they are similar to

the

type

specimen or to each other but because they are part of the

same chunk of the genealogical nexus." (1978, p.353).

Just

as

we

treat mass

terms (such as ‘gold’ or ‘water’) as an

individual, so we can/should treat species as individuals.

In both cases, just as with ‘normal’ individuals, particular

hunks of gold or particular organisms are part of the whole

(the individual) because of sameness of origin, not sameness

of

properties.

This

sounds

as

though it supports a

Kripkean-like analysis of

individuals and

of essences

(because having a particular historical origin can serve as

the essential property).

However, it does so only under the

constraints noted above concerning the public language—

dependence of individuation both synchronically and

diachronically.

While I think that a fuller treatment of

the issue of species as individuals can shed some light on

the issue of essentialism,

the point I have just made

concerning it will suffice for the present purposes.

 

197

 

  

 

 

NOTES

1I use ‘object’ here to refer to both individuals and

kinds.

For example, as I will note below, species can be

thought of as individuals.

As will be seen, the distinction

between individuals and kinds has been important.

It is a

distinction which I reject (at least as

it

is usually

presented).

21 am not making an ‘existence’ restriction here.

That

is,

I am not saying that for an object to have essential

properties,

it must have them in order for that object to

exist, but simply in order for it to be what it is.

I take

the

issue

of

essences

and essentialism as

issues

of

identity, not of existence.

31 will return to this issue, and strengthen my claims

here, shortly.

First I need to explicitly address the

distinction between individuals and kinds.

4Extensionally, there are often cases where kinds are

indistinguishable from

individuals

that

make

up

the kind

(e.g. ‘even prime numbers’ and ‘ ’).

5Actually,

there

is

a sense in which I

do not think

that Searle is committed to essentialism here.

I will make

this explicit at the end of this chapter when I discuss what

I see as a legitimate way of talking about essentialism.

6This ‘realist’ commitment is borne out in such claims

as the following: "In order that a name should ever come to

be used to refer to an object in the first place there must

be some independent representation of the object.

This may

be by way of perception, memory, definite description, etc.,

but

there must be enough Intentional content to identify

which object the name is attached to." (1983, p.259).

7I should add at this point that, just as I see much of

what Searle claims as correct to an extent, so,

too, I see

much of what Kripke claims as correct to an extent.

At the

end of this chapter I will argue for what is basically right

about both of their views on essences.

First, however, just

as I criticized Searle, I want to now criticize Kripke.

8This sounds like Searle’s claim that being a person is

analytically true of this person (e.g. de Gaulle).

90f course, for Kripke, "a clearly identifiable object"

is meant metaphysically.

Nixon is Nixon whether we know it

or not, and ‘Nixon’ is a rigid designator because it refers

to that object in all possible worlds.

10An important objection that might be raised here is

 

198

 

that I have misrepresented Kripke’s position when I accuse

him of being committed to essentialism through the notion of

rigid designation.

I claim that there must be some property

(or properties) common to, say, Nixon in all possible worlds

such that ‘Nixon’ refers to that object in all of those

worlds.

That property (or properties), I suggest, is what

is essential to Nixon.

What else, I asked, could guarantee

that Nixon is referred to by ‘Nixon’ in all possible worlds?

One answer to this question - other than the requirement of

essential properties - is that Kripke simply stipulates that

‘Nixon’ refers to

the same object in all possible worlds.

Indeed,

this

seems

to

be Kripke’s explicit position.

He

says:

Don’t ask:

how can

I identify this

table

in

another possible world, except by its properties?

I have the table in my hands, I can point to it,

and when I ask whether ii might have been in

another room,

I am talking, by definition, about

it.

I don’t have to identify it after seeing it

through a telescope.

If I am talking about it,

I

am talking about ii, in the same way as when I say

that our hands might have been painted green,

I

have stipulated that I am talking about greenness.

Some properties of an object may be essential to

it, in that it could not have failed to have them.

But these properties are not used to identify the

object in another possible world, for such an

identification is

not needed.

Nor need

the

essential

properties

of

an

object

be

the

properties used to identify it

in

the actual

world,

if indeed it

is identified in

the actual

world by means of properties (I have up to now

left the question open).

...So, we

do not begin

with worlds (which are supposed somehow to

be

real, and whose qualities, but not whose objects,

are perceptible to us), and then ask about

criteria of transworld identification; on

the

contrary,

we begin with objects,

which we

have,

and can identify, in the actual world.

We can

then ask whether certain things might have been

true of the objects. (1980. pp.52-53).

Now,

I want to say (repeat) several things about this

view.

As

I have said above,

the issue

that I see as

the

important

one

is

not

that

of

transworld

identity

or

transworld identification or transworld individuation, but

of identity/identification/individuation in this world.

Kripke’ 5 quote seems to me to indicate that we start with an

identified object ("I am talking, by definition, about is")

and we can then simply stipulate that we are talking about

that object.

Much of what I am arguing for in this chapter

and the previous one

is that Kripke is mistaken in

suggesting that he can do just that.

(This will come out

clearer in the final two chapters.)

When Kripke says that

he is, by definition, talking about ii, I want to ask: about

 

 

199

 

what?

Kripke seems to presume that we can clearly separate

out issues of individuation from issues of identification.

I want to suggest that we can’t.

We cannot, contra Kripke,

simply stipulate that we are talking about, say, Nixon and

presume that we are referring to Nixon in all possible

worlds merely by using the name ‘Nixon’.

Indeed, I am

somewhat perplexed

that Kripke would want to make such a

move, since it seems (to me, anyway) that this divorces the

name from any public, historical, linguistic context - the

very point for which he criticizes Searle’s theory of names.

While

it

may

be

true

that

a move

to postulating the

stipulation of (the identity of) an object across possible

worlds

relieves Kripke from a commitment to essentialism,

the cost, I would want to argue, is too great.

The issue is

whether or not one can stipulate in this world the identity

(or individuation) of an object.

To

the extent that this

can be done, it is possible only given a public language and

only to the extent that the language community accepts such

a stipulation.

As Kripke himself says,

I cannot refer to

Schmidt by (the use of) ‘Gddel’.

Nor, I want to add, can I

do so by stipulating that I am referring to Schmidt by (the

use of) ‘deel’ (unless the language community agrees, and,

as a result, ‘deel’ becomes another name for Schmidt).

Additionally, I cannot separate out

the issues of

the

individuation of an object (say, Nixon) from the issues of

the identification of that object.

That is, I think Kripke

is simply wrong to claim that he is talking about ii (say,

Nixon) and

this

makes no commitment to talking about the

means of identifying it.

When Kripke says he’s talking

about it,

and I ask, about what, Kripke cannot respond by

 

saying—Ehat my question misses the point because it

is

asking for the identification of the object that has already

been individuated (by stipulation).

Kripke’s position seems

to be committed to the following: in a given possible world

Nixon is not only not a president, but Nixon is not even a

human; rather Nixon is

a particular dime.

When I ask: how

do you know that this dime is Nixon and not that dime,

 

 

Kripke’s answer, it seems to me, can only be: because I said

so!

If indeed this

is Kripke’s move to avoid a commitment

to essentialism, then I suggest that he has avoided the

commitment, but only at the cost of incoherence.

11Of course, what I have said about rigid designation

and essentialism with respect to names applies mutatis

mutandis with respect to natural kinds.

 

CHAPTER SEVEN

CORRESPONDENCE AND REALISM

I

have

been arguing that questions and

issues

about

reference can be addressed only within the context of prior

questions

and

issues

about

individuation and

similarity.

Theories of reference,

I have claimed, presuppose and are

committed to theories of individuation and similarity.

In

particular,

I have maintained that the cluster and causal

theories of reference presuppose and are committed to a

‘realist’ theory of individuation and similarity.

This is

borne out in their commitment to private reference and

essentialism.

I have also been arguing that the issues of

individuation and

similarity are

ultimately epistemic

issues.

Underlying (and occasionally surfacing from) what I

have been saying is the view that metaphysical and semantic

issues and positions rest on epistemic concerns.

That is,

the basic metaphysical issues of individuation and

similarity must finally be addressed within the context of

asking how we

pick out individuals,

how we partition the

world.

What

things

there

are

in

the world and what

properties they have are,

in the final analysis, dependent

upon our language and cognitive interests.

While this seems

to

be a metaphysical claim, it

is only meaningful, I hold,

as an epistemic one.

(It makes no sense to say that there

201

 

 

 

are an even number of stars in the universe if we can never

know this.)

This openly verificationist position has been

(for

the most part)

in

the wings

throughout the previous

chapters.

It will come on stage in its fullest glory in the

next.

Besides suggesting that metaphysical issues ultimately

rest on epistemic priorities, I am maintaining that semantic

issues, too, must be framed within an epistemic context.

In

particular, issues of reference presuppose and are committed

to epistemic theories.

Likewise, views on truth depend upon

views on knowledge.

In this section I want to argue that

the cluster and causal theories of reference are committed

to

a correspondence theory of truth, which itself entails

realism

(in

the

sense

of

a

particular

conception of

individuation and similarity).

That conception, I have

argued, is mistaken.

I now want to suggest that theories of

realism and correspondence which

rest

on

it are also

mistaken.

Furthermore, theories of reference which rest on

those theories of realism and correspondence are mistaken as

well.

I will first lay out what I

take a correspondence

theory of

truth to

be

(and will discuss Tarski’s semantic

theory of truth in this context).

I will argue that such a

theory rests on a commitment to realism.

After explicating

what I take realism to be,

I will criticize it.

Having put

forth a critique of realism and correspondence, I will argue

that both Searle and Kripke are committed to correspondence

(and hence realism),

suggesting that they are consequently

202

 

 

weakened.

Finally,

I will conclude that the received

theories of reference are, because of their commitments to

the same faulty epistemic positions, mistaken and inadequate

theories of reference.

Truth

What is

a correspondence theory of

truth?

Quite

simply,

it

is

a

theory

that

takes

the

truth of

a

proposition1 to consist not in its relations with other

propositions, but in its relations to the world, to facts.

A proposition is true if and only if

it corresponds to the

facts of the world.

‘Snow is white’ is true because snow is

white.

Of course,

this definition is vacuous without some

explication of

what

‘correspondence’

means.

Perhaps

the

most famous contemporary explication is Wittgenstein’s

(1922)

picture theory of meaning,

in which correspondence

consists in an ordered isomorphism between words in (true)

atomic

propositions

and

(simple)

facts

(or

states

of

affairs) in the world.2

An

older,

and

perhaps

even more

famous,

explication

(or,

at

least,

characterization)

of

correspondence

comes

from Aristotle (Metaphysics Gamma, 7, 27):

"To say of what

is that it

is not, or of what is not that it is,

is false,

while to say of what is that it is,

or of what is not that

it

is not,

is true."

While Aristotle apparently takes the

nature of this correspondence as a given, Tarski (1956)

tries to give an account of truth which captures this sense

of ‘true’.

Tarski’s conception, and theory, of truth,

 

 ——l

however, have come to be known as a semantic theory of

truth, rather than a correspondence theory of truth.

I want

to lay out what Tarski’s theory is and consider the claim

that it is actually a version of a correspondence theory of

truth.

I want to argue that appealing to Tarski’s theory to

shore up and explicate the correspondence theory of truth

fails on two grounds.

First, even if it is a version of the

correspondence theory (and is not philosophically neutral),

then I think the theory begs the question.

Second, I think

it can be forcefully argued that Tarski’s theory of truth is

philosophically neutral and so does not do

the job of

shoring up or explicating a correspondence theory.

Tarski states that in discussing the concept of truth

in formalized languages his task was to construct "with

reference to

a given language"

(1956, p.152) a materially

adequate and formally correct definition of

the term ‘true

sentence’.

He

remarks

that

he

is

concerned "exclusively

with grasping the intentions which are contained in the so-

called classical conception of truth (‘true - corresponding

to reality’)" (1956, p.153).3

Indeed, when discussing the

concept of truth in colloquial language, he claims that a

semantic definition of ‘true sentence’ can be expressed as

follows:

A

true sentence is

one which says

that the

state of affairs is

so and so, and the state of

affairs indeed is so and so.

The task of a semantic definition is

to make this intuitive

meaning more precise and clear.

He suggests several means

204

 

 _7—W

of

doing

this.

First, quotation-mark names, denoting the

names of sentences, could be used (e.g. ‘It is snowing’ is a

true sentence if and only if

it

is snowing).

Second,

structural-descriptive

sssss,

describing

the

words

which

compose the expression denoted by

the name along with the

signs of which each word is composed and the order of

the

words, could be used (e.g. An expression consisting of three

words of which the first is composed of the two letters El

and t

o e
l

(in that order), the second word of the letters s1

and e

n
o

s

(in that order), and the third word of

the letters

ess, 53’ s, double-u, sz, 33 and
 

gee (in that order), is a

true sentence if and only if

it

is snowing).

However,

because of semantic paradoxes (such as the liar paradox, in

which a sentence is true if and only if it is false), Tarski

concludes that:

the very possibility of

a consistent use of

the

expression ‘true sentence’ which is in harmony

with the laws of logic and the spirit of everyday

language

seems

to

be

very questionable,

and

consequently

the

same

doubt

attaches

to

the

possibility of constructing a correct definition

of this expression. (1956, p.165).

Leaving natural languages aside, then, Tarski proceeds

to define the notion of truth for formalized languages (i.e.

artificial languages in which the sense of every expression

is unambiguously determined by its form).

In particular, he

restricts himself to the calculus of classes.

Before

laying

out

a

conception

of

truth,

Tarski

distinguishes between object language and metalanguage.

An

object language is

a language in which we speak; a

205

 

 

metalanguage is

a language used to speak about the object

language. With regard to truth,

the object language is the

language iss which truth is defined and the metalanguage is

the language is which truth is defined.

For example,

to

say, “Noseworthy looks bad today" is to speak a sentence in

an object language.

However, to say,

today’ contains five

"x

letters"

is

to speak metalinguistically about that object

language.

(As this example shows, natural languages, such

as English, serve as both object language and metalanguage.

This

is

a reason why semantic paradoxes such as

the liar

paradox are possible, and hence, why a semantic definition

of

truth cannot,

for Tarski,

be given for natural

languages.)

Following Haack (1978), we can say that there are two

parts to Tarski’s theory: (1) adequacy conditions (i.e.

conditions which any theory of truth must/ought to fulfill)

and (2) a definition of truth (at least for the calculus of

classes), which Tarski claims is adequate.4

There are

two adequacy conditions to

be met.

First,

material adequacy (i.e. setting the limits on the possible

content) and, second, formal correctness (i.e. establishing

the possible form of any acceptable definition of truth).

The spirit of the material adequacy condition was noted

above in remarking on Tarski’s discussion of truth in

natural languages.

Briefly, the material adequacy condition

says that any acceptable definition of truth should have as

a consequence all instances of Convention T, that being:

 

 

(T) S is true iff p,5

where ‘p’ is referred to by ‘8’ (either as a quotation-mark

name or as a structural-descriptive name).

Convention T is

actually a schema, as ‘p’ can be replaced by any sentence of

the object language (i.e.

the language for which truth is

being defined) and ‘S’

is

to be replaced by a name of

the

sentence which replaces ‘p’.

An example is "‘Snow is white’

is

true

iff

snow

is white."

Of

course,

for Tarski,

Convention T

is 533 a definition of truth, but simply a

material adequacy condition, which states that all instances

of

it must be entailed by any definition of truth which is

to count as being materially adequate.6

Tarski places this

stipulation on definitions of truth, again, because he wants

to capture the classical (i.e. correspondence) conception of

truth in his semantic theory.

The second adequacy condition, formal correctness,

follows from Tarski’s commitment to avoid the semantic

paradoxes noted above.

Assuming that the laws of logic

hold,

these

semantic

paradoxes

arise

because

natural

languages are semantically closed

(i.e.

the languages

contain

not

only

expressions,

but

also

the

means

of

referring

to

those

expressions,

along

with

such semantic

predicates as ‘true’).

While the object language/

metalanguage distinction resolves this problem (by yielding

semantically open languages),

a corollary is

that

the

metalanguage,

M,

must contain

the object language,

0,

or

translations of all sentences of 0, plus the means to refer

207

 

 

to

the expressions of 0 (since all equivalences of

the

schema Convention T must be implied by

the definition of

truth, which will occur in M).

M must have both 0 (or

translations of O) and the means to refer to the expressions

in 0 because instances of

T have an expression denoting a

sentence of O (viz. S) and a (translation of a) sentence of

0 (viz. p).

Additionally, for Tarski, the structure of the

object language and the metalanguage must be ‘formally

specifiable’, because in order to define truth in the object

language it must be possible (and is essential) to pick out

the wffs of

the object language (since they are what truth

in the object language, what ‘true-in-O’, applies to).

Before turning to Tarski’s definition of truth (in O),

a final word needs to be said about Convention T.

It might

be (and sometimes is) thought that Convention T provides an

obvious way of giving a definition of truth.

Two ways are

proposed, both motivated perhaps because Tarski admits that

each instance of Convention T could be thought of as a

partial definition of

truth.

First,

it might be

thought

that a conjunction of all the instances of T (one for each

sentence

in

0)

would constitute a complete definition of

truth (in

0).

However, Tarski claims that this is

impossible because

the number of sentences of

0 might be

infinite, so the conjunction could never be formed.

Second,

it might be thought that Convention I could be stated as a

universal

quantification

and

could

thus

constitute a

definition.

(So,

the definition D might be: (p)(S is true-

 

in-O iff p), where S=‘p’.)

This, however, is unacceptable

because

quantifying

into quotation marks

is

meaningless.

Snow is white is no more a part of ‘Snow is white’ than rat

is a part of ‘Socrates’.

Rejecting these proposals, Tarski

offers his own definition of truth (in 0).

Tarski provides his definition of truth by

the

following steps:

specifying the syntactic structure of

the

object language, 0;

specifying the syntactic structure of

the metalanguage, M (including expressions of

0,

or their

translations,

syntactic vocabulary,

and

the usual logical

apparatus);

defining

the

notion

of

satisfaction;

and

defining the notion of truth (in terms of satisfaction).7

The following characterization can serve as an example

of

a

Tarski—like

presentation:

(1) Syntax of object language 0

Individual constants: a,

b,

c,

Variables: x,

y,

z,

Predicate letters: F, G,

Logical connectives: ',

&

Quantifier: ( ...)

Grammatical punctuation: (,

)

(2) Syntax of metalanguage M

Atomic sentences of

O are strings of expressions

which consist of an n-place predicate followed by

n individual constants or variables.

All atomic sentences are wffs.

If A is a wff, then ~A is a wff.

If A and B are wffs, then (A&B) is g wff.

If A is a wff, then ( x)A is a wff.

Nothing else is a wff.

(3) Definition of satisfaction

Open sentences (i.e. sentences containing free

variables) are not

true

or false,

but are

satisfied.

For example,

‘x

i

a number’

is

209

 

 

satisfied by 2,

‘x is greater than y’ is satisfied

by <3,2>,

‘x

is between y and z’

is satisfied by

(2,1,3), where ‘(...>’ indicates an ordered n-

tuple of

the n objects appearing between the

brackets.

Satisfaction is a relation between open

sentences and ordered n-tuples of objects.

Since

open sentences may have any number of free

variables,

Tarski

defines

satisfaction

as

a

relation between open sentences and infinite

sequences.

(So, F(x1...n) can be satisfied by

(0 ...on, °n+1"°> just in case it is satisfied by

the

first n members of

the sequence;

subsequent

members can be ignored.)

The negation of an open sentence S is satisfied by

those sequences which do not satisfy S.

The conjunction of two open sentences S1 and $2 is

satisfied by those sequences which satisfy both 81

and $2.

The existential quantification of an open sentence

S

is satisfied by

a sequence of objects just in

case

there

is

some other sequence of objects

(differing from the first sequence in at most the

ith place, where the ith is the variable bound by

the quantifier) which satisfies the open sentence

S

resulting from dropping the quantifier.

(For

example,

the

sequence <4,1,3,...>

satisfies

the

sentence ‘( x)(x is between y and z)’ because

<2,1,3> satisfies ‘x

is between y and z’ and

<2,l,3> differs from <4,1,3,...> in

at most the

first place,

i.e.

the variable

bound

by

the

quantifier.)

(4) Definition of truth

Given a definition of satisfaction, Tarski quickly

offers a definition of truth.

A sentence S

is

true just in case it is satisfied by all sequencss

and false just in case it is satisfied by none.

For example,

‘x

is

a number’ is satisfied by

<2,...> regardless of

the second or subsequent

members of the sequence; ‘x is greater than y’ is

satisfied by (3,2,...) regardless of the third or

subsequent members of

the sequence;

( x)(x is

a

number)

is

satisfied by <...,...>

regardless

of

the first or subsequent members of

the sequence,

since there is

a sequence (e.g. <2,...>) which

differs from any arbitrary sequence in at most the

first place and which satisfies ‘x

is

a number’.

Any

closed sequence,

then,

will be satisfied by

all sequences or by none.

Hence, truth is defined

in terms of the satisfaction of all sequences.

210

 

While

explicitly

restricted

to formalized languages,

Tarski’s semantic definition and conception of

truth have

been profoundly

influential

in analytic philosophy.

For

example, the theory has been taken to be a vindication and

legitimation of

a correspondence theory of

truth.

Popper

(1963), for instance, says: "Tarski’s greatest achievement,

and the real significance of his theory for the philosophy

of the empirical sciences lies, I believe, in the fact that

he

re-established a

correspondence

theory

of

absolute or

objective truth which showed

that we are free

to use the

intuitive idea of truth as correspondence with the facts."

(1963,

p.223).

Now,

assuming that

this view is correct,

there are,

I want to claim, serious problems for a

correspondence

theory

of

truth as exemplified by

the

semantic

theory.

As stated above, for a correspondence theory of truth,

a proposition is true if and only if

it corresponds to the

facts

of

the world.

Such a claim seems almost a truism,

and, as mentioned, the onus is

to lay out exactly what

‘correspondence’ means.

This involves not only stating the

nature of

the correspondence,

but also indicating the

objects which correspond.

As noted above, the nature of the

correspondence is a structural isomorphism between words and

facts.

Of course, this is not very helpful if one wonders

about correspondence in the first place.

It is, at best, a

restatement of

the claim that a proposition is true if and

only if

it corresponds to the facts.

It still remains to

 

 

say what propositions are, what facts are and what

correspondence is.

While the question of what propositions

are has received a great deal of attention, it

is not

crucial to the present discussion, so it will be left aside.

The question of what facts are has also received much

attention, and it will only be addressed here in the sense

of saying that two related issues must be considered when

addressing this question.

One

issue to

be considered is

what are facts (i.e. what does ‘facts’ mean?); another issue

to be considered is what are £33 facts (i.e. what facts

obtain?).

In the next chapter I will address these issues

and suggest that the two issues are intimately interrelated,

such that the first cannot be answered without answering the

second.

For now, I will simply say that I take the answer

to both to be epistemic.

In the next chapter I will explain

what I mean.

As we saw,

truth,

for Tarski,

is defined in terms of

satisfaction.

A sentence is

true

(in a language) just in

case it

is

satisfied by

all sequences

(in

the language).

Satisfaction is

a relation between open sentences and

ordered n—tuples of objects.

Now,

in one sense,

it

is

almost

trivially true

that

this conception of

truth is

a

correspondence theory.

Correspondence (intuitively captured

by Convention T)

is revealed simply as satisfaction.

‘Snow

is white’

corresponds

to

the

facts

(of

snow being white)

quite simply because there is

a relation between ‘Snow is

white’ and (snow, white(ness),...>, viz.

the latter

 

satisfies the former.

While it might be claimed that Tarski’s theory has

fortressed the correspondence theory of truth by giving a

precise definition of truth and a clear explication of the

nature of correspondence, it has been claimed that Tarski’s

theory does not bolster the correspondence theory, rather it

presupposes

the

correspondence

theory.

The argument for

this claim also seems somewhat obvious.

Tarski’s explicit

acknowledgement that he wants to capture the intuitions of

the classical (i.e. correspondence) theory seems to indicate

a clear commitment to the latter.

Beyond this, though,

there are further reasons for asserting that Tarski

presupposes a correspondence

theory.

The major argument

goes as follows: The very notion of satisfaction requires a

specifiable set of objects to which statements are related.

That is,

without identifiable

(and individuated)

objects,

the notion of satisfaction would make no sense because there

would be nothing to which statements could be related.

The

sequence <x,y> could not satisfy the statement (3x)ny if

there were no objects x and y.

(This is not an ontological

claim about what kind of objects x and

y are,

but a

conceptual claim about the plausibility of

the notion of

satisfaction.)

So, for satisfaction to get off the ground,

there must be a prior set of specifiable objects.

But what

does this mean to say that there must be a prior set of

specifiable

objects?

Doesn’t

this

mean

that

for

satisfaction to be possible, it must be true that there is a

 

prior set of specifiable objects?

If

so, we already have

the concept of truth built in to the notion of satisfaction.

Appealing to satisfaction,

then,

to define

the notion of

truth is question-begging.

It might be objected that it

is not truth that is

presupposed by satisfaction, but simply a set of objects.

That is, what is required for satisfaction to make sense and

be applicable is only objects, not conceptions of truth.

I

think this

is

false.

The following example

(I

hope)

illustrates why.

Suppose we say that (3x)ny is satisfied

by <glub,

sosh> (and R stands for the relation ‘is flibber

than’).

In what sense can we say that glub is flibber than

sosh (or light is faster than sound or two is greater than

one)?

The answer,

it seems to me,

is because glub is

flibber than sosh.

But, then, the notion of satisfaction is

superfluous, as we’ve already said that glub is flibber than

sosh in order to say that <glub, sosh> satisfies ‘x

is

flibber than y’.

Another objection to what

I

have suggested is

that,

because

truth and satisfaction are defined in

the

metalanguage rather than in the object language, there is no

question-begging in

the semantic conception of

truth.

Indeed,

it

is for

this reason that Tarski claims that the

semantic conception is neutral with respect to metaphysical

(and epistemic)

positions.

That is,

truth in

the object

language is philosophically neutral because any commitment

to

or

connection with

other

philosophical

considerations

 

takes place in the metalanguage, not in the object language.

So, even if it were the case that the notion of satisfaction

is philosophically ‘loaded’, this would have no bearing on

truth in

the object language; satisfaction is ‘1oaded’ in

the metalanguage, not in the object language.

With respect

to

the object language,

then,

the semantic theory is just

that -

a semantic theory, with no commitments to

a

correspondence (or any other) theory.

This is

a powerful objection, and I am sympathetic to

it somewhat.

However, in the end I think it

is either

mistaken or it removes the semantic theory from redeeming a

correspondence

theory.

Speaking metaphorically,

I

think

such an objection (at best) wins the battle, but loses the

war.

(Or,

perhaps,

more topically: wins

the debate,

but

loses the election.)

Assuming that the objection is sound,

it

is just such an argument, I believe,

that has led many

writers to complain that the semantic theory is all form and

no substance.

Precisely because the semantic theory is

(said

to

be) neutral to metaphysical and epistemic issues

and positions, it

is said to

be

of little interest and

value.

After all, why do we want a theory of truth?

Isn’t

it so we know what truth is and know what is true?

That is,

why do we care about a theory of truth if not to give us a

criterion for determining whether or not a claim is

true?

(So,

a theory of truth will tell us whether or not ‘Galileo

invented the microscope’ is true.)12

If this is why we are

interested in a theory of truth, then how does the semantic

215

 

 

theory fare?

Not well, I want to say, unless it presupposes

a correspondence theory

(and even

then not so well).

Clearly, if we want to know whether or not ‘Galileo invented

the microscope’ is true, it is of no value to be told that

‘Galileo invented the microscope’ is true just in case

Galileo invented the microscope (or that ‘Galileo invented

the microscope’ is true just in case gold has atomic number

79).

Nor is

it

of any value to be told that ‘Galileo

invented the microscope’ is true just in case ‘Galileo

invented the microscope’ is satisfied by all sequences in

the language.

We want to know whether or not Galileo

invented

the microscope,

and

the semantic theory does

nothing to answer this.

(This does not mean that the

semantic theory is of no value at all, but that it does not

do

the job

that we intuitively want a theory of

truth to

do.)

It might be objected that the semantic theory 2223 give

a definition for

‘truth’

and so gsss serve as

a adequate

theory of truth.

It might not tell us what is

true

(i.e.

what ‘facts’ are

true), but it does tell us what truth is

(i.e. what ‘true sentence’ means/refers to).

But note that

the definition of truth in the semantic theory depends upon

the notion of

satisfaction, which in

turn depends upon a

specifiable

set

of

identifiable

individuals

(i.e.

on

the

fact of there being these individuals).

But this is simply

to hold that the truth of

the world is that this set of

individuals

is

the

set

of

individuals

upon

which

216

 

satisfaction depends (i.e. that it is true that this set of

individuals are in the world in order for satisfaction to

obtain).

Related to

this last point, Putnam (1981)

has argued

that the semantic theory of truth is philosophically neutral

(i.e.

it

is consistent with any theory) because it misses

the

interesting philosophical problems.

As

he

says,

the

problem is not

that we don’t understand ‘Snow is white’.

Rather, the problem is that we don’t understand what it

is

to understand ‘Snow is white’.

What is needed for a theory

of

truth

to

do

its

job

is

some

standard of rational

acceptability, of what is considered a rational way to

pursue an inquiry, what standards of objectivity there are,

when it

is rational to terminate an inquiry, what grounds

will be regarded as providing good reason for accepting one

verdict over another, etc.

In a nutshell, truth is the

idealization of rational acceptability.

Says Putnam:

..truth

is

an

idealization

of

rational

acceptability.

We speak as if there were such

things as epistemically ideal conditions, and we

call a statement ‘true’ if

it would be justified

under such conditions.

‘Epistemically ideal

conditions’, of course, are like ‘frictionless

planes’: we cannot really attain epistemically

ideal conditions,

or even be absolutely certain

that we have come sufficiently close to them.

But

frictionless planes cannot really be attained

either,

and yet

talk of frictionless planes has

‘cash value’ because we can approximate them to a

very high degree of approximation.

(1981, p.55).

(I mention Putnam’s claims here primarily to indicate the

sort of view about truth which I think is correct and which

I will elaborate upon and defend in the next chapter.)

217

 

The upshot of all this is the following.

Tarski’s

semantic theory

of

truth

in which truth depends upon

satisfaction is either committed to a correspondence theory

(if satisfaction takes place as a hook-up between the object

language and the world) or is philosophically uninteresting

(if satisfaction takes place as a hook-up between the object

language and the metalanguage).

In the former case,

it is

committed to correspondence because satisfaction is simply

another way of trying to lay out the structural isomorphism

between words and objects and presupposes a specifiable set

of identifiable individuals (furthermore, this former case,

I think, is mistaken as Putnam suggests); in the latter

case,

it

is philosophically uninteresting because it fails

to do what we want a theory of truth to do (i.e. give us a

means of distinguishing true and false sentences, or helping

us determine what is true).

We are still left with trying

to account for truth for Searle and Kripke. It is

to this

which I now turn.

Reference 329 Realism

As

with

essentialism,

Searle

says

little

explicitly

about truth and even less about a theory of truth.

Accordingly, we must cull from brief comments and passages

the

sort of theory of truth to which he

is committed.

Within the context of speech act analysis, Searle discusses

the notion of truth.

He claims that we characteristically

call something true "only if

a comment, remark, assertion,

statement, or hypothesis, or the like, has already been made

 

 r——————————————————————————————————————————————————————————————----------::::1

or

is at least in some way under consideration; in short,

only if a proposition is already in the offing." (1969,

p.153).

For example, if your house is

on fire,

I do not

announce to you that it is true that your house is on fire.

We affix the ‘it is true that’ prefix to a proposition only

in a context in which that proposition is already being

entertained.

Because of this, and because a proposition is

characteristically entertained or put under consideration by

the performance of some illocutionary act, then calling

something true places us

in

a certain relation to

the

initial illocutionary act (e.g.

in a relation of agreement,

disagreement,

endorsement, etc.).

However,

this

does not

give us a theory of truth or help us determine what truth is

(or what is true).

To get a sense of what view of truth Searle is

committed to, we must turn to a distinction made within his

theory of reference, viz.

the primary-secondary aspects

under which reference is made.

Recall that for Searle, the

primary aspect under which reference is made is that aspect

which, if not satisfied, would yield a statement that cannot

be

true.

The secondary aspect is any aspect which the

speaker expresses such that the speaker utters it in an

attempt to secure reference

to

the object which satisfies

the primary aspect, but which is not intended as part of the

truth conditions of

the statement the speaker is attempting

to make.

For example,

the speaker, looking at someone in

the room, says, ‘Smith’s murderer is insane.’

Now, the

 

 

speaker and the hearer might agree that the speaker has

referred to,

and made a true statement about,

that

particular person being looked at even though that person

(and perhaps everyone) fails to satisfy the expression

‘Smith’s murderer’.13

The speaker could,

on demand,

fall

back on another aspect, say, one expressed by ‘the person I

am looking at’.

If it turns out that there is no one being

looked at,

only a hologram perhaps,

then the speaker could

fall back on another aspect, say, one expressed by ‘the

person arrested by

the police and accused by

the District

Attorney as Smith’s murderer’.

If

it turns out that there

is

no such person, then the speaker could fall back on

another aspect.

Eventually, however, an aspect must be

reached such that if no one satisfied it, then the statement

could not be true (nor reference be made).

The primary

aspect under which reference is made is

the ‘last’ aspect,

the aspect which either works or results in a statement

which is false.

The other aspects are secondary.

What is important about this in the present context is

that this treatment of aspects under which reference is made

indicates a commitment to

a particular theory of truth,

indeed a Tarski-like theory of truth.

As Searle says:

there must always be some...aspect under which the

speaker could have referred to

the object and

which

is satisfied by

the object.

Furthermore,

this aspect is such that if nothing satisfies it

the statement cannot be true. (1979, p.144).

This insistence on there being a match in terms of

satisfaction between the object in the world and the primary

220

 

aspect

under which reference is made

is certainly

Tarskiesque, and,

I want to argue, not accidentally.

Just

as

the semantic theory of truth can be construed as being

philosophically

neutral

(and

consequently

philosophically

uninteresting), so Searle’s implicit conception of truth can

be

likewise construed.

That

is,

simply demanding

satisfaction between a primary aspect under which reference

is made and a state of affairs in the world could be neutral

concerning epistemological and metaphysical commitments, but

only if

this satisfaction is

a relation between an object

language and a metalanguage.

Since Searle (1969, p.72ff)

seems to reject such a distinction between object languages

and metalanguages,

this

move

seems unlikely.

The

alternative

interpretation of

Searle’s

view

here

is

that

satisfaction between a primary aspect under which reference

is made and a state of affairs in the world is

a relation

between, not languages, but language and the world.

This is

certainly consistent with his view of reference as a speech

act (and speech acts as a type of human action based finally

on

brain states).

If satisfaction is

a relation between

words and states of affairs, then it seems that this makes

the

same commitments

that we

saw Tarski’s

theory making,

viz.

there is

a specifiable set of identifiable individuals

(in this case states of affairs) ‘out there’ in the world.

As we saw with Tarski, this is

to presume the truth of

claims as to what individuals there are.

Two features stand

out here.

First,

as with Tarski,

Searle fails

to give a

221

  

 

theory of truth that does what we want a theory of truth to

do.

Second,

as with Tarski,

there seems to

be a further

commitment here to realism.

That is,

the presumption of a

specifiable set of identifiable individuals in order to

account for reference seems,

for Searle,

to be an implicit

commitment to a realist ontology.

At this point I want to

explain more fully what I mean by realism and then return to

my claim that Searle seems to

be committed to realism.

I

will then turn to Kripke (and Devitt) and argue likewise.

Throughout this book I have suggested that particular

views and conceptions are ‘realist’.

I have intentionally

put the word ‘realist’ within quotes because I did not

provide any sort of explanation or definition of what I took

realism to

be.

More specifically, I have argued that

particular conceptions of individuation and similarity are

‘realist’.

Most especially I have put the word ‘realist’ in

quotes in that context because,

in fact,

I think that

realism is simply a reflection of

those particular

conceptions.

That is, what I take realism to be is the view

that there is

a (specifiable) set of (identifiable)

individuals which exist independently of

our

theoretical,

cognitive and

linguistic

interests.

Furthermore,

I also

take

realism

to

be

the

view

that

similarities

among

individuals and

groups

exist

independently of

our

theoretical, cognitive and linguistic interests.

As

a

corollary to this, I take realism to hold that metaphysical,

logical and

semantic

issues

are

independent

of epistemic

 

issues, such that the question of what individuals exist is

independent of how we can know (or decide) what individuals

exist.

One of

the problems with dealing with the issue of

realism is that it means so many different things to so many

different authors.

For example,

many authors distinguish

common-sense

realism

and

scientific

realism

(e.g.

Devitt

(1984)).

Others

distinguish realism about

theories

from

realism about entities (e.g. Hacking (1983) and Cartwright

(1983)).

There is not even clear agreement as

to what

scientific realism means.

For example, Leplin (1984) notes

at

least

the

following

ten conceptions

of

scientific

realism:

1. The best current scientific theories are at

least approximately true.

2. The central terms of the best current theories

are genuinely referential.

3. The approximate truth

of

a scientific theory

is

sufficient

explanation

of

its

predictive

success.

4. The (approximate) truth of a scientific theory

is the only possible explanation of its predictive

success.

5.

A scientific theory may be approximately true

even if referentially unsuccessful.

6. The history of at least the mature sciences

shows progressive approximation to a true account

of the physical world.

7.

The theoretical claims of scientific theories

are

to

be

read literally, and

so read are

definitely true or false.

8. Scientific theories make genuine,

existential

claims.

223

 

 

9. The predictive success of a theory is evidence

for the referential success of its central terms.

10.

Science aims at a literally true account of

the physical world, and

its success is

to

be

reckoned by its progress toward achieving this

aim.

Although these conceptions of (scientific) realism are

slightly different from one another,

they all make a

connection between truth and reference.

They all, I think,

are captured by my notion of realism as being a commitment

to a particular conception of individuation and similarity,

viz.

that individuation and similarity are independent of

our interests.

In chapter five I criticized that conception

of individuation and similarity and argued that

it

is

mistaken, so I will not repeat the arguments here.

Suffice

it to say, since I identify realism with that conception, I

find realism to be wrong.

What I want to suggest at this point is that,

to the

extent that the semantic theory of truth is interesting, it

is committed to realism.

Likewise, I see Searle as being

committed to realism, as well.

Now,

it seems to me to be

obvious that the semantic theory of

truth is committed to

realism.

The definition of truth in terms of satisfaction

entails, as noted above,

that a specifiable set of

identifiable individuals exists.

Furthermore, this set

exists prior to and independent of our interests.

That is,

to

the extent that the semantic theory of truth is

philosophically interesting and does the job that we want a

theory of

truth to

do,

it

is not philosophically neutral,

but is committed, as we saw,

to a correspondence theory of

224

 

truth in which satisfaction plays the role of correspondence

between language and

the world.

In addition,

the set of

individuals which satisfy open sentences is independent of

our interests; the set is ‘out there’.14

As

we

saw,

Searle

shares

this view of

satisfaction.

There are other indications of his commitment to realism.

For example, in his discussion of the nature of Intentional

states, he claims that there is a ‘mind-to-world’ direction

of fit (i.e. our beliefs must conform to the way the world

is.

While I agree with Searle, I agree only to the extent

that

what

he

is

claiming

here

entails

a publically-

determined direction of fit.

That is, he is right, but only

if private Intentionality (and

the resulting direction of

fit) is ruled out.

However, as we saw in chapter four,

Searle’s

theory

does

not

rule

out

such

private

Intentionality.

(I will say more about this

in

the next

chapter.)

Like Searle, Kripke has said relatively little in terms

of a theory of truth and even less in connecting his theory

of reference to a

a theory of truth.

In his two books

(1980,

1982) and in his essay on speaker reference versus

semantic reference (1977) he makes no mention or proposal of

a theory of truth.

He does, however, provide an ‘outline of

a theory of

truth’

(1975).

His claims here are quite

similar to Tarski’s semantic theory, and, like Tarski,

Kripke seems motivated by seeking a resolution of semantic

paradoxes.

Kripke suggests that the notion of truth (i.e.

225

 

the assignment of truth-values to sentences) can be

explained by the concept of ‘groundedness’.

For Kripke, one

can legitimately assert that a sentence is true just in case

one can legitimately assert the sentence (and one can

legitimately assert that a sentence is not true just in case

one can legitimately deny

the sentence).

So,

if one can

legitimately assert

that

snow

is white,

then one

can

legitimately assert that ‘Snow is white’ is true.

This can

be extended

to different

‘levels’,

e.g.

one

can

then

legitimately assert that ‘"Snow is white" is true’ is true.

Groundedness is the idea that a sentence is grounded just in

case the sentence will eventually get a truth-value in this

process.

(Semantically paradoxical sentences, Kripke says,

will not get grounded.)

In Kripke’s words:

"Given a

sentence A of [language] 81, let us define A

to be grounded

if

it has a truth value in

the smallest fixed pointac;

otherwise, ungrounded." (1975, p.706).

(The smallest fixed

point is the point at which the set of true sentences is the

same as

the set of true sentences at the previous language

level.)

Kripke proceeds to give a formal account of

this

intuitive presentation of

a

theory of

truth.

For the

present purposes, what is important is

that,

like Tarski’s

theory,

Kripke’s

theory misses

the philosophically

interesting problem and his

theory really fails

to

do

the

job that we want a theory of truth to do.

As we saw,

the

theory only gets going if a speaker can legitimately assert

a sentence.

But that is exactly what is at issue!

What we

226

 

want to know, and what we want a theory of truth to do is to

let

us

know,

if

a speaker can legitimately assert a

particular sentence.

We want to know if

it

is

true that

snow is white, and it is unhelpful to be told that if we can

legitimately assert that snow is white, then ‘Snow is white’

is true.15

Another important point to note with respect to

Kripke’s outline of

a theory of truth is that it makes no

explicit connection to his theory of reference.

Devitt

(1984), however, has attempted to bolster the causal theory

of reference by fleshing out an accompanying theory of truth

and of realism.

To this attempt we will now turn.

Devitt

separates

ontic

issues

and questions

from

epistemic issues and questions from semantic issues and

questions.

One of his stated maxims is not only to separate

these different kinds of issues and questions, but also to

settle the ontic issues and questions first.

For Devitt:

Realism is an overarching ontological doctrine

about what there is and what it’s like.

It is

committed to most of the physical posits of common

sense and science, and to the view that these

entities are independent of the mental.

It has an

epistemic aspect:

the entities do not depend for

their existence or nature on our opinion; they

exist objectively. (1984, p.227).

Devitt distinguishes between ‘common-sense realism’ and

‘scientific realism’.

Common-sense realism is defined as:

tokens of most current observable common-sense, and

scientific, physical types objectively exist independent of

the

mental.

Scientific realism is defined as:

tokens

of

most current unobservable scientific physical types

objectively exist independently of

the mental.

Realism

227

 

(along with (the possibility of) human learning and

teaching) entails Correspondence Truth.

This latter concept

(i.e. Correspondence Truth (x))

is defined as follows:

sentences of

type

(x)

are true or false in virtue of

(1)

their

objective

structure;

(2)

the

objective

referential

relations between their parts and reality; (3) the objective

nature of that reality.

A crucial point to notice is that,

for Devitt, truth rests on a prior conception of (objective)

reference.

Another crucial point to notice is

that given

this definition of truth,

there is no necessary connection

between Correspondence Truth and epistemic concerns,

e.g.

evidence.

Says Devitt: "In sum,

to hold to Correspondence

Truth is

to believe in

a notion of

truth that is not

explained in terms of evidence.

Questions of evidence are

epistemic

ones

quite

distinct

from

the

constituative

question of

the nature of truth."

(1984,

p.32).

A third

crucial point to notice, related to the epistemic issue just

mentioned,

is

that the purpose of

a theory of truth is

to

explain, not reference, but communication (particularly

learning and teaching).

In

a nutshell,

then, Devitt suggests that realism

holds,

that we

‘bump

into the world’ in a causal way and

refer

to

independently existing objects

through a

causal

nexus.

This causal hook-up allows us

to posit a theory of

truth which he calls ‘Correspondence Truth’.

This view is

clearly at odds with the position which I have put forth in

this and previous chapters.

Indeed, I have spent the past

 

several chapters arguing against just this sort of view.

As

I showed in chapter five, theories of reference depend upon

prior conceptions of individuation and similarity.

The very

realist conceptions which Devitt assumes, I have argued, are

mistaken.

Furthermore, theories of reference reveal a

commitment to the ontic issue of essentialism (or anti-

essentialism)

based

on

the

underlying conceptions

of

individuation and similarity.

Finally, those conceptions of

individuation and similarity entail a commitment to

particular theories of truth, with both the cluster and

causal

theories

of

reference being committed to

a

correspondence theory of truth (because of a commitment to a

particular conception of individuation and similarity).

In

a sense, then,

I have been agreeing with Devitt that truth

is entailed by reference and reference in turn rests on

prior commitments concerning what there is.

However, I have

been arguing that realism is wrong, whereas Devitt assumes

that it is right.

In addition, by showing that realism is

mistaken,

that individuation and partitions are

interest-

dependent,

I have suggested

that ultimately ontic and

semantic issues and questions rest on epistemic issues and

questions.

Returning to Devitt’s account of

the causal theory of

reference, we can see his commitment to correspondence and

realism.

Recall that Devitt summarizes

the theory as

follows:

The central idea of

a causal theory of names

is

that our present uses of a name, say ‘Aristotle’,

229

 

 

designate the famous Greek philosopher Aristotle,

not in virtue of
 

the various things we (rightly)

believe true of him, but in virtue of a causal

network stretching back from our uses to the first

uses of the name to designate Aristotle.

It is in

this way that our present uses of a name "borrow

their reference" from earlier uses.

It

is

this

social mechanism that enables us all to designate

the same thing by a name. (1981, p.25).

In the simplest case,

the naming baptism, we simply are

perceptually (causally) connected to an object and have

‘grounding

thoughts’

(including

a

mental

representation)

about the object.

So, for example, I look at my newly-

acquired kitten

(say,

newly-born)

and,

having

the

appropriate mental representation, I say, "I call you Nana.“

Thus Nana is named.

I perceive the cat and correctly

believe it to be a cat (this latter epistemic aspect is

necessary, for Devitt; see above, chapter two, p.52 and

corresponding footnote).

By being causally connected to the

cat in this way and by saying,

"I call you Nana," I have

named (and referred to) Nana.

Now, notice the various commitments involved here.

First,

there is

a clear commitment to realism.

I can only

be causally connected to the cat if

in fact there is a cat

there to which to be causally connected.

Quite simply, the

cat has to be there.

Devitt is explicit about this when he

insists

that

I must see

the object which I

am naming

(referring to) as a member of a very general category.

Not

only is Devitt committed to saying that the cat is there,

but he is committed to the view that I see the object as,

say, a cat.

For my concerns, what this means is that there

 

 

must be

a prior partitioning established so that I am

referring to and causally connected with a cat instead of an

object of

some

other very different kind of category

(single-headed entity? cat time-slice? non-green entity?).

So,

not

only

am

I committed

to

a particular view of

individuation

(with

respect

to whether and what

this

individual is), but also I am committed to a particular view

of partitionings (with respect to what kind of

thing this

object is).

There is

a further, epistemic, commitment in

the sense

that,

according to Devitt,

I must believe that

there is an individual there and an individual which is

a

member of

a certain category.

Finally,

there

is

the

commitment that my mental representation corresponds to the

state of affairs in the world (i.e.

that it is true that I

am causally connected to this particular object which I

believe is a cat).

As can be seen, the simple act of naming

is, philosophically, not so simple, after all.

There is one final point I want to make with respect to

the

causal

theory

of

reference and

the

issues

of

correspondence and realism.

The point is a major one and a

full treatment of

it is beyond the present purposes, but I

want to raise the point, anyway.

The point is this:

to the

extent

that

the causal

theory of reference relies

on

the

concept of causality, it divorces itself from correspondence

and realism.

Let me explain briefly.

Obviously, the causal

theory of reference rests on there being causality.

We must

be caused,

for example, to have the appropriate mental

231

 

representations of the objects to which we refer.

However,

what sense can we make of causality?

It is not enough to

say that we are causally connected to an object by

perceiving it.

For one thing, this would rule out reference

to abstract and ‘future’ entities, but, as we saw in chapter

two,

the causal theory wants and needs to account for our

ability to apparently refer to those sorts of entities.

Besides this,

it won’t do

to say that perception explains

causal connections because we need causality to explain how

perception would do the job.

That is, how does perception

connect us with

the object?

We still need a causal

explanation for perception.

So, Devitt’s identification of

perceiving an object with our being causally connected to

the object won’t work.

Indeed, when we look at what sense

can be made of causality, we see that the causal theory of

reference which holds objectivity as

a desideratum is

weakened.

I want to suggest (though not fully defend the

view) that the notion of causality depends upon a prior

conception of normalcy.

That is,

to say that ‘x causes y’

is

to say,

in effect, that x was a variation from the norm

for a given situation.

For example, to say that the cause

of A’s death was A being shot makes sense because it is not

normal for A (or things like A)

to be shot.

If A were shot

during a bank robbery, it would not be appropriate to claim

that the cause of A’s death was A’s being in the bank (or

even A’s being in the bank at that time) because A’s being

in the bank (or things like A being in the bank) is not

232

 

abnormal.

Had A not been in the bank, A might not have

died, but simply A’s being in the bank is not the cause of

death.

On the other hand, A’s being shot is abnormal for A

(and things like A).

That could, then, be the cause of

death (and in many contexts would count as the cause of

death).

Nor does the emphasis on normalcy drop by moving to

a different level of explanation.

For example, it won’t do

to object that it wasn’t being shot that was

the cause of

death, but the loss of

a certain amount of blood that was

the cause of death.

Again,

this would make sense only

because it is not normal for A to lose that amount of blood.

If A was found dead having lost a teaspoon of blood, we

would not assign loss of blood as

the cause of death.

Losing a teaspoon of blood is not abnormal.

So, causality

(at

least

for

the

purposes

of

this

brief

and

simple

characterization), and the identification of causes, depends

upon there being variation from ‘normal’ conditions and

situations.

However,

normalcy - what counts as being normal -

depends upon prior individuation and partitioning.

That is,

what counts

as

normal situations

for A depends upon what

sort of thing A

is

(i.e. what partitioning is done).

If A

is

a bank robber or a bank security guard, then what is

normal for A is different than if A is someone who has never

been in a bank before.

Take an example from ‘nature’.

Suppose we ask, "What is the cause of quasars travelling at

the amazing speeds that they do?"

Well, what is normal

233

 

 

travelling speeds for quasars is different from what is

normal travelling speeds for non-quasars.

We might appeal

to particular laws

of physics and theories about the Big

Bang and suggest that the group of objects about which we

are referring is those objects created in the Big Bang.

So,

we can ask for the cause of those amazing quasar velocities

because what is normal in this situation is not what is

normal for quasars, but what is normal for things created in

the Big Bang.

But, again, what is crucial is how we break

up the world, and, as I argued in chapter five, how we break

up

the world

is

interest-dependent.

We

have certain

interests

(say,

cognitive,

theoretical

interests)

in

breaking the world into quasars versus non-quasars rather

than things created in the Big Bang versus not-(things

created in the Big Bang).

(Of course, at one time it might

be

in our interest to partition the world one way and at

another time

in our interest to partition the world in a

different way.)

The point here is

that causality depends

upon the prior notion of normalcy, which, in turn, depends

upon

the particular individuation and partitioning of

the

world.

The latter is interest-dependent.

To

the extent,

then,

that the causal theory of reference rests on

causality, it needs to either give a realist, objective

account of causality or to give up causal talk as a way of

accounting for reference (and truth).16

234

 

 

NOTES

1I

am

not

interested

here

in whether or

not

propositions (as opposed to, say, sentences) are the bearers

of truth.

I will use the term ‘proposition’, but in doing

so, I am not taking a stance on this issue.

2Graphically,

the

relation,

for

Wittgenstein

in

the

Tractatus, between language and the world is as follows:

Language

Propositions

Elementary propositions

Namese\‘§‘-‘\‘\“‘\

Picturing relation

Naming relation

States of Affairs

Facts

World

3However, in the shorter, informal presentation of his

semantic

theory

(1944),

he

insists

that

the

semantic

conception of truth is "completely neutral" with respect to

epistemic and metaphysical attitudes.

As

I argue below, I

believe he

is mistaken about this;

his theory is committed

to realism.

4The following synopsis of Tarski draws heavily from

Haack’s (1978) discussion of him.

5While Tarski’s (1944) definition of Convention T

is:

"(T) X

is true if, and only if, p," his original (1956)

definition was

as

follows:

"Convention T.

A formally

correct definition of

the

symbol

‘Tr’,

formulated

in

the

metalanguage, will be called an adequate definition of truth

if

it has the following consequences: (a)all sentence; which

 

are obtained from the expression ‘x 6 Tr if and only if p’

by substituting for the symbol ‘x’ a structural-descriptive

name of any sentence of the language in question and for the

symbol

‘p’

the expression which forms

the

translation of

this sentence into

the metalanguage; (b)the sentence,

‘for

any x,

if

x

6 Tr then x

e

S’

(in other words ‘Tr s S’)."

(1956, pp.187-188).

[‘8’ here refers to an arbitrary class

sequence.

6As Haack put it,

in a manner of speaking, Convention T

235

 

.ixes the extension of ‘true’.

7The reason that Tarski defines satisfaction first is

because closed, compound sentences are formed from open

sentences rather than

from closed, atomic sentences.

For

example, ‘(3x)(Fx v Gx)’ is formed from ‘Fx’ and ‘Gx’ by the

operations

of

disjunction and existential quantification.

‘Fx’ and ‘Gx’ are not true or false; they are satisfied by

objects.

The definition of satisfaction is recursive (i.e.

definitions are given first for the simplest open sentences

and then the conditions are stated in which compound open

sentences are satisfied).

8Actually, if A is a wff, then (3¢)A is a wff, where ¢

stands for any variable.

9Tarski’s

stated

definition

of

‘satisfies’

is:

"The

sequence f satisfies the sentential function x

if and only

if

f

is an infinite sequence of classes and x

is

a

sentential function and these satisfy one of

the following

four conditions: (a)there exist natural numbers k and 1 such

that x=‘k,l and fk 5 f1; (b)there is a sentential function y

such

that

x=y

and

f

does not satisfy the function y;

(c)there are sentential functions y and 2 such that x=y+z

and f either satisfies y

or satisfies z; (d)there is

a

natural number k and a sentential function y such that x=nky

and every infinite sequence of classes which differs from f

in at most the k-th place satisfies the function y." (1956,

p.193).

10Tarski’s

(1956)

stated definition is:

"x

is

a

true 

sentence -

in symbols x

e

Tr -

if and only if

x e

S and

every infinite sequence of classes satisfies x." (1956,

p.195).

11Another,

perhaps

simpler,

way

to

present

Tarski’s

procedure is

as follows:

Let

I

be

a function from

expressions (in O)

to objects (in the world, or in a model),

such that to each individual constant I assigns an object;

to each variable I assigns an object;

to each n-place

predicate I assigns a set of ordered n-tuples of objects.

Then,

‘Fa’

is

true under I

iff

I(‘a’) E

I(‘F’);

‘Fab’

is

true under I iff <I(‘a’), I(‘b’)>

é

I(‘F’); ‘“Fa’ is true

under I

iff I(‘a’) i

I(‘F’); ‘A & B’

is true under I iff A

is true under I and B

is true under I; ‘Fx’ is true under I

iff I(‘x’) e I(‘F'); ‘(3x)Fx’ is true under I iff there is

an interpretation 1’ which differs at most in what it

assigns to ‘x’ which makes ‘Fx’ true under I.

12To suggest at

this point that what the semantic

theory of

truth (or any theory of truth,

for that matter)

aims at giving is not a criterion (or criteria) for telling

us what is true, but rather a definition of truth will not,

I think, work.

I still want to ask why we would want a

236

 

 

definition of

truth.

It

is not simply to give us

a

definition, but because we want to

be able to distinguish

between true claims and false claims.

The purpose for this

(or, at least, a purpose) is

to give us a criterion for

accepting inferences based on

true (but not false) claims

and for acknowledging, say, warrantability of actions and

practices based on particular (true) claims.

Arguing that

the semantic theory gives us

a definition of truth, not a

criterion of demarcating true claims from false ones, does

not, I think, meet the objection I am raising in the text.

l3Or,

consider

the

case

of

humoring

someone

and

speaking about that person as

the present king of France.

So a speaker says, ‘I see the present king of France is

hungry.’

The speaker and hearer (assuming both know that

the speaker is humoring the person being looked at) might

agree that the speaker has referred to, and made a true

statement about, that particular person being looked at.

14One might want to argue that the semantic theory of

truth does

not entail that

the set of individuals exists

independent of

our interests.

One might say

that we can

agree that we determine what individuals there are and in

what ways

they

are

(dis)similar and furthermore hold

the

view that - given those individuals - truth is definable in

terms of their satisfying open sentences.

So, satisfaction

is just as legitimate for gruesome speakers as for English

speakers.

I am somewhat sympathetic to this claim, but I am

hesitant to endorse it

for

two reasons.

First (and less

importantly), I don’t think that this is what Tarski had in

mind in proposing his theory.

Second (and more important),

this begins to rob the notion of correspondence of any

significance and uniqueness.

The semantic theory fails to

serve as a vehicle for determining what is true if it allows

correspondence

between

language

and

any

individuals

(and

partitionings).

Besides,

we would still be

left with no

clearer meaning of what correspondence is than our original

intuitions.

Yet the semantic theory was meant, in part, to

clarify

those

intuitions,

to

give

us

a

more

precise

conception of what correspondence is.

Nonetheless, if one

argues that the semantic theory of truth and its corollary

conception of correspondence is consistent with an interest-

determined

principle

of

individuation

and

similarity

relation,

then I find it unobjectionable, because it seems

to

me

to

then

be indistinguishable from a coherence,

verificationist point of view.

15Kripke may be admitting this when he confesses that,

for ordinary statements, "there is no intrinsic guarantee of

their safety (groundedness) and that their ‘level’ depends

on empirical facts" (1975, p.707).

16Putnam (1981) also suggests that the causal theory of

reference fails to

the extent that it relies on causality,

237

 

 E::::__________________________________________________________________________________:=III

or at least on appropriate causal chains.

He states:

...if

I say

‘the word "horse" refers to objects

which have a property which is connected with my

production of the utterance "There is

a horse in

front of me" on certain occasions by a causal

chain of the appropriate type’,

then I have the

problem that, if I am able to specify what is the

appropriate type of causal chain, I must already

be able

to refer to

the kinds of

things and

properties that make up that kind of causal chain.

But how did I get to be able to do this?

The conclusion is not that there are no terms

which have

the logic ascribed by

the similitude

[i.e.

cluster]

theory,

any

more

than

the

conclusion is that there are no terms which refer

to things which are connected to us by particular

kinds of causal chains.

The conclusion is simply

that neither similitude nor causal connection can

be

the only, or

the fundamental, mechanism of

reference. (1981, p.66).

Precisely!

238

 

A CONSTRUCTIVIST ALTERNATIVE

"I have never doubted the truth of signs, Adso; they are the

only things man has with which to orient himself in the

world.“

 

INTRODUCTION

In this final section I want to suggest an alternative

view of

reference,

a view based not on realism and a

correspondence theory of truth, but on verificationism and a

coherence theory of

truth.

Like Kripke, I claim to offer

only a ‘better picture’ of

a theory of reference.

Like

Kripke,

I also claim not to offer a set of necessary and

sufficient conditions for

(a theory of) reference.

In

chapter eight I will present (admittedly minimally) a case

for coherence and verificationism.1

In chapter nine I will

present a glimpse of what a view of reference would be based

on coherence and verificationist commitments.

The case I

present in these final two chapters is admittedly minimal

partly because I have already been indicating my reasons for

such views in

the discussion up

till now of my misgivings

with the received theories of reference and their underlying

commitments.

Besides this reason for giving here a minimal

case

for coherence and verificationism

(and

a subsequent

view of reference),

I also see

the project of fully (and

adequately) giving my case as beyond the scope of this

present book.

What I have been hoping to

do here

is

primarily to show the received theories of reference as both

being committed

to privacy,

essentialism,

realism and

correspondence.

As such I see them as two sides of the same

240

 

theory.

I have tried to identify their shared underlying

commitments and,

at

the same

time, criticize those

commitments.

Having done that,

I now want to consider an

alternative view of reference, based on alternative

commitments.

The

form that

this consideration will

take

here is simply that of initial feelers which suggest a

plausible alternative.

These final two chapters, then, are

little more than a statement of a scope of work to be done,

not a final theory of reference based on coherence and

verificationism.

As mentioned in the preface,

this work is intended to

do at least two (related) things.

One is

to show that

theories of reference are not simply semantic matters, but

carry with them commitments to other issues and positions.

To fully understand and appreciate theories of reference, we

need to make as explicit as possible these underlying

commitments.

The second intention I have in mind in

this

work is to show that commitments to, say, theories of truth,

bear directly on other matters (for example, on theories of

reference).

These two intentions are, perhaps, merely two

angles on the single point to which I am committed, viz. our

philosophical positions are understandable only as coherent

wholes, with no sharp separation between issues of ontology,

epistemology, semantics, etc.

Now,

before

suggesting an alternative view of

reference, I want to review my account of

the received

theories of reference.

I will do this by indicating what I

 

 

see

as

the

strengths

and weaknesses

of

the

received

theories.

First, what I see as

the major strength of

Searle’s cluster theory is

that he insists that reference

must be understood as

a speech act,

that it

is intimately

tied up with our extra-semantic, pragmatic concerns.

I

agree

completely and,

in

fact,

see

one

of

the major

weaknesses of Searle’s position as his abandoning of

this

insistence when he allows private

reference.

The

formal

presentation of Searle’s theory has,

I think, been soundly

criticized by Kripke (though, as

I argued in the first

chapter,

I

think

that Kripke attacked something of

a

simplified version of Searle).

What I see as the strength

of Kripke’s causal theory is his insistence that reference

is public, that we do not refer simply because we intend to.

Of course, as I argued in chapter four, I think that while

Kripke talks as though publicity is necessary for successful

reference to occur,

his

own theory (especially in

its

fullest - Devittian - form) allows private reference.

The

(somewhat) formal presentation of Kripke’s theory faces

difficulties with ordinary cases of reference to,

say,

abstract and ‘future’ entities

(cases which are

of

no

problem to Searle).

Furthermore, without some satisfactory

account of causality, I don’t see

the causal theory as

offering a very plausible alternative to Searle or the

descriptivist accounts of refcrence.

As I suggested in

chapter seven,

appealing to causality as

a basis

for

(or

even as an element of) a theory of reference will not cohere

242

well with the sort of ‘objective, realist’ view which Kripke

(and Devitt) want to hold.

Opting for talk of an historical

theory of reference (a

la Donnellan) rather than a causal

theory, fares no better unless one can make the assumption

of an ‘objective, realist’ account of history.

However, as

I argued in chapters five and six,

such an assumption is

suspect (at least).

In sum, then, I think that at best the

received theories each offer a necessary condition of

successful reference.

A Searlean cluster theory, I want to

say,

is fine,

but only to

the extent to which it requires

public reference.

So, when Searle claims that at least one

description associated with a name (and used in the primary

aspect) must be true of an object for the speaker to refer

to the object by the use of the name, then I agree, but only

to

the extent that it holds, not for the speaker, but for

the language community.

I can refer to Feynman by (the use

of) ‘Feynman’ even if I have no true descriptions associated

with ‘Feynman’ to enable me to pick out Feynman.

However,

this is only because names are public and I am a member of a

language community.

If

the descriptions associated with a

name failed for the entire language community, then I would

argue that reference fails.

(I am not here suggesting that

Searle

is

right

to

claim

that

there

is

a cluster of

descriptions which each speaker associates with a given

name.

As

I will argue in chapter nine,

this view of

the

speaker having a

‘ready made’ cluster of descriptions

associated with a name distinguishes his account of

243

 

reference from Wittgenstein’s account, and weakens his

(Searle’s) position.)

So, while Searle’s account may, at best, provide a

necessary condition for successful reference,

the same

is

true for Kripke’s causal account.

Objects are historical

entities and reference is

an historical process; both are

public.

For reasons already mentioned, I think that in the

end,

the causal theory fails to give a plausible theory of

reference, however,

the intuitions which Kripke pointed to

in explicating his causal view are, I think, correct in

large part.

Kripke is

right,

I

think,

to recognize that

reference must be understood diachronically.

Where he goes

wrong

is

in

thinking

that

this diachronic feature of

reference is also a realist feature.

In the following two

chapters, I will draw on what I see as the positive aspects

of

the received theories.

As

they have been presented by

their adherents, they are fundamentally mistaken.

However,

they are not completely mistaken.

To the extent that they

are not mistaken, they can (and will) be incorporated into

the alternative view which I will now offer.

244

 

NOTE

1I want to stress at this point that the sense of

coherence I

will employ

in

the following chapters

is

coherence as consistency and completeness.

However, by this

I do not mean simple logical consistency and completeness.

(Likewise, when I speak of verification(ism), I do not mean

logical verification(ism).)

Such a purely logical

conception of coherence and verification(ism), divorced from

human practices, behaviors and institutions, is,

I will

argue,

a

mistaken

way

of

seeing

coherence

and

verification(ism).

245

 

 

CHAPTER EIGHT

COHERENCE AND VERIFICATIONISM

In this chapter and the next, I want to suggest a view

of

reference

that

is

an alternative

to

the

received

theories.

The alternative takes

the form primarily of an

alternative with respect to underlying commitments.

On the

surface, this view of reference looks much like the theories

proposed by Searle and Kripke.

It is the nature and meaning

of reference, not its theoretical structure, which is most

important in the present context.

Indeed,

the theoretical

structure of

reference can only be understood,

I want to

claim,

within the context and confines of

its nature and

meaning.

Ultimately,

reference

is

(and can only be)

understood within the behavior and practices (i.e.

the

Wittgensteinian ‘forms of life’) that encompass it and give

it not only its meaning but also its possibility.

As I have

argued up to now, theories of reference carry with them, and

are based on, underlying (epistemic) commitments.

I have

argued

that

the

received

theories

of

reference

share

particular epistemic commitments and rest in large part on

particular conceptions of individuation and similarity.

Those conceptions I have identified as being correspondence

and realism (and, subsequently, essentialism and privacy),

and I have indicated that I find those conceptions as being

246

l
r
.
|
.
|
1

 .

fraught with difficulties.

In this chapter I want to

consider

alternative

conceptions,

viz.

coherence

and

verificationism.

I will present what I consider to be

plausible views of these two positions and their conceptions

of individuation and similarity.

I will then consider

objections to those views.

In the next chapter I will

propose (albeit tentatively) a view of reference consistent

with these views and suggest how I see such a view of

reference as being superior to the received theories of

reference.

First, then, I want to present what I take a coherence

theory of truth to be.

In simplest terms, a coherence

theory of truth is the view that the truth of a statement is

a function of its coherence (i.e. consistency) with the set

of statements accepted as true.

Several features stand out

from this simple statement.

First, the coherence theory of

truth emphasizes the issue of justification.

That is, what

justifies the assertion that a claim is

true

is

that

the

claim coheres with the set of accepted statements.

Second,

the coherence theory of truth is

(or, at least, appears to

be) strictly epistemic.

Note that there is no mention in

the ‘definition’ of coherence of reality or the world.

(As

will be addressed below, this feature has been criticized by

opponents of the coherence view.)

The

coherence

theory

of

truth historically has

been

closely associated with philosophical idealism.

However,

coherence also historically has been associated with logical

positivism.

This might indicate that the theory is truly

epistemic and makes no ontic commitments to,

say, idealism

or realism.

(Both White (1970) and Haack (1978),

for

example, discuss coherence as separable from ontic

commitments.

I disagree, and will return to this issue

below.)

Contemporary versions of a coherence theory include

those of Blanshard (1939), Quine (1972) and Rescher (1973).

While coherence is usually spoken of

in terms of

consistency,

coherence

theorists

stress

that

consistency

alone is not the mark of truth or a good view of truth.

For

example,

a

theory which contained a single atomic

proposition would,

on

just

the

basis

of consistency,

constitute a consistent set of (accepted) statements.

Yet,

intuitively, this is not a plausible view of truth.

What is

also

required

(minimally)

is

completeness

or

comprehensiveness.

What is desired is not simply a set of

accepted statements, but the most complete set of consistent

(accepted) statements.

So, for example, in Rescher’s (1973)

presentation of a coherence theory of truth he refers to a

maximal consistent subset (MCS) of beliefs (of a set of

beliefs), which he defines as follows: 8’

is

a maximal

consistent subset of S

if

it

is a non-empty subset of S

which is consistent; and to which no member of 8 not already

a member of

S’

can

be added without generating an

inconsistency.

(As

will

be

noted

below,

this

characterization of coherence, and ones like it, have been

criticized as,

in Russell’s words,

failing to distinguish

248

 

truth from consistent fairy tales.

I will discuss this

criticism below.)

Minimally,

then,

a coherence theory of

truth demands not only consistency but also completeness.

Whether or not there is a necessary connection between

coherence and ontic commitments, there has definitely been a

connection between coherence and other epistemic concerns

such as verifiability.

For example, often (if not always),

a test of someone’s statement as being true or false depends

upon the statement’s consistency (or lack of

it) with

accepted statements.

For instance,

if

a speaker insists

that she saw a ghost (or, for that matter, a rose—breasted

grosbeak), a minimal test for the acceptance of that claim

as

true depends upon its consistency with accepted

statements.

We would probably be sceptical about the truth

of seeing a ghost given the accepted statements we hold

today.

While we would probably tend to accept a claim about

the sighting of a rose—breasted grosbeak, we wouldn’t if the

claim were made in particular contexts (for example, in my

basement or in New York City during February).

Again,

consistency with accepted statements seems crucial for the

acceptance of a given statement as true.

This is especially

true for, say, mathematical claims.

To say that Goldbach’s

Conjecture is

true seems

to rest fundamentally on

its

consistency with accepted mathematical statements.

(Indeed,

as Russell’s paradox of set membership indicates, internal

inconsistency is sufficient for rejecting not only a given

mathematical statement, but an entire system.)

Another

249

 

   

intuitive sort of statement for which consistency is primary

is that of historical claims.

For example, the claim (see

Gould (1983)) that Teilhard de Chardin was responsible for

the Piltdown hoax is determinable only on the criterion of

consistency.

We accept the claim as true only if there is

sufficient confirming or corroborating evidence and reject

it as false if there is disconfirming evidence (i.e. if the

claim is inconsistent with accepted statements).

In none of

these cases is there any mention of correspondence to facts.

Besides this, however,

there is

an apparent connection

between coherence and verification (or,

at

least,

testability).

(An objection to the presentation just given

is that coherence is not necessarily connected with

verification because a claim can cohere with (accepted)

statements whether or not we could verify this coherence.

I

will consider this objection below.)

Before considering the various criticisms of coherence

which I have mentioned above,

I want to address an issue

associated with coherence which I do not consider to be

particularly significant.

The issue is that of degrees of

truth.

If

coherence

is

simply a matter of

internal

consistency along with completeness,

then,

it

is

said,

to

the extent that a given statement is consistent with more or

fewer other accepted statements that given statement is more

or less true.

So, a claim (S) that is consistent with more

accepted statements than is another statement (8’)

is more

true than that other statement.

(The calculation of ‘amount

of consistency’ could be the following: the degree of truth

of statement S

is

a function of the number of accepted

statements with which S

is consistent minus the number of

accepted statements with which S is inconsistent.

Call this

value ‘the relative truth’ of the statement (Tr(S)).

80, S

is more true than S’

if and only if

Tt(S) > Tr(S’).)

(Of

course, in one sense this definition is inadequate - and the

objection is incoherent - because given any statement S, its

‘relative truth’ with respect to other statements depends

upon the other statements.

If a statement S is inconsistent

with other statements in a set, then it is inconsistent with

statements in any larger set (which includes ~S).)

I don’t

think that the issue of degrees of truth is significant for

the present concerns.

In part, I say this because this way

of looking at coherence seems to presume that consistency is

a matter of degree and that there is already a specifiable

set

of accepted statements with which to compare a given

statement.

As will be seen, in the conception of coherence

which I propose below, the notion of consistency as a matter

of degree doesn’t make sense, nor does the notion that there

is

an already-established specifiable set

of accepted

statements.

(Indeed, as

I will argue in the next chapter,

this is analogous to a difference between Searle’s theory of

reference and Wittgenstein’s view of reference; for Searle,

it

makes

sense

to

speak of

an already-established

disjunctive set of descriptions associated with a name,

while, for Wittgenstein, this makes no sense.)

I will

mention this issue of degrees of truth again later, but for

now I will turn to criticisms of coherence and try to

respond to those criticisms.

Criticisms of Coherence

I will consider four objections to coherence.

First,

it is claimed that advocates of the coherence theory of

truth confuse a definition of ‘truth’ with a criterion for

truth.

Second, it is claimed that coherence as consistency

and completeness fails

to make contact with the world;

it

fails to distinguish truth from consistent fantasy.

Third,

it is claimed that coherence fails to offer any connection

with verification.

Fourth, it

is claimed that ‘truth’ and

‘true’ apply not to particular statements, but only to

systems

or

sets

of

(consistent and

complete)

statements;

this

is minimally unintuitive.

I will flesh out and

consider each of these objections.

Objection One: The coherence theory of truth confuses a

definition of ‘truth’ with a criterion for (the acceptance

of)

truth.

It may well be that we demand coherence among

our statements as

a criterion for testing statements,

but

this does not serve as a definition of ‘truth’.

That is, it

may well be that when determining whether or not it is true

that Teilhard de Chardin was responsible for

the Piltdown

hoax we demand coherence between this claim and other

accepted statements, but this only entails that we want

coherence,

not

that truth is identifiable with coherence.

Furthermore, coherence is a desideratum because the world is

 

coherent.

Coherence does not entail truth, it reflects

reality (and truth).

Objection Two: The coherence theory of truth fails to

make any hook-up with the world.

Taken literally, the

coherence theory of truth fails to distinguish between truth

and coherent fantasy.

Two complete and internally

consistent sets

of

statements

could

be

inconsistent with

each other, yet, under the coherent theory of

truth,

they

would both be

true.

For an external point of view,

then,

the

coherence

theory

of

truth

would

countenance

inconsistency (and hence be incoherent).

Objection Three: The coherence theory of truth, while

apparently relying on verification for determining the truth

of statements, fails to make any necessary connection with

verification.

The theory simply demands consistency (and

completeness), but

this

says nothing about whether we can

verify any coherence among statements.

Objection Four:

The coherence theory of

truth allows

truth to apply only to systems or sets of statements, not to

particular statements.

It

is not the case,

then,

that

‘Teilhard de Chardin is responsible for the Piltdown hoax’

is a true (or false) statement.

What is true (or false) is

the set of accepted statements of which one member is

the

statement ‘Teilhard de Chardin is responsible for

the

Piltdown hoax.’

Minimally,

this

is unintuitive.

Surely,

truth applies to particular statements.

I now want to respond to these objections.

(While I

253

 

will specifically address a particular objection at any

given time, my response to each overlaps with my response to

the

others.

In

effect,

I

see

the

objections

as

interrelated;

consequently,

my

responses

will

be

interrelated and will be presented as responses to a single,

disjunctive objection.)

The first objection is

that

the coherence

theory of

truth confuses a definition of truth with a criterion for

(the acceptance of) truth.

(My response to this objection

will usher in the verificationist elements of my view.)

It

has

been argued that coherence serves as

a criterion for

(the acceptance of)

truth or truth claims,

but

that

this

does not constitute a definition of ‘truth’.

The conflation

here,

it

is claimed,

is reflective of a conflation of

two

related but separate questions: (1) what is coherence (i.e.

what does

‘coherence’ mean)?,

and

(2)

what coheres

(i.e.

what (kinds of) things exhibit coherence)?

An answer to the

first question is

a definition of ‘truth’, while an answer

to the second question is a criterion of truth.

One might

accept the view that coherence is important (essential,

necessary,

etc.)

for serving as

a criterion for

(the

acceptance of) truth.

Indeed, we do demand coherence among

our claims.

However,

this hardly shows

that what ‘truth’

means is coherence.

The definitional question can be

answered by,

say, ‘correspondence to the facts’, while the

criteriological question can

be answered by

an appeal

to

coherence among our statements.

In fact, under this view,

254

since ‘truth’ means ‘correspondence to the facts’, we would

insist on the criterion of coherence, since we would want to

assert (assume?) that facts cohere with one another.

(Here

I

am not only arguing that the sharp distinction between

coherence as criterion and as definition cannot be drawn,

but also

that coherence can

(and does) serve as

a

definition, as well as a criterion, of truth.)

Now, what could it mean to say that there is

a

distinction between saying that something is true and saying

that we are justified in classifying something as true?

What can it mean to say that it really is true that Teilhard

de Chardin was responsible for the Piltdown hoax without at

the same time saying that we are justified in classifying

the statement ‘Teilhard de Chardin was responsible for the

Piltdown hoax’ as

true?

Could we ever meaningfully assert

that a statement was true (really, actually true!) without

being committed to claims concerning the justifiability of

such an assertion?

At least in this case,

the difference

between definition and criteria does not seem to me to make

a difference.

What, after all, would a (proper) criterion

(or definition)

of

truth do

or seek other than,

say,

necessary and sufficient conditions for a statement’s being

warranted and justifiably classified as

true?

A (proper)

criterion, we would suppose, establishes the conditions for

justifiably accepting a statement as

true.

However, if

it

is not sufficient for establishing the statement as

true,

why is

is sufficient for establishing the statement as

 

justifiably acceptable (as true)?

What I have done so far is

to suggest that the

difference between coherence as a definition of ‘truth’ and

as a criterion for (the acceptance of) truth is not obvious.

Minimally,

the establishment of

a difference needs greater

motivation.

What provides this greater motivation comes, I

think,

from considerations which prompt the second

objection, viz. there must be some hook-up between coherent

systems

and

the

world.

The

criterion for warranted

assertability might well be coherence with accepted

statements (belief, theory, etc.), but the hook-up with the

world comes with correspondence with

the facts

(or,

say,

with Tarskian satisfaction).

This brings us

to the second

(and, perhaps, most important) objection: the coherence

theory fails

to make any hook-up with the world.

As

mentioned above, Russell claimed that the coherence theory

of truth fails to distinguish between truth and consistent

fantasy.

Along

these

lines,

Russell

(1912)

suggests

the

following difficulty with coherence:1

...there is no reason to suppose that only one

 

coherent body of beliefs is possible.

It may be

that,

with sufficient imagination, a novelist

might invent a past for

the world that would

perfectly fit on to what we know, and yet be quite

different from the real past.

In more scientific

matters, it is certain that there are often two of

more hypotheses which account for all

the known

facts

on some subject,

and although,

in such

cases, men of science endeavour to find facts

which will rule out all the hypotheses except one,

there is no reason why they should always succeed.

(1912, p.122).

In part, because the coherence theory fails to guarantee any

256

 

hook-up with the world, any number of coherent systems could

be true.

The coherence theory offers no criterion for

distinguishing between coherent systems.

Furthermore,

although Russell does not bring in this point explicitly, it

will

not work

to

say

that

the coherence

theory does

distinguish

between

coherent

systems

by

insisting on

completeness as well as consistency.

The reason that this

defense won’t work is because there could be numerous

(perhaps infinite) consistent and complete systems.

Coherence,

then,

even conceived of

as consistency coupled

with completeness,

fails

to offer a criterion for

distinguishing between different coherent systems.

Quite

simply, even if coherence is

a necessary condition for

truth, it is not a sufficient condition.

Now, it seems to me that this objection (i.e. coherence

fails to establish a connection with the world) is

fundamental as

a challenge to coherence.

I want to offer

two responses to

this challenge.

First,

I want to argue

that the coherence theory of truth does connect with ‘the

world’.

Second, the objection itself presupposes realism by

presupposing that there is

a world ‘out there’ with which

coherence

must

connect.

I

want

to question

this

presupposition on two grounds; first, that there is a (i.e.

one) world

out there’, and, second, that there is

a world

\

out there (i.e. independent of our systems).

First,

then,

I want to argue that coherence is not

simple logical consistency (and completeness).

Rather,

257

 

coherence

rests

on

our

behaviors,

practices

and

institutions,

what Wittgenstein has

called our

‘forms

of

life’.

Coherence, I want to claim,

is

a semantic, linguistic

concept.

As such,

it is

a public concept.

That is,

as I

argued in chapter four, language is necessarily public;

consequently,

notions of

coherence,

being linguistic,

are

also necessarily public.

The arguments which point to the

publicity of language and coherence also point to the extra-

linguistic nature and basis of language and coherence.

A

brief

summary

of

Wittgenstein’s

arguments

concerning

language, publicity and forms of life will help here.

Wittgenstein (1953) says:

"So

you

are

saying

that human agreement

decides what is true and what is false?" - It is

what human beings say that is true and false; and

they agree in the language they use.

This is not

agreement in opinion but in form of life.

(1953,

#241).

The conceptual context for this claim is the following.

Wittgenstein saw the meaning of a word (at least for "a

large class of cases") as its use in a language.

Words and

sentences are ‘tools’ which are used for specific purposes

within a language.

The employment of words and sentences is

understandable only within ‘forms of life’.

The notion of

meaning as use within (different) contexts led Wittgenstein

to his famous conception of ‘language games’, a notion which

is meant to capture

the fact

that language is both rule-

governed

(like

games)

and

a human artifact

(like

games).

This is not to say that words can mean anything we want them

258

to.

As children, when learning a language, we learn to use

it correctly by

the rules which govern the language.

We

learn these

rules

to an extent explicitly (e.g.

we study

grammar in elementary school), but for the most part we

learn these rules implicitly (e.g.

by the way people speak

to us).

These rules are learned by how we and others use

expressions, and obeying a rule (which is necessary for

language) is a practice.

The simple fact that we learn the

meaning (and form) of expressions by how they are used

prevents us from making up whatever rules we wish, because

we would not be able

to communicate with others by using

private rules

(besides,

conceptually,

as noted in chapter

four, private rules make no sense).

We follow (perceived)

regularity of behavior.

This (perceived) regularity of behavior points to two

important features, the fundamental nature of forms of life

(and

language

games)

and

the

importance

of verification.

For Wittgenstein, what has

to

be accepted, what is given,

are

forms

of

life

(i.e.

our behaviors,

practices,

institutions, etc.).

There are social contexts within which

our concepts and language games operate.

(E.g.:

"‘I know

all that.’

And that will come out in the way I act and in

the way

I speak about the things in question." (1969,

#395).)2

Consider the following analysis of

doubting as

illustrating the fundamental nature and importance of forms

of life.

According to Wittgenstein,

the language game of

doubting cannot be universal in scope.

That is,

we can’t

259

 

doubt everything.

If we tried doubting everything, we would

not get as far as doubting anything.

As

he puts it,

the

game of doubting presupposes certainty and this comes from

the very nature of language:

That is

to say,

the questions that we raise

and

our doubts depend on

the fact that some

propositions are exempt from doubt, are as it were

like hinges on which those turn.

That is

to say,

it belongs to

the logic of

our scientific investigations that certain things

are in deed not doubted.

But it isn’t that the situation is like this:

we just can’t investigate everything, and for that

reason we are forced to rest content with

assumption.

If

I want the door to

turn,

the

hinges must stay put.

My life consists in

my being content to

accept many things. (1969, ##341-344).

For language to be possible rules must be obeyed.

This

obedience is

a practice;

as

such,

our behavior is

consistent.

Of course, Wittgenstein is not simply appealing

to common sense here as support for or proof of the notion

of forms of life.

As he says:

Giving

grounds,

however,

justifying

the

evidence, comes to an end;

- but the end is not

certain

propositions

striking us

immediately

as

true;

i.e.,

it

is not a kind of seeing on our

part;

it

is our acting, which lies at the bottom

of our language-game. (1969, #204).

This

leads,

then,

to

a

conception

of

truth.

Wittgenstein remarks that what a proposition is

is

in one

sense determined by the rules of sentence formation and in

another sense by the use of

the sign in the language game.

The

words

‘true’

and

‘false’

function

just

as

other

constituents of the game and are not separate from it.3

Wittgenstein uses an example of

a chess game to illustrate

 

this.

If one says ‘The king in chess is the piece that one

can check’, one has said no more than that in chess we only

check the king.

To speak of the truth of the statement

outside of

its context of chess is nonsense.

(Also,

‘One

can steal second base but not first base’ is true, but not

outside the context of baseball.)

When Moore tried to

determine the truth or falsehood of

‘I know here is my

hand’, for example, as though the truth value were part of

the context preceding language,

he was mistaken.

He was

trying to treat truth as something outside of language (and

perhaps language as something outside of our practices, or

forms of life).

This leads back to the quotation above initiating this

discussion of Wittgenstein.

Does

our

language determine

what truth is (and what is true)?

Are we led to a

subjectivist, relativist view of truth?

As noted above, his

response is that what is true, for him, is what is rooted in

our

language,

which

is

rooted

in

our

forms

of

life.

Communal agreement and use (based on communal practices) is

fundamental to

the notion of truth.

Now,

it

is

a simple

move from use as basic to utility as basic for the

conception of truth, and indeed Wittgenstein remarks that he

is

trying

to

say something that "sounds like pragmatism."

(1969, #422).

For Peirce (1955),

talk of truth is

meaningful

only within

the actual and practical bearings

which frame human activity.

(For Peirce, however, the

practical

is

the

sort

of

inquiry which characterizes

261

 

 

scientific activity.)

James (1977) saw the practical as the

importance

of

beliefs

to

the

values

and

behavior of

individuals.

(For James,

to think of some sort of

independent truth is beside the point.

If beliefs are

verified, if they satisfy what is expected or intended, they

are true.

If something works, it is right; if it satisfies,

it is true.)

Dewey (1950) stressed more the use of beliefs

than their utility, within the context of our actions and

inquiries.

(The traditional correspondence view of the

relation of fact and belief is not enough, for Dewey, since

we must know what purpose, what goal is being aimed at in

our inquiry or activity.

We cannot separate

the inquiry

from its context.)

The similarity between these views and

Wittgenstein’s are, I think, obvious.

Now,

this demand for use within context seems to lead

toward the emphasis of

a coherence view, and this does

indeed seem to be Wittgenstein’s leaning.

He states:

When we first begin to believe anything, what

we believe is not a single proposition,

it

is

a

whole

system

of

propositions.

(Light

dawns

gradually over the whole.)

It is not single axioms that strike me as

obvious, it

is

a system in which consequences and

premises give one another mutual support.

(1969,

##141-142).

Quite simply, for Wittgenstein, we do not have experiences,

or make statements, or hold beliefs in isolation.

They all

occur within our language games and are based on our forms

of life.

The significance of this for the present concerns

is that it is not the case that a coherence theory of truth

fails to make any hook-up with ‘the world’ (or, as I prefer

262

 

to say, it is not the case that a coherence theory of truth

fails to incorporate a non-linguistic component).

The hook-

up

(or

component),

however,

is

not

some

sort

of

correspondence with facts; rather, it is an acknowledgement

that truth (as coherence) emerges from communal language,

which in turn emerges from communal behaviors, practices and

institutions.

Coherence as consistency and completeness is

not,

then, simply logical consistency and completeness, but

communal,

practical,

consistency and completeness.

(Note

that

this also illustrates the verificationist element of

coherence.

Because truth (as coherence) is ultimately based

on language and practice, truth claims must, in effect, be

verifiable through that language and practice.

Unverifiable

truth claims make no sense in a context in which truth is

based on language and practice.)

In considering the objection that coherence fails

to

make any connection with the world (and allows an infinite

number of possible coherent systems), I remarked that I had

two responses to this objection.

The first (just discussed)

was

that

I

think

one

can argue

for

a non-linguistic

component of

coherence

(viz.

forms

of

life).

The second

response, I said, was that the objection rested on realist

assumptions.

As I indicated in the previous chapter, I find

realist assumptions questionable.

I now want to flesh out

my second response to the objection.

Russell thought there was a problem with coherence

because it allowed an infinite number of possible coherent

(but different) systems.

Glibly, my response is:

so what?

Following Goodman (1972, 1978), I think it is a mistake to

speak of the gay the world is.

Instead, we should speak of

the ways the world is.

Since our most fundamental ontic

concerns, viz. individuation and similarity, are, as

I

argued in chapter five,

theory-,

language-,

and interest-

dependent, and since our communal practices ultimately

determine notions of truth, then to suggest that there is a

way the world is,

is,

I think, mistaken.

English speakers

might well decide

that

the world is

‘really’ partitioned

into green things and not-green things, while gruesome

speakers might well decide that

the world is

‘really’

partitioned into grue things and not-grue things.

But is

grass ‘really’ green or grue?

Well, it’s neither and it’s

both.

It’s neither if

that means that the statement that

grass is green is true independent of a language community;

it’s both if that means that different language communities

view it as one color or the other and that view is coherent

with

their

other beliefs

(both synchronically and

diachronically).

The recognition that there are different

ways of individuating objects and different ways of

partitioning the world removes the motivation for insisting

that there is a way the world is.

5

Furthermore, it seems to

me

that nothing is lost in making this recognition.

That

there might be various legitimate interpretations of a given

event

does

not

mean

that

each

interpretation

is

illegitimate.

The

plausibility

of

alternative

264

interpretations does not show the disvalue, simple-

mindedness or falsity of any particular interpretation;

rather it shows the enormous complexity of the event in

question.6

Besides the fact that Russell’s objection to coherence

presumes that there is a way the world is,

it also presumes

that there is a world out there.

The above discussion,

along with my remarks in chapter seven, speak against this

realist assumption, so I will not repeat them here.

Suffice

it

to

say,

I don’t see

that Russell has shown his

presumption to be warranted.

The raising of verificationism above leads into the

third objection that was mentioned regarding coherence.

The

objection is that there is no necessary connection between

coherence and verification(ism).

Statements can cohere (or

fail to cohere) independent of whether or not we can verify

them as coherent or as warrantable.

Yet the discussion of

coherence seems to rely on verification(ism) as a criterion

for acceptability (since correspondence with facts

is not

available to the view of coherence presented here).

I want

to suggest that this objection fails and it fails because it

presumes that this sharp distinction can be made between the

logical or semantic issue of whether or not statements

cohere and the epistemic issue of whether or not (and how)

we can verify that statements cohere.

As I have indicated

throughout this book, one of the points I am arguing for is

that the former issues cannot be isolated from the latter;

265

 

epistemic issues and concerns underlie ontic and semantic

issues and concerns.

With regard to the issue at hand, I

want to suggest that

it makes no sense to claim that

statements can and do cohere apart from saying that we can

verify this.

Consider again that semantic issues are

linguistic and that language is

a public process based on

communal activities and practices.

As such, semantic and

linguistic issues

arise from and are determined by

those

underlying forms of life.

The move from verifiability to coherence is,

it seems

to me, quite straightforward.

As Waismann (1965) argues:

Suppose there is a metal ball in front of me, and

I have the task of finding out whether the ball is

charged with electricity.

To do

that I connect

the balls with an electroscope and watch whether

the gold leaves diverge.

The statement ‘The gold

leaves

of

the instrument diverge’

(3)

describes

the verification of the statement ‘The ball is

charged’ (p).

Now what exactly am I doing when I

describe the verification of

the statement p?

I

establish a connection between two statemenfs by

declaring that the one (s)

is

to follow from the

other (p).

In other words, I lay down a rule of

 

inference which allows me

to pass from the

statement ‘The ball is charged with electricity’

to another that describes an observable situation.

By doing this I connect the statement with another

one, I make it a part of a system of operations, I

incorporate it

into language,

in short,

I

determine the way it is to be used.

 

In this sense

giving

the verification of

a statement is

an

important part of giving its use,

or,

to put it

differently, explaining its verification is

a

contribution to its grammar. (1965, p.123).

The problem is not

to make

the move from verification to

coherence - this is what Waismann is arguing for here - but

to make the move from coherence to verification(ism).

That

is,

while

there might be

a necessary connection from

266

 

verification to coherence,

is there a necessary connection

from coherence to verification(ism)?

It seems to me

that

there is and it is based on the same motivations as for the

converse connection.

To be committed to coherence, as we

saw above, is to be committed to a conception of truth that

emerges ultimately from communal language and forms of life.

To claim that a statement is

true

or

that two statements

cohere

is,

in

the final analysis,

meaningless apart from

some communally-based understanding (i.e.

apart from some

public, social, behavioral - verifiable - criterion of

understanding).

To make the claim that the semantic issue

of coherence can be dealt with apart from the epistemic

issue of verification is

to presume exactly what has

been

argued is a misconception of language, that it is not rooted

in shared human practices.

Coherence and verification(ism),

then, are inherently connected.7

Finally,

I want

to consider

the fourth objection to

coherence noted early

in

this chapter.

The objection is

that under a coherence theory of truth it is systems of

statements, not particular statements, that are true.

This,

it

is said,

is unintuitive.

Surely, individual statements

have truth values.

Now, it seems to me that this objection

is simply mistaken (and I must admit that I am somewhat

bewildered as

to why such an objection has ever been

raised).

A coherence theory of

truth does not deny that

particular statements have truth values, but the truth

values

that

they

have

are determined their relation with

267

other statements.

To say that truth is determined by or a

function of correspondence with facts is not to deny that

particular statements have truth values.

Likewise,

to say

that truth is determined by or a function of coherence is

not to deny

that particular statements have truth values.

It

is simply to give an account of how the particular

statements have the truth values that they have.

Where,

then, does this all lead?

Even if we can make

sense of

a coherence theory of truth,

so what?

What work

can this do for us?

As I will argue in the next chapter,

one thing that this can do for us

is

to provide a view of

reference that is an alternative to the received theories of

reference.

Given a coherence theory of truth and a related,

non-realist, conception of individuation and similarity we

can begin to construct an alternative view of reference that

does

not suffer

the problems of

the received theories of

reference.

To this task, then, we will now turn.

268

NOTES

1Russell sees a second difficulty, viz. coherence rests

on the notion of correspondence.

He suggests that coherence

presupposes the truths of

the laws of logic.

His argument

is as follows:

Two propositions are coherent when both may be

true, and are incoherent when at least one must be

false.

Now

in order

to know whether two

propositions can both be true, we must know such

truths as the law of contradiction.

For example,

the two propositions, ‘this tree is

a beech’ and

‘this tree is not a beech’, are not coherent,

because of

the law of contradiction.

But if

the

law of contradiction itself were subjected to the

test of coherence, we should find that,

if we

choose

to suppose it false, nothing will any

longer be incoherent with anything else.

Thus the

laws of logic supply the skeleton or framework

within which the

test of coherence applies,

and

they themselves cannot be established by this

test. (1912, p.123).

I

want

to

make

three points with regard to

this

argument.

First, Russell seems to be identifying coherence

with non-contradictoriness.

One of

the points of this

entire chapter is

to suggest that

this

identification is

somewhat mistaken, so

I will say no more about it at this

point.

Second, given Russell’s own theory of types and his

endorsement

of

Tarski’s

object-language/metalanguage

distinction, I am somewhat surprised by his comments here.

What

is wrong with saying that coherence pertains

to

the

object-language?

His move to considering whether or not the

law of contradiction is coherent can be construed as a move

to the metalanguage.

If it is, it’s not obviously fatal to

the

object-language notion of

coherence.

Third,

Russell

seems to

be speaking as though the laws of logic are ‘out

there’ prior to the rest of our beliefs, independent of any

other system.

Yet, as Quine remarks:

Mathematical and logical laws themselves are

not immune to revision if

it

is found that

essential simplifications of our whole conceptual

scheme will ensue....Logical laws are the most

central and crucial statements of

our conceptual

scheme, and for this reason the most protected

from revision by

the force of conservatism;

but

because again of their crucial position, they are

the laws an apt revision of which might offer the

most sweeping simplification of our whole system

of knowledge. (1972, p.3).

Russell’s conception of the status and nature of the laws of

logic is,

I think, mistaken.

Hence, his conclusion that any

abrogation of

the

laws

of logic so

as

to yield coherence

incoherent is also mistaken.

 

2This theme runs throughout Wittgenstein’s later works.

For example, cf. (1953, ##43, 133,

150, 595) and (1969,

##31, 229, 472, 476, 509, 554).

3cr. J.H. c111 (1966).

4cr. (1969, ##225, 410).

5If there are,

say,

two internally consistent systems

which are inconsistent with one another, then we can either

accept that they are inconsistent with each other (yet

internally consistent) or we can try

to remove the

inconsistency (and show one system to be false) by proposing

that perhaps for the sake of completeness each system adopt

a

new

statement

S,

which would

result

in

one

system

remaining internally consistent,

but

more complete,

while

the other system becoming inconsistent.

For example,

the

debate between evolution and creationism might be seen as

this sort of situation.

So, it might be said, carbon dating

has shown that the world was not created, say, in 4004BC, or

was not created in seven days.

The difficulty with this is

that creationist do not see

this

sort of evidence as

inconsistent with their view.

Following Duhem and Quine,

one can argue that it is never a single hypothesis that is

tested or falsified.

Instead a whole system is tested and

can gel and transform if need be.

In actual cases of

conflicts between internally consistent systems, it

is

rarely (if ever) the case that, external to those systems, a

statement can be presented which will demonstrate that one

system is false and the other true.

6Paulos (1985) presents a humorous example showing that

there is no single,‘right’ interpretation of events:

Since a Roman scholar was visiting Timur’s court,

the emperor asked the Mulla to ready himself for a

battle of wits with the scholar.

The Mulla piled books with ficticious,

but

impressive-sounding

titles

on

his

donkey,

books

such

as The Theory of Universal

 

Bifurcants,

Erosion and Civilization, A Critique of Tolerant

Purity, §3E1a1 Origins of MEntal De-actTViation.

The day of

the COSTESt the Mulla appeared in

court with his donkey and books.

His native wit

and

intelligence

overwhelmed

the

Roman

scholar,

who finally decided to test the Mulla’s knowledge

of theoretical matters.

The Roman scholar held up

one finger.

The Mulla answered with two fingers.

The Roman held up three fingers.

The Mulla responded with four.

The scholar showed his whole palm to which

Mulla responded with a closed fist.

The scholar then opened his briefcase and

 

out an egg.

The Mulla responded by digging an

onion out of his pocket.

The Roman said, "What’s your evidence?"

The Mulla answered, "The Theory of Universal

Bifurcants, A Critique of TETErant Purity," etc.

When the Roman spittered that he’d never

heard of those titles, the Mulla responded “Of

course,

you haven’t.

Look and you will see

hundreds of books you’ve never read."

The Roman looked and was so impressed that he

conceded defeat.

Since no one had understood any

of

this,

later,

after

refreshments

had

been

served,

the Emperor leaned over and asked the

Roman scholar the meaning of it all.

"He

is

a brilliant man,

this Mulla," the

Roman explained.

"When I held up one finger,

meaning that there was only one God,

he held up

two

to say

that He created heaven and earth.

I

held up three fingers, meaning the conception-

life-death cycle

of

man,

to which the Mulla

responded by showing four fingers, indicating that

the body is composed of four elements - earth,

air, water, and fire."

"Well then, what about the egg and the

onion," the Emperor pressed.

"The egg was the symbol of the earth (the

yolk)

surrounded by

the heavens.

The Mulla

produced an onion indicating the layers of heavens

about the earth.

I asked him to support this

claim to assign the same numbers of layers of

heavens as there are layers of onion skin and he

supported his claim by all those learned books of

which I alas an ignorant.

Your Mulla is

a very

learned man indeed."

The dejected Roman then

departed.

The Emperor next asked the Mulla about the

debate.

He replied, "It was easy, Your Majesty.

When he lifted a finger of defiance to me,

I held

up two, meaning I’d poke both his eyes out.

When

he held up three fingers indicating, I’m sure,

that he’d deliver three kicks, I returned his

threat by threatening four kicks.

His whole palm,

of course, meant a slap in my face, to which I

responded with my clenched fist.

Seeing I was

furious, he began to be friendly and offered me an

egg so I offered him my onion."

7The connection I am making here between coherence and

verification(ism) is not the connection made by the ‘logical

positivists’, viz.

the meaning of

a term is

its method of

verification.

(A number of authors have suggested the sort

of non—positivist verification(ism) that I am considering in

the text.

Cf. Waismann (1965), Dummett (1978) and Luntley

(1982).)

CHAPTER NINE

COHERENCE AND REFERENCE

In

this final chapter I want to expand on

the

conception of coherence explored in the previous chapter by

considering a view of reference based on such a conception

(and which is committed to such a conception).

This view of

reference is, following Kripke, suggested as

a ‘better

picture’ and is only in its conceptual nascency.

This view

of reference, which I see as

a genuine alternative to the

received theories, is not, again following Kripke, a theory

of reference in the sense of offering a set of necessary and

sufficient conditions for successful reference to occur.

I

do not, however, see this as a weakness of the view, because

I

think that reference is

too complex a process to

be

capturable by

the elaboration of necessary and sufficient

conditions.

Furthermore, I don’t even see the establishment

of

a synchronically-determined, permanent set of necessary

and sufficient conditions as

a desideratum (as

I

think my

stand against essentialism, correspondence and realism

suggest).

In presenting this view of reference, I want to

first argue that the view of reference which the later

Wittgenstein proposes is very much in line with the sort of

view

I

am proposing.

I want

to show how and why

Wittgenstein’s conception of reference is not

the same as

272

Searle’s (contra Kripke) and not the same as Kripke’s

(contra Cappio (1981)).1

I will then remark on what I take

Wittgenstein’s view of reference to

be.

Finally,

I will

suggest what I see as

the strengths and weaknesses of his

view of reference,

indicate why I think it

is

a more

plausible view than the received theories of reference, and,

in conclusion, argue for the (initial) plausibility and

superiority of

my alternative view of

reference.

First,

then,

I want to turn to Wittgenstein and show that he does

not endorse a cluster theory of reference.

Kripke (1980) claims that a good example of the cluster

theory of reference is found in Wittgenstein (1953).

Kripke

quotes part of paragraph 79 as introducing the idea of

family resemblances (which is incorrect; family resemblances

are introduced in paragraph 67).

In addition, says Kripke,

this paragraph is a forceful statement of the view that "the

reference of

a name

is determined not by

a single

description but by some cluster or family" (1980,p.3l) and

is

a good example of

the view that "what we really

associate" with a

name

is

a family or cluster of

descriptions.

Kripke quotes the following part of paragraph

79:

Consider this example.

If one says ‘Moses did not

exist’, this may mean various things.

It may

mean:

the Israelites did not have a single leader

when they withdrew from Egypt-or: their leader was

not called Moses—or: there cannot have been anyone

who accomplished all that the Bible relates of

Moses-...But when I make a statement about Moses,-

am I always ready to substitute some one of those

descriptions for ‘Moses’?

I shall perhaps say: by

‘Moses’ I understand the man who did what the

273

  

Bible relates of Moses, or at any rate a good deal

of it.

But how much?

Have I decided how much

must be proved false for me

to give up

my

proposition as false?

Has the name ‘Moses’ got a

fixed End unequivocal use for me in all possible

cases?

Now,

there is

an important jump made here which is

a

clear misunderstanding of Wittgenstein’s view.

Kripke

doesn’t

say

much

more

directly

related

to

this

misunderstanding, but I think it is fair to believe that he

is guilty on this point.

Swartz (1977), however, explicitly

commits

the error,

viz.

taking Wittgenstein to

be saying

that we only associate names with a family of descriptions;

taking Wittgenstein to

be saying

that words

have meaning

only in virtue of family resemblances.

Says Swartz:

According

to

one

contemporary version

[of

a

descriptionist theory of reference] there is not a

 

conjunction of properties associated with each

term,

but there is a 5133335 of properties

associated with it.

For example, it is held that

we

cannot define

‘game’

by

a conjunction of

properties such as having a winner and a loser,

being entertaining, involving the gaining and

losing

of

points,

because

some

perfectly

acceptable games lack some of these features.

According to

the cluster theory,

something is

a

game because it has enough features from a cluster

of properties like these.

A cluster theorist

would claim that there need not be any property in

the cluster that is sufficient for the application

of

the

term,

but he nevertheless holds

that the

cluster taken as a whole determines the extension

of

the

term.

Wittgenstein’s position that there

are only family resemblances among the individuals

in the extension of many ordinary terms can be

construed as

a version of

the cluster theory.

(1977, p.15).

True, Wittgenstein’s position could be construed as

a

version

of

the

cluster

theory,

but

only

on

a

misunderstanding and false representation of his position.

274

 

Wittgenstein’s

point

in introducing

the

notion of

family

resemblances is

to show that it is a mistake to insist that

because we use the same word to refer to different objects

that there must be some essential property - some common and

peculiar characteristic(s) - of those objects in virtue of

which we use the same word for them all.3

This is not to

say that words have only family resemblances; some words may

well have a single (conjunction of) description(s) giving

the meaning of the word, e.g. ‘bachelor’ (‘adult human male

who has never been married’) or ‘mule’ (‘offspring of a

female horse and a male donkey’).

Instead,

it

is

to say

that many words do not have a single (conjunction of)

description(s) giving the meaning of the word.

Furthermore, if indeed the cluster theory "holds that

the cluster taken as a whole determines the extension" of a

term,

it

is obvious that

this

is not Wittgenstein’s view.

The major reason for this is that Wittgenstein denies that

it makes any sense to even speak of "the cluster taken as a

whole.‘

This is because words (at least some words) are not

"closed by a frontier" (1953, #68); they have, in Waismann’s

(1965) terminology, an ‘open texture’.4

This notion of ‘open texture’ is important and needs to

be fleshed out before expounding on Wittgenstein’s view of

reference.

In paragraph 68 of (1953) Wittgenstein says:

"All right:

the concept of number is defined for

you as the logical sum of these individual interrelated

concepts: cardinal numbers, rational numbers, real

numbers,

etc.;

and

in

the

same way

the concept of

a

game as

the logical sum of a corresponding set of sub-

concepts." -It need not be so.

For I can give the

275

 

 

concept ‘number’ rigid limits in this way, that is, use

the word "number" for a rigidly limited concept, but I

can also use it so that the extension of the concept is

not closed by a frontier. (1953, #68).

What Wittgenstein is getting at is,

on the one hand,

the projectibility of words (not Goodman’s sense of

projectibility);

that is,

the propriety and legitimacy of

using a word or term in a novel manner or context (e.g.

‘feed’ in ‘feed the parking meter’); and, on the other hand,

the lack of distinct boundaries or a clear extension of

a

word or term (e.g. ‘game’).

Waismann,

who coined the

term ‘open texture’,

states

that open texture may mean that we could get acquainted with

some totally new experience or that some new discovery could

be made which affects our interpretation(s) of certain facts

and events.

He offers the following example:

The notion of

gold

seems

to

be defined with

absolute precision,

say by

the spectrum of

gold

with its characteristic lines.

Now what would you

say if a substance was discovered that looked like

gold, satisfied all

the chemical tests for gold,

whilst it emitted a new sort of radiation?

‘But

such things do not happen.’

Quite so,

but

they

might happen, and that is enough to show that we

can never exclude altogether

the possibility of

some unforeseen situation arising in which we

shall have to modify our definition.

Try as we

may, no concept is limited in such a way that

there is no reason for any doubt.

We introduce a

concept and limit it

in some directions;

for

instance, we define gold in contrast to some other

metals such as alloys.

This suffices for our

present needs,

and we

do not probe any farther.

We tend to overlook the fact that there are always

other directions in which the concept has not been

defined.

And if we did, we could easily imagine

conditions

which

would

necessitate

new

limitations.

In short, it

is not possible to

define

a

concept

like

gold

with

absolute

precision, i.e.

in such a way that every nook and

cranny is blocked against entry of doubt.

That is

276

 

what is meant by

the open texture of

a concept.

(1965, p.126).

Now, the point here is this: when it comes to giving an

account of (most) words (and names), we are not able to

state a set of necessary and sufficient conditions for the

applicability of those words.

Specifically, for names, we

are not able - contrary to both the cluster theory and the

causal theory -

to establish such conditions prior to and

independent of the use of names in language games (and,

finally,

in our forms of life).

We cannot say ‘ahead of

time’ as

it were what conditions must be met for a name to

refer to an object and for what, in general, naming is.

A

mistake of

the cluster theory is

the attempt to establish

these conditions through the notion of a disjunctive set of

descriptions associated with a name; a mistake of the causal

theory is

the attempt to establish these conditions through

the notion of

a causal chain connecting an utterance of a

name with an original baptismal act in which an object is

referred to by the (use of the) name.

This aspect of open

texture is, for both Wittgenstein and Waismann, an empirical

fact about our use of language (and names); we do

in fact

often use names without any fixed meaning.6

This leads to a question which was touched upon earlier

in laying out Searle’s cluster theory,

and Wittgenstein’s

answer to

it which not only differentiates his view from

Searle’s, but also shows its greater plausibility over

Searle’s.

The question was: does at least one member of the

disjunctive set of descriptions associated with a name need

277

 

to be

true for reference to successfully occur?

Searle’s

answer seemed to be ‘yes’.

Wittgenstein’s ‘answer’ is that

the question is

a wrong-headed one;

it displays a

misunderstanding of

the relationship between names and

descriptions.

For Wittgenstein, names can be used as

abbreviations for descriptions, although in such cases the

name would not function as

a name,

but as

a description.

The mistake which Searle (in his theory) and Kripke (in his

criticisms) make is

to think that names are synonymous with

(one or more) descriptions.

For Wittgenstein, names can be

explained by descriptions, but they are in no way synonymous

with them (cf. (1953, ##43-45)).

For example, suppose Jones

says something about N and Smith wants to know who N is,

to

whom reference is being made).

Jones remarks that N is the

man at the bus stop over there.

Now this description does

not define ‘N’

in any sense -

the description is not a

synonym and no necessary and sufficient conditions of

reference are laid down - rather it helps to identify N for

Smith by providing instructions (by giving an explanation in

a sense) of finding out who N is.

If the explanation fails

-

if Smith cannot pick out N

- another description may be

given (perhaps, ‘N has on a grey hat’).

It is these sorts

of considerations that Wittgenstein has in mind when he

speaks of various descriptions that we could use to give the

meaning of ‘Moses’ and that we use a name without any fixed

meaning.

Although descriptions

explain

(and

perhaps,

in

some cases, make possible) the use of names,

the meanings

278

 

(or uses)

of

the descriptions are not identical with the

meaning (or uses) of the names.7

The above remarks sound quite similar to what Kripke

has to say about descriptions as reference fixers and indeed

it is similar.

However, Kripke’s account of descriptions as

reference fixers seems

to miss Wittgenstein’s main point,

viz. the open texture of names.

Kripke says:

Now another view might be

that even though

the description in some sense doesn’t give the

meaning of the name, it is what determines its

 

reference and although the phrase ‘Walter Scott’

isn’t synonymous with ‘the man such that such and

such and such and such’, or even maybe with the

family (if something can be synonymous with a

family),

the family or the single description is

what is used to determine to whom some-one is

referring when he says ‘Walter Scott’.

...some of

the attractiveness of the [descriptionist] theory

is lost if

it isn’t supposed to give the meaning

of the name; for some of the solutions of problems

that I’ve just mentioned will not be right, or at

least won’t clearly be right,

if

the description

doesn’t give

the meaning of

the name.

...If,

on

the other hand, ‘Moses’ is not synonymous with any

description, then even if its reference is in some

sense determined by a description, statements

containing the name cannot in general be analyzed

by replacing the name by a description, though

they may

be materially equivalent to statements

containing a description. (1980, pp.32-33).

Attractiveness notwithstanding,

one

of Wittgenstein’s

concerns

in

talking about open

texture

is precisely

that

it’s mistaken to

think

that a name can in general be

analyzed by replacing it with a description (cf. (1953,

#63).

An identifying description helps to pick out the

referent of a name for a given speaker or listener in a

given context;

in

a different situation with a different

speaker or listener the identifying description might not

 

succeed

in picking out

the

referent,

whereas another

description might.

Nor is

it Wittgenstein’s view that

offering

descriptions

as

identifying

descriptions

rather

than as synonyms of names

is intended to replace talk of

meaning of names.

That is,

to replace talk of descriptions

as synonyms of names with talk of descriptions as reference

fixers

(i.e.

identifying descriptions)

is

not

to abandon

talk of meanings for Wittgenstein.

Again, Kripke seems to

be assuming that for names to have meanings, they must have

fixed meanings, which, as noted above in the discussion of

open texture, Wittgenstein denies.

Besides

the error of supposing that names are

synonymous with their identifying descriptions, philosophers

have committed the error of supposing that they are able to

uniquely pick out objects prior to or separate from some use

of identifying descriptions.

This, for Wittgenstein, is

a

mistaken picture of how reference (and language) works.

As

he says:

For naming and describing do not stand on the same

level:

naming is

a preparation for description.

Naming is so far not a move in the language-game—

any more than putting a piece in its place on the

board is a move in chess.

We may say: nothing has

so far been done, when a thing has been named.

It

has not even

ot a name except in

the language-

game. (1953, #%9).

A candidate which is commonly offered in support of our

ability to uniquely pick out objects prior to

or separate

from some use of identifying descriptions is ostension.

However, as shown in chapter five, ostension only works

within agreed-upon contexts

(i.e.

in language-games).

280

   

 

Rather than the (objective, realist) identity of an object

giving us the use (referent) of a name, the converse is the

case; it is the use of a name which gives the identity of an

object.

This points to the important social nature of names

and naming, and it also points to the epistemic primacy of

naming.

This epistemic primacy becomes apparent with the

elaboration of

the

issue of

the place

of evidence in

reference.

Consider this situation: Jones is

an incoming

student in the first term of graduate study in philosophy.

The faculty which Jones has so far encountered at Philosophy

University have all been bearded and elderly.

Jones decides

to enroll in a course taught by Noseworthy.

That is, Jones

enrolls in a course, the description of which in the college

catalog states is taught by Noseworthy.

At the first class

meeting a clean-shaven, young-looking person enters and

begins lecturing.

During the course of

the lecture Jones

overhears several students

remark

that

they

think the

lecturer is

a graduate assistant (being so young-looking).

Later that day Jones sees

the same lecturer sitting in an

office which Jones knows is not the office of any other

faculty member.

The following day Jones passes

the

same

office and sees the same lecturer there.

As Jones continues

down the hall, Jones meets Smith, who asks, "Have you seen

Noseworthy around today?"

Jones replies, "He’s in his

office right now, down the hall."

Now,

in this case Jones

successfully refers; indeed Noseworthy is

the clean-shaven,

young-looking lecturer.

Here is what I find important about

this example: the causal theorist will claim that Jones

referred

to Noseworthy in virtue of

the causal chain

connecting Noseworthy and Jones’s contact with the name

‘Noseworthy’ which Jones read on the course description in

the college catalog.

But what seems involved in this case

is the relevance of

the evidence Jones had which led Jones

to infer that this lecturer was in fact Noseworthy.

It

seems much more likely in this case than in many others

which the causal theorists cite

that evidence plays a

genuine role in Jones being able to hook up a name with an

object.

On

the relation between naming and evidence,

Steinman (1982) has argued that the causal chain connecting

an object and an utterance of its name is not necessary and

suggested a spectrum of cases that involve the relationship

between naming and evidence.

Steinman offers the following

example to support his position.

Suppose Bob is at a party

and

sees

the wife

of

his

friend Dave.

During their

conversation Dave’s wife informs Bob that Dave names all of

his pets ‘Sal’.

She also tells Bob that, Dave being a one-

pet man,

there is never any confusion as

to what he is

referring to when he utters the name ‘Sal’.

At any time

Dave has at most one pet and that pet’s name is ‘Sal’.

Several days later Bob comes over to Dave’s house and Dave

shows Bob his new pet goldfish.

Upon seeing the fish for

the first time, and not having heard Dave mention the pet’s

name, Bob remarks, "Sal sure is

a fine specimen."

In this

case, Bob’s utterance of the name is not causally connected

282

 

to

the object, even though he successfully referred to the

object.

The evidence that Bob had about the name of

the

object (e.g. Dave is a one-pet man; Dave names all his pets

‘Sal’; Dave’s wife is known to be a reliable, truthful

person, etc.) is what made reference possible.

Back to

the Noseworthy example (in which evidence

played a part for reference to have occurred), the point is

not to deny that a causal chain existed between Noseworthy

and Jones’s utterance of ‘Noseworthy’.

The presence of

a

causal chain connecting utterances of

a name with the

referent of

the

name

does

not seem to

be

a point of

contention between

the

causal

theorists and Wittgenstein;

both would admit that a causal chain exists between an

object and

the name-utterance(s), at least in many cases.

Where the conflict of views occurs (at least in part) is in

the referring status granted to descriptions.

(Of course,

this

is

only symptomatic of

deeper differences.

For

example, as Putnam (1981) argued,

causal chains cannot be

established or determined without reference having already

taken place.)

Quite simply, for Wittgenstein, there is no

essence to naming.

He directly speaks to this in paragraphs

37 and 38 of (1953):

What is

the relation between a name and the

thing named?...This relation may...consist, among

many other things,

in

the fact that hearing the

name calls before our mind the picture of what is

named;

and it also consists, among other things,

in the name’s being written on the thing named or

being pronounced when the thing is pointed at.

...we call very different things ‘names’; the word

‘name’

is used to characterize many different

kinds of use of a word, related to one another in

283

 

many

different ways.

...Naming

[under Russell’s

account] appears as a queer connection of a word

with an object.-And you really get such a queer

connexion when the philosopher tries to bring out

the relation between name and thing by staring at

 
an object in front of him and repeating a name or

even the word ‘this’ innumerable times. ...And

here we may indeed fancy naming to be some

Pemarkable act of mind, as it were a baptism of an

object. (1953, ##37-38).

Tied in with this belief that there is no essence to

naming

is

the

main

point

of

difference

between

Wittgenstein’s view and the causalist view.

The difference

is found in paragraph 87 of (1953):

Suppose I give

this explanation: "I

take

‘Moses’ to mean the man, if there was such a man,

who led the Israelites out of Egypt, whatever he

was called then and whatever he may or may not

have done besides."-But similar doubts [about

meaning and reference] to those about ‘Moses’ are

possible about the words of this explanation (what

are you calling ‘Egypt’, whom the ‘Israelites’

etc.?).

Nor would these questions come to an end

when we got down to words like ‘red’, ‘dark’,

‘sweet’.-"But then how does an explanation help me

to understand if after all it is not the final

one?

In

that case

the explanation is never

completed; so I still don’t understand what he

means, and never shall!"

-As though an explanation as

it were hung in the

air unless supported by another one.

Whereas an

explanation may indeed rest on another one

that

has been given, but none stands in need of

another-unless we require

it

to prevent a

misunderstanding. __One might say:

an explanation

serves

to remove or to avert a misunderstanding—

one,

that is,

that would occur but for the

explanation; not every one that I can imagine.

...The sign-post is

in order-if, under normal

circumstances,

it fulfils its purpose.

(1953,

#87).

An immediate point of these remarks in relation to the

problems of naming and reference is this: in some cases (at

least) the presence of

a causal chain is unnecessary as an

explanation of how reference occurs.

Causal chains may well

284

serve as descriptions of how reference occurs (i.e. Searle’s

‘external observer’) and they may even serve (in some cases)

as explanations of how reference occurs, but if what we want

is an understanding of how we refer by naming, then causal

chains are not needed.

Wittgenstein’s reasoning is that we

require explanations in order to solve problems of failure

to understand.

Explanations serve to remove or avert

misunderstandings, but these explanations are needed only in

such a capacity; if there were no misunderstanding that

needed removing or averting,

the explanation would not be

needed.

When it comes to naming, the problems that actually

occur, such as knowing the correct name of someone or when

to use one name rather than another when referring, are

problems such that any misunderstandings that might arise

(at least, generally) can be handled without having to bring

in talk of underlying causal chains.

This is not to deny

that causal chains may be invoked in some explanations, it

is only to claim that in general such explanations are not

necessary to provide an understanding of how naming and

referring occur and are successful.

Now,

I

think

that

the

view

of

reference

that

Wittgenstein has suggested is a more plausible view than the

views proposed by

the received theories of reference, not

only because it gives a more plausible account of ordinary

cases of referring, but also because it both captures and

reflects the coherence, verificationist perspective outlined

in

the previous chapter (and does not suffer the flaws of

285

 

commitments

to

privacy,

essentialism,

correspondence

and

realism).

I will now try to support this claim by

considering the sorts of cases which Searle and Kripke

discuss as well as suggest other cases which indicate the

greater plausibility of the Wittgensteinian view.

Examples

As noted in chapter one, Kripke thinks

that Searle’s

theory of reference is "wrong from the fundamentals" to the

extent that the theory holds

that "we give ourselves some

properties which somehow qualitatively uniquely pick out an

object and determine our reference in that manner" (1980,

pp.93-94).

Kripke and others have offered counter-examples

which are intended to show that a speaker can refer to an

object even though the speaker has no descriptions

associated with a name which are true of the object to which

reference is made.

(Kripke’s Columbus, Einstein and deel

examples,

for

instance,

are meant

to show

this.)

Searle

responded

that

the Kripkesque counter-examples fail

to

do

their job because it’s not the case that a speaker has no

descriptions associated with a name which are

true of

the

object

to

which

reference

is

made

and

yet

reference

succeeds.

For Searle, the particular descriptions which we

have associated with a name and which are not true of

the

object to which reference is made are usually descriptions

which reflect the secondary aspect under which reference is

made.

As such,

reference can succeed for Searle because

8

the speaker can always fall back on another description to

 

secure

reference.

Whichever description succeeds

in

securing reference reflects the primary aspect under which

reference is made and can do so even if the description is

not true of the object to which reference is made.

For

Searle,

then,

Kripke’s

counter-examples

fail

to

show his

theory to

be mistaken.

However,

as

I claimed in chapter

one, given Searle’s theory it does seem (at least for the

primary aspect under which reference is made) that at least

one member of the disjunctive set of descriptions associated

with a name must be true of an object for the name to refer

to that object.

Whether or not Kripke’s particular counter-

examples work,

one might construct a counter-example that

does

work.

I

then offered two such counter-examples

(at

least

two plausible candidates as counter-examples).

The

first was the case of my neighbor complaining to her friend,

"My weird neighbor kept me up last night yelling something

about ‘that damn turtle’ or ‘deel’ or something like that.

I had to pound on the wall to get him to quiet down."

Has

my neighbor,

I

asked,

referred to deel?

Given Searle’s

theory as presented in his (1969), his rule 5 (and perhaps

rule

6)

is violated,

so reference,

it would seem,

is not

made.

[Rule 5:

S intends that the utterance of R will pick

out or identify X

to

H.

Rule

6:

S

intends

that

the

utterance of

R will identify X

to

H

by means

of

H’s

recognition of

S’s intention to identify X and be

(S)

intends this recognition to be achieved by means of H’s

knowledge of

the rules governing R and his (H’s) awareness

287

 

of (context) C.)

In this case, my neighbor does not seem to

intend to pick out any object by her utterance of ‘Gddel’,

so reference is not made.9

(It should be noted, however,

that in (1983) Searle backs away from this view and suggests

that in some cases - perhaps this case - reference can be

made

"where

one

is almost

totally dependent on other

people’s prior usage to secure reference" (1983, p.261).

To

the extent that Searle moves to this sort of view, he is,

I

think, abandoning his (1969) theory and adopting the sort of

view which I have identified with Wittgenstein.)

Now,

again,

has

my neighbor referred to deel?

For

Searle,

no,

because she didn’t intend to pick out deel.

This

answer

seems

(to

me,

anyway)

intuitively plausible.

However, it runs into the sort of problems that Kripke and

others point out,

viz.

that names

are historical,

public

entities and whether or not one intends to pick out deel by

(the use of) ‘deel’, that name (usually) picks out deel.

For Wittgenstein, as opposed to Searle, my neighbor has not

referred because she

is not in

the appropriate context to

refer to deel in this case.

That is, her lack of intention

to refer to deel might play a part in

the lack of

reference, but it certainly is not sufficient.

Rather, she

was not engaged in

the language game of referring.

Here,

the use/mention distinction (cf. note 9 above) would be

relevant because it indicates a difference of language games

in which the speaker is engaged.

A second counter-example to Searle’s position which I

288

proposed for consideration was

the

following:

Suppose my

obsession with deel has gotten to

the point that I often

blurt out my innermost thoughts concerning him.

On one such

occasion, while standing at a bus stop, I mutter, "Ha!

Gddel couldn’t hold a candle to Cantor.

Now,

there was a

man!"

Upon hearing this,

a bemused and intrigued stranger

next to me asks, "Who is GUdel?"

Has the stranger, I asked,

referred to deel?

Searle, I suggested, would say no,

because it’s not clear that the stranger intended to pick

out a particular object (deel).

Indeed, the very question,

‘Who is deel?’ shows that the speaker has no clear referent

associated with the name,

and in fact the question can be

paraphrased with no

loss

of meaning with a different

question,

‘Whom are

you

talking about?’ or

‘Whom are you

referring to when you say "deel"?’

(Again, in his (1983)

Searle seems to back away somewhat from his earlier theory.)

For Wittgenstein,

it

seems

to

me,

reference

could easily

occur

in

this

case.

Within the language game

of

questioning, one could refer.

Indeed, given the public

nature of names, the stranger could refer to deel, but only

because

the publically- and historically-determined nature

of names allows a speaker to use names which have an already

publically-determined sense (referent).

This, of course, is

much of

the motivation for Kripke’s critique of Searle, and

to

this extent be

is,

I

think,

right.

For other various

reasons (which I have indicated in the previous several

chapters), however, Kripke’s alternative account is

289

 

mistaken.

Not only can Wittgenstein’s view account for

reference being made in this case by the stranger to deel,

but his view can also account for reference being made to

abstract and future objects, something troublesome for the

causal theory.

Because Wittgenstein’s view does not require

the intentionality of the speaker for reference to succeed,

it

is not constrained in the ways that Searle’s theory is.

In addition,

because Wittgenstein’s view does

not require

appropriate causal chains for reference to succeed,

it

is

not constrained in

the ways

that Kripke’s theory is.

For

example, Wittgenstein’s view can account in similar ways for

successful reference in the following three cases:

‘Who is

G5del?’, ‘Who is God?’ and ‘What is infinity?’, each spoken

by

a speaker who has never heard the name (or concept) in

question.

Reference can occur in each of these cases (and

the myriad of cases of which they are instances) because the

‘objects’ referred to and the names (or, in the case of

infinity,

the

term)

used are shared,

public,

communal

conventions.

A speaker who has no (true) descriptions

associated with Gddel (other than, perhaps, ‘the name I just

heard’) can nonetheless refer to Gddel by (the use of)

‘Gddel’,

not

in virtue of

some appropriate causal

connection,

but because

the speaker is

a member of

a

linguistic community and understands (at least tacitly) the

language game (and the practices and conventions upon which

it rests) of naming and of referring.

Not only is the

ability to name and refer something that is social, but so,

290

 

too,

is

the object to which reference is made (as I argued

in chapter five).

Likewise, reference to God is made in the

same way and for the same reasons.

The name/term ‘God’

functions in certain ways in certain communities and given

certain practices.

As such, speakers can refer to God

because who and what God is

is determined socially.

The

name/term has a ‘life’ independent of any given speaker, but

not independent of all speakers.

‘God’, like ‘Gddel’ is a

public, social, historical entity (as is, in one sense, God)

and reference can be made only because of

this public,

social,

historical nature.

(This holds also for infinity

and other abstract or ‘future’ entities, i.e.

the sorts of

‘objects’ which the causal theory of reference has

difficulty.)10

The Wittgensteinian view of reference, then,

I have argued is not troubled by the sorts of counter-

examples which plague

the cluster and causal

theories

of

reference.

However, I think that the strengths of this

alternative View can be seen more sharply by considering two

current

issues

which

implicitly

involve

questions

of

reference.

These examples also show the close conceptual

connection between

issues

of

reference and

issues

of

individuation and similarity.

As I will argue, these ‘real-

life’ examples will demonstrate the strengths of

this

alternative view of reference (and its underlying epistemic

commitments).

The first example is that of the current

discussion of the hypothesis that our sun is part of a

binary solar system.

The second example is that of the

291

 

controversy concerning the discovery of

the cause of AIDS

(acquired immune deficiency syndrome).11

In

1977, Alfred Fischer and Michael Arthur (both of

Princeton University) suggested that a periodic interval of

about 32 million years elapsed between mass extinctions of

life on

the earth.

Six years later, David Raup and John

Sepkoski (both of

the University of Chicago), after

examining paleontological records

dating back 250 million

years, argued that preliminary evidence, backed by

statistical tests, suggested a regular, periodic extinction

of life on the earth every 26 million years.

Included in

the series of extinctions is

the extinction of

the

dinosaurs, approximately 65 million years ago (which marks

the end of

the Cretaceous period and the beginning of

the

Tertiary period).

Numerous proposed explanations for

the

extinction of

dinosaurs have been advanced,

including

oncoming,

major

climatic

changes

and

early mammalian

predation on dinosaur eggs.

However, more than 75 percent

of species on the earth were extinguished, so predation

seems unlikely to be an adequate explanation.

Climatic

change also seems questionable, given recent views that

dinosaurs (at least some species of dinosaurs) were not

reptilian, but were in fact warm-blooded.

In any case,

a

recent account for

the mass extinction at

the end of

the

Cretaceous period is that a comet struck the earth, creating

a prolonged period of

darkness

over

the earth’s

surface,

killing off basic photosynthetic elements of the food chain,

292

and finally resulting in the extinction of most life forms

on the planet.

An important piece of evidence offered in

support of this account is the widespread and high levels of

iridium found in the clay separating the Cretaceous and

Tertiary geologic periods.

Iridium is rare on the earth’s

surface, but more abundant on extra-terrestrial bodies.

Also, the distribution of this iridium is such that the

difference in levels across the globe suggest a highest

concentration near Iceland, the levels gradually decreasing

away from this localization.

In early 1984 scientists met at a conference to discuss

the work of Raup and Sepkoski and the proposed cataclysmic,

extra-terrestrial account of periodic mass extinctions.

Among the scientists at the conference were Marc Davis,

Richard Muller (of

the University of California, Berkeley)

and Piet Hut

(of

the Institute for Advanced Studies) who

suggested that our sun is part of a binary system.

(Others,

e.g. Daniel Whitmire and Albert Jackson IV,

have also

suggested that our

sun

is part of

a binary system.)

The

companion star to our sun,

it

is claimed, is

the cause of

these noted periodic mass extinctions of terrestrial life

forms.

Davis, Muller and Hut have dubbed the companion star

‘Nemesis’ and have suggested that its orbit brings it close

enough to

our solar system to encounter (or

at least

disturb) the Oort Cloud, a halo of comets and matter on the

fringes of our solar system.

The presence of Nemesis

creates cometary havoc and results in comet showers within

293

our solar system, with enough comets striking the earth so

as to cause prolonged darkness and intense volcanic activity

on

the earth’s surface.

The orbit of Nemesis,

it

is

claimed,

brings

it within the crucial proximity every 26

million years.

Whether or not this hypothesis concerning the existence

of

a companion star

is

true is not important for

the

concerns

here,

except in

a sense which I will explicate.

What is important is the reference to Nemesis and the use of

‘Nemesis’ as a name.

There are several points which I think

emerge and are highlighted by this example.

First, the

introduction of

the name ‘Nemesis’ was done within the

larger context of proposing a scientific hypothesis.

Naming

Nemesis (and naming in general) was not done privately,

external to any other epistemic concerns, but as part of a

larger practice and activity.

Prior to the particular

scientific context,

established by

the various public

publications and positions espoused, the naming of Nemesis

would have made no sense.

Even with questions and

criticisms

about

the hypothesis and

the existence of

Nemesis, the introduction of the name ‘Nemesis’ is plausible

(including the choice of

the name) and conceptually

coherent.

The conceivability of someone proposing to

introduce

the name

‘Nemesis’

for

(or,

at

the very least,

successfully referring

to)

a companion star prior

to

the

introduction of a reasonable hypothesis which could serve as

a background context for this name is doubtful.

294

 

This leads

to

a second point which emerges from this

example,

viz.

it

is

the acceptance by

the community that

secures reference, not any intention by a particular speaker

nor any specific, appropriate causal chain.

The name

‘Nemesis’ has,

in a sense,

stuck, and even critics of

the

hypothesis use

the

name.

For example,

in scientific

newsmagazines Nemesis is often a topic worthy of report (cf.

Science

News

4/21/84,

9/1/84,

11/3/84,

3/16/85).

Scientists,

both in favor of and in opposition to the

hypothesis,

use

the

name

‘Nemesis’

in discussing

the

(de)merits of the hypothesis.

The simple introduction of a

name does not suffice for naming to take place.

Rather the

communal acceptance and adoption of

a name that is

introduced is necessary for naming to occur (and reference

to succeed).

For a speaker to intend to pick out an object

by using a name

(or,

better,

by dubbing it)

does

not

guarantee that naming has occurred.

Naming, as

a social,

linguistic practice,

requires social,

linguistic adoption.

(This will be seen more readily in

the discussion of

the

AIDS case below.)

A

third

point

is

that naming

(and reference)

is

intimately

tied

up with epistemic and verificationist

issues, ultimately reflecting a coherence position.

Whether

‘Nemesis’ actually names anything, whether it continues to

function as a name, is not determined by any a priori set of

conditions that are

met,

but by

the scientific community.

For example, consider this scenario (which has been

 

proposed):

Astronomers

are

presently

attempting

to

infer

where and how Nemesis might be found so as

to test the

claims of the hypothesis that such an object exists.

Even

if scientists can reasonably predict where they ought to be

able to ‘find’ Nemesis, the failure to find it does not mean

that Nemesis does not (did not) exist.

It could have been

destroyed since

it

last approached the Oort Cloud

(approximately eleven million years ago).

Or, it could have

finally escaped its orbit, leaving our sun a unary sun.

(Or,

Nemesis might not exist and

there

is another

explanation for the periodic mass extinctions, e.g. the

wandering of our solar system throughout the Milky Way

relative to other stars and systems.

As

the solar system

enters

different

areas

of

the

galaxy,

in

a

periodic

oscillatory pattern, cometary activity could result, causing

mass extinctions of life on earth.)

Now, suppose it were

the case that Nemesis is not found and no other evidence of

its existence is found.

The name would, I suggest, drift

out

of

the

scientific

vocabulary

(except

perhaps

anecdotally,

like Vulcan,

the supposed

‘invisible’

planet

thought

to explain the perturbations of Mercury’s orbit).

Without some verifiable evidence adding credence to the

existence of Nemesis,

‘Nemesis’ will fade from serious

scientific discussion and will not be seen as the name of an

object at all (i.e. will not be accepted as a name).

Quite

simply, if there is nothing to support and substantiate the

use of

the name,

the name will not be used;

in effect,

it

296

will not be a name.

This is true even if, as proposed,

Nemesis caused the periodic mass extinctions, but has now

left its old orbit and is now undetectable.12

The causal

theory of reference, of course, must hold that if

the

appropriate causal connection ‘is there’, then reference is

made.

In this case, we might actually be referring to

Nemesis by (the use of) ‘Nemesis’ even if, because of lack

of

corroborating evidence,

we

reject

the hypothesis

that

Nemesis exists.

On the causal theory, we might really be

referring even if we aren’t aware of

it and decide that we

aren’t really referring.

This strikes me as an obvious

problem with the causal theory.

(What sense would it make,

I want to ask,

to claim that we really did refer to Vulcan,

even though we never found any evidence to suggest that we

referred to Vulcan by (the use of) ‘Vulcan’?)

Now,

this

example

of

Nemesis nicely

illustrates

the

strength of

the alternative view of reference which I

associate with Wittgenstein and for which I am arguing.

The

linguistic,

semantic issue

is not whether or

not Nemesis

exists,

but whether or not the name ‘Nemesis’ is accepted

and adopted by

the scientific community (and later by

larger,

different communities).

For the causal theory of

reference, however, the ontic issue is significant, because

causal chains can only be established to objects in the

world.

Names, for the causal theory, name things; if those

‘things’

don’t exist,

then those

‘names’

don’t name

(and

aren’t really names).

In this case, I want to suggest that

297

whether or not Nemesis exists, ‘Nemesis’ can still be used

as

a name

(e.g.

if

the scientific community decides

that

Nemesis did cause the mass extinctions, but it has left its

orbit,

so we won’t find it).

Besides,

in practice right

now, ‘Nemesis’ is functioning as a name.

Not only is it not

the case that ontic questions must be settled for semantic

(referential) questions to be settled, it is the settling of

the latter that can determine the former.

The diachronic

acceptance of ‘Nemesis’ as a name of an object ushers in the

diachronic acceptance of

the individuation of

the object,

Nemesis.

This example also exhibits the significance of

coherence for reference.

As noted,

only with publically

verifiable corroborating evidence granting support for the

individuation of Nemesis will the name ‘Nemesis’ continue to

be used and accepted.

The claim that ‘Nemesis caused the

mass extinctions on the earth’ must cohere with our accepted

theories or we will not accept Nemesis as an object or

‘Nemesis’ as a name.

If

it

is decided that the claim that

‘Nemesis caused the mass extinctions on the earth’ does not

cohere with other (accepted) statements,

then it will also

be decided that reference was not made to Nemesis (despite

any descriptions associated by a given speaker or intentions

by a given speaker and despite any causal chains that might

possibly be established).

At present,

‘Nemesis’ is

a

theoretical term for

the scientific community, and whether

it

remains

so

or becomes an observational term or

is

discarded as non-referring depends finally on the coherence

of the claims within the scientific community and their

acceptance by the scientific community.

13

The determination

of reference is made by a public, verifiable process, not by

a given speaker, and, at least in this case, independent of

any causal connection.

The dubbing by Davis, Muller and Hut

of

‘Nemesis’

to

the

(theoretical)

object Nemesis

did not

establish any baptism or any causal chain to an object, nor

did

it establish a name

to

be used by

the community.

‘Nemesis’ became a name, not by

a dubbing (accompanied by

intentions or mental representations),

but by

the gradual

adoption by the scientific community of

a linguistic label

to an ‘object’ whose existence is

in question.

(This

clearly

relates

to

natural

kind

terms

and

issues

of

individuation and partitionings.

For example, see Kuhn

(1962)

on

the gradual acceptance of

reference

to oxygen.

The diachronic acceptance of oxygen as

a natural kind

correlates with

the diachronic adoption of

‘oxygen’

as

a

natural kind term.

The acceptance of both was determined by

the scientific community and was shaped by their coherence

with the theoretical frameworks of that community.

Again,

this sort of claim will be illustrated perhaps more clearly

in the AIDS case to be discussed below.)

This

Nemesis

example

illustrates

the socially

constructed character of reference (and of individuation).

The following AIDS example also illustrates this character,

however, where the Nemesis case deals with a presently-

299

 

considered theoretical individual, the AIDS case deals with

issues of partitionings and the establishment of causal

agents (both within a social context).12

AIDS (acquired immune deficiency syndrome) supposedly

first appeared in 1981.

Efforts to uncover a (the) cause

led to two independent announcements of the discovery of a

(the) cause.15

In the spring of 1983, French scientists at

the Institut Pasteur announced that they had discovered the

cause of AIDS.

The cause, they claimed, was a virus, which

they dubbed LAV (lymphadenopathy-associated virus).

Several

months (before:

this announcement,

scientists

reported

that

SAIDS (simian acquired immune deficiency syndrome) appeared

to be caused by a virus.

SAIDS, a disease held to be

similar to AIDS in symptoms and pathology, having a viral-

infectious agent, led researchers to look for a similar

cause for AIDS.

The viral organism responsible for AIDS,

according to the French researchers, was so-named because of

its initial discovery in a patient with the swollen lymph

glands and flu-like symptoms often associated with AIDS.

Almost a year later, in April 1984, American scientists

announced that they had discovered the cause of AIDS.

The

virus which

these

scientists

claimed was

responsible was

dubbed. HTLV-3 (human. lymphotrophic. retrovirus-3).

At the

public announcement of the American researchers’s discovery,

it was claimed that HTLV-3 was probably the same organism as

that

identified as

LAV,

but

genetic

analysis

and

immunological. comparisons would.

be necessary

to determine

300

u'

 fi'

this.

It was suggested that both HTLV-3 and LAV belong to a

class of retroviruses which use RNA rather than DNA as their

basic genetic material.

The American scientists dubbed the

culprit organism as HTLV-3 because they linked it to another

retrovirus called HTLV-1, which is associated with a rare

form of human leukemia (found primarily in Japan,

the

Caribbean and

parts

of Africa).

According to public

reports, the French researchers were not sure whether or not

the virus they construed as the causal agent was a member of

the HTLV family.

When the American announcement was made,

however, the French researchers claimed: "We are very happy

that another team has found the same thing" (Science N315

125(1984): 285).

"From its effect, and the descriptions of

it, it can only be the same one [virus]." (Ibid.)

LAV, they

said, was only a "provisional name."

Several interesting ‘events’ occurred over the course

of the year following the American announcement.

The first

is that the scientific community continued to use both ‘LAV’

and ‘HTLV-3’ to designate the virus which they investigated

and studied.

For example, in a noted science newsmagazine

(Science N313), fourteen articles appeared on AIDS between

July 7, 1984 and April 27, 1985.

In five of those articles,

both ‘LAV’ and ‘HTLV-3’ were cited as two names for the same

virus (probably).

In four articles HTLV-3 was cited as the

viral cause and LAV was not mentioned.

In no articles was

LAV cited as the cause and HTLV-3 not mentioned.

(In five

articles neither was cited as

the cause,

but in those

301

 

articles,

the cause of AIDS was not discussed.)16

No

official announcement has yet been issued that LAV and HTLV-

3 either are or are not the same virus.

Related to the debate over whether or not LAV and HTLV-

3 are the same virus, at least one researcher, Jay Levy (of

the University of California at San Francisco) even opted

for a

third name, ARV (AIDS-associated retrovirus).

Levy

and collegues isolated a virus from an AIDS victim and

cloned the entire genome while trying to develop a screening

test (and, potentially, a vaccine).

Reportedly, pending a

comparison of LAV and HTLV-3, Levy called his isolates ARV.

Another interesting ‘event’ associated with the naming

and

referring

issues

here

is

that,

following

the

identification of HTLV-3 (and of LAV), researchers have gone

back to look for a (the) historical cause for AIDS.

In

March 1985,

it

was reported that AIDS "or something very

like it" existed in Africa in

1972.

(As mentioned above,

originally AIDS was said to have emerged in 1981.)

Scientists thawed frozen blood samples from Uganda taken in

1972 and 1973.

They looked for the presence of antibodies

to HTLV-3, found it in two thirds of the samples, suggesting

that "the virus detected may have been a predecessor of

HTLV-3 or HTLV-3

itself

but existing in

a population

acclimated to its presence." (Science N313 127 (1985): 173).

What is significant about this

is

that, given the apparent

social identification of HTLV-3 as

the cause of AIDS,

scientists then looked for earlier instances of AIDS, and,

302

in effect,

located (or determined) a different historical

nexus for the disease.

(On April 20, 1985 it was reported

in Science News that Max Essex (of Harvard University) has

suggested that African green monkeys may harbor AIDS virus

and could be the source of AIDS in humans.

If this is so,

the historical nexus could again be altered.)

Now, with regard to the issues of naming and referring,

what is

the significance of

the AIDS example?

The

significance, I think, is that, as with the Nemesis example,

the Wittgensteinian view of naming and of reference gives a

better account of

this case

than does

the Searlean or

Kripkean view.

Several features

are worthy of attention

here.

First, as with the Nemesis case,

the identification

of

a natural kind (the cause of AIDS), was

(is) a social,

diachronic process.

Although two groups of researchers

dubbed some virus type as the cause of the disease, the name

that actually gets accepted and used is one that depends not

on the intentions of any particular speaker (a la Searle),

but on

the agreement of

the social community.

Nor is

the

name established (a la Kripke) by an appropriate causal

chain, initiated by a speaker’s perceptually causal hook-up

to an object (conjoined with the appropriate mental

representation) because much of

the debate itself concerns

just what

the causal hook-up connects.

Determining what

object is there to be causally hooked up with is what’s at

issue.

This

issue

will

be

settled over

time

by

the

scientific

community

and

in

so

doing,

the scientific

303

community will establish the proper reference, along with

the appropriate partitioning (into, minimally, the cause of

AIDS versus not the cause of AIDS).

The variability over

the past couple of years about the appropriate name for the

virus

indicates

that

the naming has not been settled and

that reference has not been determined because it hasn’t

been determined what is being referred to.

This seems to

suggest the Wittgensteinian view of reference as being more

appropriate here than the cluster or causal theories.

As

the Wittgensteinian view contends, reference seems to be a

matter of communal agreement,

not simply

in

the

‘linear’

manner associated with

the

causal

theory,

but

in

the

language-community consensus sense.

That is,

the communal

agreement is not just that once reference has been

established (a

la Kripke), then future reference is

parasitic upon

that initial naming,

but

that communal

agreement in the sense of the community as a whole (or,

given Putnam’s notion of linguistic division of labor,

the

relevant portion of the community) over time comes to adopt

a name for reference, as well as an object (and kind) for

reference.

Conclusion

I want

to conclude by restating two points mentioned

earlier in this chapter.

First, I have presented this

Wittgensteinian view of reference not as a fully worked-out

theory, but as

a tentative, better, view of reference.

It

captures, I think, the strengths of both the cluster and the

304

causal

theories,

without making the (questionable)

underlying commitments of each.

Furthermore, as I have

tried to show, it can account for the cases which were

troublesome for each of the received theories as well as the

two ‘real-life’ scientific cases

just presented.

The

second, final, point I want to make is that, at least in one

sense,

this Wittgensteinian view does not deny that there

might be cases which both the cluster and causal theories

can account for very well.

It could well be that in some

contexts

reference

can

be

determined

by

descriptions

associated with a name or by a causal chain.

However,

to

the extent that these accounts work,

I want to hold,

they

work

because

they

can

be

subsumed within

the

larger

Wittgensteinian view

and work only when

their underlying

commitments are abandoned.

To the extent that a cluster or

causal theory of reference can be separated from privacy and

realism (along with corollary notions of essentialism and

correspondence), such a theory might be appropriate and

adequate in given situations.

As general theories of

reference, however, they are inferior to the Wittgensteinian

(constructivist) view.

305

1Much of

the following discussion of Wittgenstein is

based on my (1978).

NOTES

2The entire paragraph 79 reads as follows:

Consider this example.

If one says ‘Moses did not

exist’, this may mean various things.

It may

mean: the Israelites did not have a single leader

when they withdrew from Egypt-or: their leader was

not called Moses-or: there cannot have been anyone

who accomplished all that the Bible relates of

Moses-or: etc. etc.-We may say, following Russell:

the name

‘Moses’ can be defined by means

of

various descriptions.

For example, as

‘the man

who

led

the

Israelites

through

the wilderness’,

‘the man who lived at that time and place and was

then called “Moses"’, ‘the man who as a child was

taken out of

the Nile by Pharaoh’s daughter’ and

so on.

And according as we assume one definition

or another the proposition ‘Moses did not exist’

acquires a different sense, and so does every

other proposition about Moses.-And if we are told

‘N did not exist’, we do ask: ‘What do you mean?

Do you want to say......or......etc.?’

But when I make a statement about Moses,—am I

always

ready to substitute some one

of

these

descriptions for ‘Moses’?

I shall p5?haps say: By

‘Moses’ I understand the man who did what the

Bible relates of Moses, or at any rate a good deal

of it.

But how much?

Have I decided how much

must be proved false for me

to give up

my

proposition as false?

Has the name ‘Moses’ got a

fixed and unequivocal use for me in all possible

cases?-Is it not the case that I have,

so

to

speak, a whole series of props in readiness, and

am ready to lean on one if another should be taken

from under me and vice versa?-Consider another

case.

When I say ‘N is dead’, then something like

the following may hold for the meaning of the name

‘N’:

I believe that a human being has lived, whom

I

(1) have seen in such-and-such places, who (2)

looked like

this

(pictures),

(3)

has

done such-

and-such things, and (4) bore the name ‘N’

in

social life.-Asked what I understand by ‘N’, I

should enumerate all or some of these points, and

different ones on different occasions.

So my

definition of

‘N’ would perhaps be ‘the man of

whom all this is true’.-But if some point now

proves false?-Shall I

be prepared to declare the

proposition ‘N

is dead’ false-even if

it

is only

something which strikes me as incidental that has

turned out false?

But where are the bounds of the

incidenta1?-If I had given a definition of the

306

.   

n
I

name in such a case, I should now be ready to

alter it.

And

this

can be expressed like this:

I use

the name ‘N’ without a fixed meaning.

(But that

detracts as little from it usefulness, as

it

detracts from that of

a table that it stands on

four

legs

instead of

three and

so sometimes

wobbles.)

Should it

be said that I am using a word

whose meaning I don’t know, and so am talking

nonsense?-Say what you choose, so long as it does

not prevent you from seeing the facts.

(And when

you see them there is

a good deal that you will

not say.)

(The fluctuation of scientific definitions:

what to-day counts as an observed concomitant of a

phenomenon will to-morrow be used to define it.)

3Wittgenstein’s immediate concern here is

to show that

it

is mistaken to suppose that there is

a general form of

propositions, something common to all that we call language.

That is,

it is mistaken to object to the notion of language

games because no essence of a language game is (can be)

given.

That issue is beyond the present concerns.

4I will say more about the notion of ‘open texture’,

but I want at this point to remark that this is precisely

one

of

the

issues

of

greatest

divergence

between

Wittgenstein’s and Searle’s views.

Where Searle admits the

sense and usefulness of speaking of

the logical sum or

inclusive disjunctive set of descriptions associated with a

name, Wittgenstein does not.

Rather Wittgenstein speaks of

the lack of any fixed meaning of

a name.

(Cf. note 2

above.)

It

is

worth noting

that Linsky

(1977)

also

recognizes this difference between Wittgenstein and Searle,

as do Ishiguro (1974), Carney (1979) and Cappio (1981).

Says Linsky:

[Wittgenstein and

Searle]

say

that

the

main

difficulty with [Russell’s]

theory is that it

ignores a certain looseness in our use of proper

names.

Proper names,

they

say,

are

not,

in

general, associated by their users with unique

descriptions as Russell’s theory requires.

Here

Searle and Wittgenstein part company.

Searle

thinks that in place of the unique description

required by Russell there are a cluster of

descriptions which the users of

a name associate

with it.

He thinks that it is a necessary truth

that

some

"sufficiently

large"

(but

indefinite)

number of

the

open sentences

of

which

these

descriptions are built are uniquely satisfied by

the referent of

the

name.

Wittgenstein’s

correction

to

Russell’s

theory

is

rather

different.

He

thinks

proper

names

are

307

 

 

characteristically used without any fixed meaning;

there

is neither a unique description nor

a

cluster of

them which fixes the sense of our

names.

We use these names without a fixed sense.

(1977, p.93).

5This sentence is presented by ‘the antagonist’, i.e.

someone objecting to Wittgenstein’s claims.

(Note the

similarity between this sentence and Searle’s disjunctive

set of descriptions associated with a name.)

6This insistence that we often use names without fixed

meaning is tied in with Wittgenstein’s notions of language

games and rule-following, and his rejection of

the

distinction between the contexts of discovery and contexts

of justification.

This

is not to say that no names have

fixed meanings.

We can (and do) fix the meaning of some

names (or, at least, some words, such as ‘bachelor’ and

‘mule’).

But it is mistaken to think that we always can (or

do).

What is also important in connection with this fixing

of meaning when it does occur is that it is done as part of

an activity (within a language game), for certain purposes.

(In this regard, see Linsky (1977, pp.104-105).)

In other

words, a word is a name only when it is used as one (as

opposed to being used as, say, a demonstrative), and to be

used as

a name means that we employ it in calling for,

asking about, saying something about, etc., the object being

referred to.

Cf. Canfield (1977), where he considers

Kripke’ 3 case in which a speaker mistakenly says that Jones

is raking leaves (when actually Smith is):

"Is

it correct to say that either he’ d attribute

the predicate [‘ is raking leaves’ ]

to Jones or to

Smith depending on how we take the question.

How

we take the question is something we have to learn

to do.

We can take the question in a way that

will give only the results that Donnellan intends.

But then the right thing to say is that it is the

series of examples he gives that teaches us how to

take

the question.

In Wittgensteinian terms

Donnellan teaches us a certain language game, one

of establishing the speech act referent by

answering a certain question in

a certain way."

(1977, pp.108-109).

(Also,

see Ishiguro (1969),

p.25.)

7Wittgenstein’s remarks here sound very much like

Searle’s distinction between the primary aspect and

secondary aspect under which reference is made.

However,

Wittgenstein’s view is not the same as Searle’s because, as

argued in chapter four,

Searle’s view

is consistent with

private language, whereas Wittgenstein’s view is

not.

As

Ishiguro (1969) says:

Throughout his later works,

Wittgenstein

discusses the temptation to look in the wrong

308

direction for a mark which would tell us why a

word refers to a certain object.

We look for

mental processes that go on as we utter the words

rather than the rule governing the use of the

words that we can come to grasp. (1969, p.23).

As noted previously, Searle is guilty of looking for mental

processes (brain states) which would tell us why a word

refers

to

a certain object.

His primary aspect-secondary

aspect distinction captures

in form Wittgenstein’s point,

but within the context of Searle’s

theory,

only

the form

(and not the substance) is captured.

8There are cases, for Searle, in which the descriptions

which reflect the primary aspect under which reference is

made can nonetheless not be true of the object to which

reference is made (e.g. referring to

a usurper as ‘the

king’).

gone might argue that in this case the name ‘deel’ is

not used,

but

is only mentioned.

Consequently, reference

isn’t made for the causal theory, either, even though there

is

a causal connection between my neighbor’s utterance of

the name and the object.

However,

it

is not clear to me

that such a move could legitimately be made by a causal

theorist.

Why,

for example,

should reference not be made

(to deel) in

the following case:

"‘Gddel’ is

the name of

the man who proved the incompleteness of arithmetic."?

If

the use/mention distinction is relevant for the causal

theory in giving an account for reference, then it is not at

all clear to

me

how

the causal theory can retain its

uniqueness as a theory of reference.

The only account I can

foresee would be such that

the

theory would have to

take

into account the pragmatic (Wittgensteinian) aspects of

reference.

A full fleshing out of my claims here is beyond

the present point.

(The point of raising this issue at all

is that given a ‘pure’ causal theory of reference, it seems

that my neighbor has referred to deel, since one could

demonstrate an appropriate causal connection.)

10Clearly, one feature of this Wittgensteinian view of

reference is that reference to fictional entities is not an

important issue, nor are negative existential statements a

problem.

These sorts of cases can be treated in

the same

way as ‘normal’ cases because the ontological status of the

‘objects’ in question is not crucial.

Cf. Elgin (1983) for

a sustained treatment of reference to various sorts of

‘objects’ and various kinds of modes of reference (such as

reference to fictional entities and metaphorical reference).

lle. Fleck (1979) for an account of the socially-

determined ‘fact’ of the cause and nature of syphilis.

12This last sentence, of course,

is

the sort of claim

that I find somewhat unpalatable (i.e. Nemesis really caused

309

  

those extinctions, but we’ll never find it because it has

left its orbit).

Without some sort of verifiable evidence

to support this claim, it serves no function in our set of

truths and fails to cohere with other (accepted) statements.

It ranks,

with me, with statements about the Absolute or

ghosts, which are there, but we’ll never detect them.

13I am certainly not intending to suggest by this claim

that I hold to a theoretical/observational distinction (at

least a sharp one).

The distinction, I think, is untenable.

I am simply suggesting that

the status of

the claim that

‘Nemesis caused the mass extinctions on the earth’ changes

based on the criterion of coherence.

14Another example of

the social nature of naming, at

least within a scientific context,

is

illustrated by

the

following item from a science newsmagazine:

Physicists are currently pondering the ‘Proposal

for the Systematic Naming of Hadrons’, yet another

attempt to agree on names for

the growing

population of subatomic particles.

It’s a tedious

chore, and it has physicists recalling the lament

attributed to Enrico Fermi

in

the early 1950s -

back when particles were

just starting to

proliferate.

"If

I could remember all

those

names,"

said

Fermi,

"I would have

been

a

botanist." (Science 85 6(1985): 9).

15As will be noted below, one point at issue is whether

or not AIDS has one cause or more than one, hence my use of

‘a (the)’.

16Of course, the fact that Science News is an American

publication could explain why HTLV-3, and not LAV, was more

often cited as the cause.

I am unacquainted with the French

literature and, consequently, am hesitant to claim too much

here about

the

sociology of naming in

the scientific

community.

However,

my main points about the social and

diachronic character of naming and referring does not rest

on the French reports.

CONCLUSION

A journey of a thousand miles begins with a single step.

(Lao Tzu, Tao Te Ching)

311

 

CONCLUSION

The overriding thesis

of

this

long

(and,

for the

reader,

torturous)

book

has

been

the

proposal

of

constructivism as

an alternative to what

is

today called

realism.

In

a statement, I have been arguing for a view

which asserts that reality is socially constructed.

I have

argued, through the vehicle of the issue of reference, that

epistemic issues underlie ontic and semantic issues,

that

theories

of

reference

rest

on

(and

are committed to)

conceptions of individuation and similarity, and that these

conceptions of

individuation and similarity are dependent

upon our theories, languages and cognitive interests.

This

thesis is by no means new; parts of

it (at least) are found

in Kuhn, Goodman, Wittgenstein, Quine,

Marx,

the American

pragmatists and many others.

What I think is new is

the

sustained analysis of

a constructivist position through an

analysis of

the specific contemporary theories of reference

and their epistemic commitments.

(Even parts of

this are

preceded

by

the

work

of

Goodman,

Elgin

and

Putnam.)

Specifically,

I have argued that the received theories of

reference -

the cluster theory associated with Searle and

the causal theory associated with Kripke and Devitt - are

more alike than different and that they share the same

fundamental epistemic commitments.

They share particular

312

 

conceptions of individuation and similarity, and subsequent

conceptions of

(and commitments to) privacy, essentialism,

correspondence and realism.

Each of these conceptions and

commitments, I have argued, is mistaken.

Consequently, the

theories

of

reference

reflected

by

these conceptions and

commitments are mistaken as well.

I then suggested the

outline and beginnings of an alternative view of reference,

a view based on constructivist conceptions and commitments.

Throughout this study, then, I have suggested not only that

theories of reference must be looked at in light of their

underlying commitments, but also (the other side of the

coin)

that particular epistemic commitments and positions

are

borne

out

in,

for example,

theories

or

views

of

reference.

To

fully understand a

theory

or

view

of

reference,

then,

we must look

at

its commitments and

underlying

conceptions;

to

fully

understand

a

particular

epistemic position, we must look at its manifestations, e.g.

in

the

view(s)

of

reference

that

follow

from

it.

Constructivism as an overarching and guiding position (along

with its corollaries, coherence and verificationism) is,

I

think,

the most plausible epistemic view,

though one which

needs to be explored and fleshed out in much greater detail.

This book, I hope, is one small contribution to that effort.

313

  

BIBLIOGRAPHY

 

 

BIBLIOGRAPHY

Armstrong,

G.E.M.

Intention.

Oxford:

Oxford University

Press, 1957.

Barnes,

Barry.

T.S.

Kuhn and Social Science.

New York:

Blanshard, Brand.

The Nature of Thought.

London: Allen and

Unwin, 1939.

Boer, Steven E.

"On Searle’s Analysis of Reference."

Analysis 32 (1972): 154-159.

Boersema,

David

B.

"Wittgenstein on

Names."

(Unpublished

manuscript.)

Brody,

Baruch

A.

Identity_ and

 

Essence.

Princeton:

Princeton University Press, 1980.

Canfield, John V.

"Donnellan’s Theory of Names."

Dialogue

16 (1977): 104-127.

Cappio,

James.

"Wittgenstein

on

Proper

Names."

Philosophical Studies 39 (1981): 87-105. 

Carney, James D.

“Wittgenstein’s Theory of Names.

Australasian Journal of Philosophy 57 (1979): 59-68. 

Cartwright, Nancy.

How the Laws of Physics Lie.

Oxford:

Claredon Press, 1983.

Cartwright,

Richard.

"Some Remarks on Essentialism."

Journal 3: Philosophy 65 (1968): 615-627. 

Davidson, Donald. "The Individuation of Events."

In Essays

in Honor

of

   

Carl

G.

Hempel.

Edited by Nicholas

REscher.

Ddfdrecht: D. Reidel, 1970.

Pages 216-234.

Devitt, Michael.

Designation.

New York: Columbia

University Press, 1981.

   

Realism and Truth.

.

Princeton: Princeton

University Press, 1984.

Dewey, John.

Reconstruction in Philosophy.

 

New York:

Mentor, 1950.

Donnellan,

Keith.

"Reference and Definite Descriptions."

Philosophical Review 75 (1966): 281-304. 

 

"Proper

Names

and

Identifying

Descriptions."

In Semantics of Natural Language.

 

Edited

by

Donald Davidson a

d Gilbert Harman.

314

 

Dordrecht: D. Reidel, 1972.

Pages 356-379.

Donnellan,

Keith.

"Speaking of

Nothing."

Philosophical

Review 83 (1974): 3-31.

Duhem, Pierre.

The Aim and Structure of Physical Theory.

New York: Atheneum, 1962.

Dummett, Michael.

Truth and Other Enigmas.

 

Cambridge:

Harvard University Press, 1978.

Dupre,

J.

"Natural

Kinds

and

Biological

Taxa."

Philosophical Review 90 (1981): 66-90.

Elgin, Catherine Z.

"Lawlikeness and the End of Science."

Philosophy of Science 47 (1980): 56-68.

 

Indianapolis: Hackett Publishing Co., 1983.

.

With

Reference

£_

Reference.

Evans, Gareth.

"The Causal Theory of Names."

Aristotelian

Society Supplementary Volume 47 (1973): 187-208.

Fleck, Ludwig.

Genesis and Development of a Scientific

Fact.

Chicago: University of Chicago Press, 1979.

Flew, Anthony.

A Dictionary pf Philosophy.

New York: St.

Martin’s Press, 1979.

Ghiselin, M.

"A Radical Solution to the Species Problem."

Systematic Zoology 23 (1974): 536-544. 

Gill,

J.H.

"Wittgenstein’s

Concept

of

Truth."

International Philosophical Quarterly 6 (1966): 71-80.

Goodman,

Nelson.

Fact,

Fiction,

and Forecast

(2nd

ed.).

Indianapolis: Hobbs-Merrill, 1965.

Bobbs-Merrill, 1972.

.

Problems

and

Projects.

Indianapolis:

Hackett Publishing Co., 1978.

.

Ways _£ Worldmaking.

Indianapolis:

Gould, Stephan J.

The Mismeasure of Man.

New York: W.W.

Norton & Co., 1981.

W.W. Norton & Co., 1983.

Hen’s Teeth and Horse’s Toes.

New York:

Haack, Susan.

Philosophy of Logics.

Cambridge: Cambridge

University Press, 1978.

315

  

 

Hacking,

Ian.

Representing and

 

Intervening.

Cambridge:

Cambridge University Press, 1983.

Hanna, Joseph F.

"Explanation, Prediction, Description, and

Information Theory."

Synthese 20 (1969): 308‘344.

of Explanatory Powerfl'

Philosophy 23 Science 45

.

"On Transmitted Information as

a Measure

(1978): 531-562.

Harding, Sandra.

"The Norms of Social Inquiry and Masculine

Experience."

In PSA 1980, Volume 2.

Edited by P.D.

Asquith and R.N.

Giere.

East Lansing: Philosophy of

Science Association, 1981.

Pages 305-324.

Hesse,

Mary.

The

Structure 3; Scientific

Inference.

 

London: Macmillan, 1974.

Revolutions and Reconstructions in the

 

 

Philosophy 33 Science.

Brighton: Harvester Press,

1980.

.

"Comment on Kuhn’s

‘Commensurability,

Comparability, Communicability."

In PSA 1982, Volume

   

2.

Edited by P.D. Asquith and T. Nickles.

East

Lansing:

Philosophy of Science Association, 1983.

Pages 704-711.

Hirsch, Eli.

"Essence and Identity."

In Identity and

Individuation.

Edited by Milton K. Munitz.

New York:

New York University Press, 1971.

Pages 31-49.

Holsinger, Kent E.

"The Nature of Biological Species."

Philosophy 2: Science 51 (1984): 293-307.

Hull, David.

"The Effect of Essentialism on Taxonomy."

British Journal for

the Philosophy of Science

15

(1965):

314-326.

.

"A Matter of Individuality."

Philosophy pf

Science 45 (1978): 335-360.

Ingber, Warren.

"The Descriptional View of Referring: Its

Problems and Prospects."

Journal pf Philosophy 76

 

(1979): 725-738.

Ishiguro, Hide.

"Use and Reference of Names."

In Studies

in the Philosophy of Wittgenstein.

Edited by Peter

 

Winch.

London: Routledge & Kegan Paul,

1969.

Pages

20-50.

Philosophy 23 Language." Philosophy 49 (1974): 438-442.

.

"Review of Michael Dummett’s Frege:

316

   

James, William.

The Writings of William James.
 
 

Edited by

John J. McDermott.

Chicago: University of Chicago

Press, 1977.

Jones, O.R. (ed.)

The Private Language Argument.

New York:

Macmillan, 1971.

Kitcher, Phi lip

"Species."

Philosophy of Science 51

(1984): 308p-J33

Kitts, 0.8. and Kitts, D.J.

"Biological Species as Natural

Kinds."

Philosophy of Science 46 (1979): 613-622.

Kripke, Saul.

"Outline of a Theory of Truth."

Journal of

Philosophy 72 (1975): 690-716.

Referencefl'

Midwest Studies in Philosophy 2

(1977):

.

"Speaker’s

Reference

and

Semantic

255-276.

 

University Press, 1980.

Naming and Necessity.

Cambridge: Harvard

Cambridge: Harvard University Press, 1982.

Kuhn, Thomas S.

The Structure of Scientific Revolutions.

 

Chicago: University of Chicago Press, 1962.

Structure of Scientific Theories.

Edited by Frederick

.

"Second Thoughts on Paradigms."

In The
 

Suppe.

Urbana:

University of

Illinois

Press,

1977.

Pages 459-482.

"Commensurability, Comparability,

Communicability.

In PSA 1982, Volume 2.

 

Edited by

P. D. Asquith and T. Nickles.

East Lansing: Philosophy

of Science Association, 1983.

Pages 669-688.

Leonard, Henry and Goodman, Nelson.

"The Calculus of

Individuals."

Journal of Symbolic Logic 5 (1940): 44-

55.

Leplin, Jarrett (ed.)

Scientific Realism.

Berkeley:

University of California Press, 1984.

Linsky, Leonard.

Names and Descriptions.

Chicago:

University of Chicago Press, 1977

Luntley,

Michael.

"Verification,

Perception,

and

Theoretical Entities."

Philosophical Quarterly 32

 

(1982): 245-261.

317

Mayr,

Ernst.

Systematics and the Origin of Species.

New

York: Columbia University Press, 1942.

McKinsey, Michael.

"Kripke’s Objections to Description

Theories of Names."

Canadian Journal of Philosophy 8

(1978): 485-497.

McMullin, Ernan.

"Values in Science."

In PSA 1982, Volume

2.

Edited by P.D. Asquith and T. Nirkles.

East

Lansing:

Philosophy of Science Association,

1983.

Pages 3-28.

Nagel,

Ernest.

The Structure of Science.

New York:

Hartcourt, Brace & World, 1961.

Paulos, John Allen.

I Think, Therefore I Laugh.

New York:

Columbia University Press, 1985.

Peirce,

Charles S.

Philosophical Writings of Peirce.

Edited by Justus Buchler.

New York: Dover, 1955.

Plantinga, Alvin.

The Nature of Necessity.

Oxford: Oxford

University Press, 1974.

Popper, Karl R.

Conjectures and Refutations: The Growth of

Scientific Knowledge.

New York: Harper & Row, 1963.

Putnam, Hilary.

Reason, Truth and History.

Cambridge:

Cambridge University Press, 1981.

Quine, W.V.O.

From A Logical Point _£ View.

 

Cambridge:

Harvard University Press, 1953

1960.

Word and Object.

Cambridge: M.I.T. Press,

New York: Columbia University Press, 1969.

Ontological

Relativity

and

Other

Essays.

.

Methods of Logic (3rd ed.)

New York: Holt,

Rinehart and Winston, 1972.

.

"Whither Physical Objects?"

In Essays in

Memory of Imre Lakatos.

 

Edited by R.S. Cohen et al.

Dordrecht: D. Reidel, 1976.

Pages 497-504.

.

"States of Mind."

Journal of Philosophy 82

219855: 5—8.

Rescher, Nicholas.

The Coherence Theory of Truth.

Oxford:

Oxford University Press, 1973.

Rhees,

Rush.

"Can There

Be

A Private Language?"

Proceedings of the Aristotelian Society Supplement 28

 

318

 

(1954): 77-94.

Rothbart, Daniel.

"The Semantics of Metaphor and the

Structure of Sciencefl'

Philosophy of Science 51

(1984): 595-615.

Russell, Bertrand.

“On Denoting."

Mind 14 (1905): 479-493.

Henry Holt & Co., 1912.

The Problems of Philosophy.

New York:

Salmon, Nathan U.

Reference and Essence.

 

Princeton:

Princeton University Press, 1981.

Schwarz, Stephan P.

(ed.)

Naming, Necessity, and Natural

Kinds.

Ithaca: Cornell University Press, 1977.

 

Searle, John R.

"Proper Names."

Mind 67 (1958): 166-173.

University Press, 1969.

.

Speech Acts.

Cambridge: Cambridge

Cambridge University Press, 1979.

.

Expression

and
 

Meaning.

Cambridge:

University Press, 1983.

.

Intentionality.

Cambridge: Cambridge

Harvard University Press, 1984.

.

Minds, Brains and Science.

Cambridge:

Slote, Michael A.

Metaphysics and Essence.

New York: New

York University Press, 1975.

Steinman,

Robert.

"Naming and

Evidence."

Philosophical

Studies 41 (1982): 179-192.

Strawson, P.F.

Individuals.

London: Methuen & Co., 1959.

Tarski,

Alfred.

"The Semantic Conception of Truth."

Philosophy and Phenomenological Research 4 (1944): 341-

3

Languages."

In Logic, Semantics, Metamathematics.

Oxford: Clarendon Press, 1956.

Pages 152-278.

"The Concept of Truth in Formalized

Thomson, Sir George.

"Some Thoughts on the Scientific

Method."

In Boston Studies

in

the Philosophy of

Science, Volume 2:

In Honor of Philipp Frank.

Edited

by Robert S.

Cohen and Marx W. Wartofsky.

New York:

Humanities Press, 1965.

Pages 81-92.

Waismann,

Freidrich.

"Verifiability."

In Logic and

319

 

 

 

Language.

Edited by'.A.G.N..Flew.

Garden City:

Doubleday Anchor Books, 1965.

Pages 122-151.

White, Alan R.

Truth.

Garden City: Doubleday Anchor Books,

1970.

Whorf, Benjamin Lee.

Language, Thought 3 Reality.

Cambridge: M.I.T. Press, 1956.

Wittgenstein, Ludwig.

Tractatus Logico-Philosophicus.

   

London: Routledge & Kegan Paul, 1922.

 

York: Macmillan, 1953.

.

Philosophical Investigations.

New

   

The Blue and Brown Books.

Oxford:

Blackwell, 1958.

 

Zettel.

Oxford: Blackwell, 1967.

 

Experiences’ and ‘Sense Data’."

Philosophical Review

.

"Notes

for Lectures

on

‘Private

77 (1968): 275-320.

 

 

1969.

92 Certainty.

Oxford:

Blackwell,

 

1932-1935.

Edited tur Alice Ambrose.

Totawa: Rowman

Wittgenstein’s Lectures, Cambridge,

and Littlefield, 1979.

320

 
 

 

  

 

"ImmmmES