The Tensed Theory of Time WilliamLane Craig

8/20/2019 The Tensed Theory of Time WilliamLane Craig

1/298

THE TENSED THEORY OF TIME


2/298

SYNTHESE LIBRARY

STUDIES IN EPISTEMOLOGY,

LOGIC, METHODOLOGY, AND PHILOSOPHY OF SCIENCE

Managing Editor:

JAAKKO HINTIKKA, Boston University

Editors:

DIRK VAN DALEN, University of Utrecht, The Netherlands

DONALD DAVIDSON, University of California, Berkeley

THEa

A.F.

KUIPERS,

University of Groningen, The Netherlands

PATRICK SUPPES, Stanford University, California

JAN

waLEN-sKI,

Jagiellonian University, Krakow, Poland

VOLUME

293


3/298

THE TENSED THEORY

OF TIME

A Critical

Examination

by

WILLIAM LANE CRAIG

Talbot School

of

Theology,

Marietta,

GA,

U.S.A.

SPRINGER-SCIENCE+BUSINESS MEDIA, B.V.


4/298

Library of Congress Cataloging-in-Publication Data

Craig, William Lane.

The tensed theory of time : a critical examination / by William Lane Craig.

p. cm. -- (Synthese library; v. 293)

Includes bibliographical references and index.

1.

Time--Philosophy. I. Title. II. Series.

BD638 .C73 2000

115--dc21

ISBN 978-90-481-5585-9

Printed on acid-free paper

All Rights Reserved

00-064723

© 2000 Springer Science+Business Media Dordrecht

Originally published by Kluwer Academic Publishers in 2000

Softcover reprint of he hardcover 1st edition 2000

No part

of

the material protected by this copyright notice may

be

reproduced

or

utilized in any form

or

by any means, electronic or mechanical,

including photocopying, recording

or

by any information storage and

retrieval system, without written permission from the copyright owner.

ISBN 978-90-481-5585-9 ISBN 978-94-015-9345-8 (eBook)

DOI 10.1007/978-94-015-9345-8


5/298

To

ALVIN PLANTINGA

who by his work and his life

has pointed the way


6/298

TABLE OF CONTENTS

Preface

1X

PART

I.

ARGUMENTS

FORAN

A-THEORY OF TIME

Section 1: The 1neliminability

of

Tense

Chapter 1 Introduction: Language, Tense, and Ontology

3

Chapter 2 The Old B-Theory of Language 23

Chapter 3 The New B-Theory of Language 66

Chapter 4 The B-Theory and Theories of Direct Reference 97

Section 2: The Experience

of

Tense

Chapter 5 Our Experience

of

Tense

131

PART II. ARGUMENTS AGAINST AN A-THEORY OF TIME

Chapter 6 McTaggart's Paradox

169

Chapter 7 The Myth of Temporal Passage

218

Bibliography

259

Subject Index

279

Proper Name Index

283

vii


7/298

PREFACE

T

he present book and its companion volume

The Tenseless Theory of

Time:

a

Critical Examination

are an attempt to adjudicate what one recent discussant

has called "the most fundamental question in the philosophy of time," namely,

"whether a static or a dynamic conception

of

the world

is

correct."

I had originally intended to treat this question in the space

of

a single volume;

but the study swelled into two. I found that an adequate appraisal

of

these two

competing theories

of

time requires a wide-ranging discussion

of

issues in

metaphysics, philosophy of language, phenomenology, philosophy

of

science,

philosophy

of

space and time, and even philosophy

of

religion, and that this simply

could not be done in one volume.

If

these volumes succeed in making a

contribution to the debate, it will be precisely because of the synoptic nature

of

the

discussion therein. Too often the question of the nature of time has been

prematurely answered by some philosopher or physicist simply because he is largely

ignorant

of

relevant discussions outside his chosen field

of

expertise. In these two

complementary but independent volumes I have attempted to appraise what I take to

be the most important arguments drawn from a variety of fields for and against each

theory of time.

The two rival theories of time which are the subject of our examination have

been known under a variety

of

names: the A- versus the B-Theory, the tensed versus

the tenseless theory, the dynamic versus the static theory. None of these labels

is

wholly adequate. The terminology of A- and B-Theory has the advantage of being

the traditional designations inspired by

J. M.

E. McTaggart; but these names are

descriptively opaque.

D.

H.

Mellor changed the vocabulary

of

the debate by

speaking instead of tensed and tense less theories, but

he

has now reverted to

speaking of the A- and B-Theories because his labels aroused confusion in the

minds

of

many concerning tense as an ontological category and tense as a linguistic

phenomenon. Michael Tooley prefers to speak of dynamic versus static theories,

but this terminology, too, can be misleading, since the vast majority

of

A-theorists

do not think of time as literally moving. In these volumes, I use such labels

interchangeably but have for the most part stuck with the traditional A and B

terminology.

I have spoken

of "the" tensed or tenseless theory of time, but this expression

is

purely stylistic.

As

we shall see, there is actually a family

of

A-Theories of time,

and B-theorists, too, differ among themselves on certain key issues. I shall argue

that many

of

these versions

of

the A- or B-Theory are,

in

fact, inconsistent and that a

unique A- or B-theoretical paradigm exists; but I should not want to be thought to

prejudice the issue in advance by my choice of words.

I am intellectually indebted in this study to too many persons to recall by name;

but I should like to acknowledge my special gratitude to Quentin Smith, from whom

Michael Tooley, Time, Tense,

and

Causation (Oxford: Clarendon Press, 1997), p. 13.

IX


8/298

x

I have learned a great deal about language and time, and to the late Simon

J.

Prokhovnik, the eminent Australian physicist, who helped me to see the wisdom of

Lorentz. I should also like to thank The University

of

Chicago Press and Wesleyan

University Press for permission to reproduce figures.

I am indebted to my wife Jan for her faithful labor in production of the typescript

and to my research assistants Ryan Takenaga, Mike Austin, and Narcis Brasov. I

should also like to thank Edward White and the Day Foundation for their generous

grant which helped to fund the production

of

the camera-ready copy and to Mark

Jensen and Jennifer Jensen for meticulously bringing this book into its final form.

Atlanta, Georgia William Lane Craig


9/298

PART I

ARGUMENTS

FOR

AN A-THEORY OF TIME


10/298

SECTION 1: THE INELIMINABILITY OF TENSE

CHAPTER 1

INTRODUCTION:

LANGUAGE, TENSE, AND ONTOLOGY

T he main dispute in the philosophy of time," writes one recent combatant, "is

about the status of the present."l Is the present an objective, independent

feature of reality or is it merely a subjective feature of consciousness or, at best, a

purely relational feature

of

events? Is time characterized by objective tense

determinations like pastness, presentness, and futurity, or are the moments

of

time

ordered only by tense less relations like earlier than, simultaneous with, and later

than?

Since

J. M.

E. McTaggart fIrst distinguished clearly between these two kinds

of time, labeling them the A- and B-series respectively, philosophers of time have

found it useful to adopt McTaggart's nomenclature, referring to theories of tensed

time as A-Theories and theories of tenseless time as B-Theories. One of the most

hotly contested issues in the struggle between A-theorists and B-theorists concerns

the alleged ineliminability of tense from language or thought and the implications

which this has for the nature of time. Accordingly in this section we shall consider

what implications linguistic tense has for an adequate ontology

of

time.

TENSED AND TENSELESS SENTENCES

By "tense," one means in the fIrst instance that familiar aspect of language which

serves to express something's pastness, presentness, futurity, or combination thereof

(as in the future perfect tense), all of which fall under what McTaggart called A

determinations, in contrast to tenseless B-relations like

earlier than, simultaneous

with, and later than. In English, tense is usually expressed by altering the form of

the verb (for example, "I write," "I wrote," "I shall write"),2 but tense may also be

expressed by a rich variety

of

adverbial phrases (for example, "now," "yesterday,"

"three days ago," "soon"), adjectives (for example, "past," "present," "future"),

prepositional phrases (for example, "at present," "in yet two days' time," "by next

1. Butterfield, "Spatial and Temporal Parts,"

Philosophical Quarterly

35 (1985): 32.

For an account

of

English grammatical tense, see Otto Jespersen,

Essentials

of

English Grammar

(London: Allen

&

Unwin, 1933). A very helpful discussion

of

the range

of

verbal tenses, some

of

which

are not found in English, is provided by Hans Reichenbach,

Elements

of

Symbolic Logic

(London:

Collier-Macmillan, 1947), pp. 288-297. The A-theorist would want to replace Reichenbach's "point of

speech" with something like "the present moment" but would otherwise find his nine fundamental forms

of verbal tense quite illuminating.

3


11/298

4

CHAPTER 1

Saturday"), and nouns (as in, for example, "Today

is

Saturday," "Now

is

when he

leaves"). In certain languages, tense is not a feature of verbs at all, other devices

being exclusively used to show the tense

of

a sentence.

3

But, as Gorman and

Wessman point out, "all four thousand

or

so known languages enable their speakers

to designate temporal relationships and to distinguish between past, present, and

future

events-though

with varying degrees

of

difficulty.,,4 Thus, as Mellor points

out, it makes no difference to the discussion

how

tenses are indicated; the point is

that a tensed sentence indicates "how near

or

far from the present, past, or future,

something is."s

In equating grammatical tense with that feature

of

language which expresses A

determinations, we imply either that sentences which do not express A

determinations are tense less

or

that no such sentences exist. Although in ordinary

language we typically assume that all our sentences are tensed, there are nonetheless

classes

of

sentences which are plausibly regarded as tenseless, for example:

(1)

mathematical or logical sentences like "2+2=4"; (2) sentences referring to types as

opposed to tokens, such as

"The

second movement of Dvorak's New World

Symphony utilizes American folk melodies"; (3) sentences expressing certain

universal generalizations like "All swans have necks"; (4) sentences expressing

certain dispositions, which

mayor

may not ever be realized, like "This glass breaks

easily"; (5) sentences about dated times

or

events related by B-determinations, such

as "The 1996 presidential election is earlier than the 2008 presidential election.,,6

The verbs in such sentences, while sharing the same linguistic syntactical form as

present-tense verbs, differ from them in their semantic content

in

that they convey

no tense.

7

It

is not incumbent upon the defender

of

tenseless sentences to maintain

that such sentences (or the propositions they express) are timelessly true (or false);

he may hold that true tenseless sentences are, in virtue

of

their tenselessness,

omnitemporally true. Thus, a true tenseless sentence may be presently true, but that

fact does not imply that the sentence is present-tensed. If a contingent, tenseless

sentence is presently true, it is logically impossible for it to be false at another

moment of time; but this is not the case for contingent, present-tensed sentences

which are presently true. A tenseless sentence, then, can imply a tensed sentence

(for example, "All swans have necks" implies that "Any swans which now exist

have necks").

It might be rejoined that while purportedly tense less sentences fail to have a

single tense, nonetheless they are in fact tensed because they each possess multiple

See Richard M. Gale, The

Language

of

Time,

International Library

of

Philosophy and Scientific

Method (London: Routledge, Kegan

&

Paul, 1968), pp. 45-46. For example, in Eskimo, the noun

puyok

(smoke) has a past-tense form puyuthluk (what has been smoke) and a future-tense form puyoqkak (what

will become smoke).

4 Bernard Gorman and Alden Wessman, "The Emergence of Human Awareness and Concepts of

Time," in

The

Personal Experience

of

Time,

Emotions, Personality, and Psychotherapy (New York:

Plenum Press, 1977),

pp.

44-45.

5 D. H. Mellor, Real Time (Cambridge: Cambridge University Press, 1981), p. 4.

See discussion in Yehoshua Bar-Hillel, "Indexical Expressions,"

Mind

63 (1954): 359-379; Nelson

Goodman,

The

Structure of Appearance, 2d ed. (New York: Bobbs-Merrill, 1966), pp. 366-368; Andros

Loizou, The

Reality of Time

(Brookfield, Ver.: Gower, 1986), pp. 9-10, 99-101.

7

For a discussion of this distinction, see Quentin Smith, Language and Time (New York: Oxford

University Press, 1993),

p.

189.


12/298

INTRODUCTION: LANGUAGE, TENSE, AND ONTOLOGY

5

tenses, either disjunctively or conjunctively.s The semantic content of a purportedly

tenseless verb will be either a disjunction of the past, present, and future tenses (for

example, "was, is, or will be") or else a conjunction

of

these same tenses (for

example, "was, is, and will be"). Thus, the meaning

of

a sentence like "The 1996

presidential election is earlier than the 2008 presidential election"

is

"The 1996

presidential election was, is, and will be earlier than the 2008 presidential election."

The meaning of "All swans have necks"

is

"All swans had, have, or will have

necks."

Now the defender of tenseless sentences might well agree that such disjunctive

or conjunctive sentences are entailed by the relevant tense less sentences; but do

such multiply tensed sentences really express what the tenseless sentences mean?

This seems dubious. Think

of

sentences employed in possible worlds semantics.

When it

is

said that "The modal operator

'D'

acts as a universal quantifier which

takes as its range the possible worlds in a sphere

of

accessibility," do we mean by

"takes" the conjunctive tense "took, takes, and will take"? That seems bizarre. If

we assert "Socrates is not wise in W*" are we really asserting "Socrates was, is, or

will be not wise in

W*"?

Are we not rather speaking tenselessly? Or if we say "In

the actual world God foreknows p, do we really mean "In the actual world God

foreknew, foreknows, or will foreknow p, or are we not rather deliberately

abstracting from considerations

of

tense in order to reflect upon a certain question?

Or consider sentences about event-types as opposed to their tokens: when I say "In

Shakespeare's Julius

Caesar

Brutus's speech precedes Antony's funeral oration," I

am talking about event-types which as such never occur. While the relevant event

tokens in a performance

of

the play can be rightly said to exist in temporal relations

with one another, must I, indeed, can

I,

in speaking

of

the event-types, properly

assert that the one either preceded, precedes, or will precede the other? Or what

about universal generalizations involving fictitious entities, such as "Centaurs have

the torso

of

a man on the body

of

a horse"? Is this really a conjunctively tensed

statement? It might be said that the meaning of this sentence

is

something like "If

Centaurs exist, existed, or will exist, then they have, had, or will have, etc." But is

the meaning so arcane? Do we not mean simply to abstract from any tense in such a

generalization? Or consider sentences about dispositions which are never realized,

like "This chemical substance kills instantly." The substance may never actually kill

anyone precisely because such a warning is believed.

It

might be said that this

sentence has the same meaning as the present-tense counterfactual,

"If

anyone were

to imbibe this chemical substance, it would kill him instantly." But while such a

counterfactual is entailed by this sentence, so is the indicative conditional, "If

anyone imbibes this chemical substance, it will kill him instantly." And these do not

have the same meaning. So how can we with confidence claim that either of these

sentences means the same as the purportedly tenseless sentence rather than

is

entailed by it? Or take sentences about B-determinations such as the one given

above. When it

is

asserted in 1998 that "The 1996 election is earlier than the 2008

election," the disjuncts "was" or "will be" of the disjunctive tense are clearly

inappropriate, since the relevant events are neither both past nor both future. But

See Smith, Language and Time, pp. 188-192; Roderick M. Chisholm and Dean W. Zimmerman,

"Theology and Tense,"

Nous 31

(1997): 262-265.


13/298

6

CHAPTER

1

how is the present-tense "is" any more appropriate in 1998, when the one event is

past and the other future? A tense less "is" seems more appropriate. These sorts of

examples supply

prima facie

evidence that there are tense less sentences.

It

certainly

seems open to us merely to stipulate that such sentences are to be taken as tenseless,

unless there is some overriding reason why a sentence must be tensed.

9

Smith attempts to provide such a reason. But the argument to which he repeatedly recurs in order to

show that purportedly tenseless sentences not merely entail, but are synonymous with, multiply tensed

ones is remarkably weak, indeed, counterproductive, viz., that the tenseless version and the multiply

tensed version are mutually substitutable

salva veritate

in belief contexts. Smith claims that if it is true

that "John believes that Socrates is wise," which reports his belief

in

a purportedly tenseless sentence,

then it is also true that "John believes that Socrates was, is, or will be wise" (Smith, Language and Time,

p. 191). Or again, with respect to sentences ascribing B-determinations to events, he claims that truth

value is preserved if

we

substitute for "John believes that Plato's birth

is

earlier than Aristotle's" the

multiply tensed sentence "John believes that Plato's birth

is

and always will be earlier than Aristotle's"

(Ibid.,

p.

198). Or again, with respect to universal generalizations like "All humans are under nine feet

tall," Smith asserts that this sentence is synonymous with

"Being human and not under nine feet tall

always was, is, and always will be unexemplified" because these are mutually substitutable in belief

contexts (Ibid., pp. 200-20 I).

But as recent literature concerning theories of Direct Reference (with which Smith

is

thoroughly

familiar) has reminded us, belief contexts provide very slippery ground for inferences concerning

entailments

in

such contexts. Poor, old John may never have even so much as entertained multiply tensed

statements, much less believe them. Maybe he thinks Socrates is (like Athena, the goddess of wisdom) a

mythological figure, who never did nor will exist. Maybe

he

is a die-hard defender oftenseless sentences

who refuses to believe that Plato's birth

is

now and always will be earlier than Aristotle's.

It

seems very

likely that Smith's unique and original analysis of material implication has never even occurred to John,

so that he might

well

believe that universal generalizations are tenseless, rather than multiply tensed

sentences. Since Smith holds that substitutability in belief contexts is a necessary condition of synonymy,

the failure of purportedly tenseless and multiply tensed sentences to be mutually substitutable

in

such

contexts shows that they

do

not mean the same thing. Thus, Smith's argument backfires.

In any case, Smith notes that mutual substitutability in belief contexts is only a necessary, not a

sufficient condition of synonymy.

He

attempts to close the gap by arguing that purportedly tenseless

sentences are semantically correlated with multiply tensed ones

in

such a way that their parts have the

same semantic content; for example,

in

"Socrates

is

wise" and "Socrates was, is, or will be wise," the

semantic content of "Socrates" and "wise" in the two sentences is identical,

as

is the semantic content of

"is" and "was,

is,

or will be" (Ibid., pp. 190-191). But that is precisely the question

in

dispute Smith just

assumes that

if

a semantic correlation

is

consistent with the undisputed semantic content

of

the two

sentences, then this counts as a confirming reason in favor of synonymy. But there is no reason to think

this assumption true, since non-synonymous sentences could be mutually entailing.

Smith does present another consideration in favor of his position, viz., that no timeless states of

affairs exist. This seems to be the principal argument for the meaning equivalence of typical

mathematico-Iogical sentences and their conjunctively tensed counterparts, since the belief context

argument would in this case be manifestly mistaken, as many philosophers do believe the tenseless

sentences but do not believe the multiply tensed versions. Smith argues that abstract objects like

numbers, properties, and so forth, are all temporal because they undergo relational change in being

referred to successively by temporal agents. Since they are temporal, the purportedly tenseless sentences

about such entities are really synonymous with conjunctively tensed sentences (Ibid., pp. 204-214).

Similarly, purportedly tenseless sentences ascribing B-determinations to events are synonymous with

conjunctively tensed sentences ascribing everlasting presentness to the events' state

of

being B-related.

Since there are no timeless states (as seen above), this state must be temporal. If two events are B

related, they are presently B-related, whether the events are past or present (Ibid.,

p.

197).

With regard to the argument against timeless abstract objects and states, I think one might be

justifiably skeptical whether the sort

of

"Cambridge change" envisioned by Smith is a sufficient condition

of

its subject's being temporal. The argument assumes without justification that

being referred to by x

is

a real property that a thing acquires or loses. But the defender of timeless entities could regard sentences

involving reference to such entities as true

de dicto,

not

de

reo Thus, "John believes that 2+2=4" means


14/298


7

It would be helpful to have some criterion for distinguishing between what are

undisputedly A-sentences and sentences which do not ostensibly ascribe A

determinations. In his classic treatment

The Language

of

Time,

Richard Gale draws

the following distinction between what he calls A-statements and B-statements:

Any statement which is not necessarily true (false)

is

an

A-statement if, and only if, it

is

made through the use

of

a sentence for which it

is

logically possible that two non

simultaneous uses

of

this sentence make statements differing in truth value, even if both

statements refer to the same things and the same places.

Any statement

is

a B-statement if, and only if, it describes a temporal relation between

events and

is

made through the use

of

a sentence for which it

is

the case that

if

it can

now be used to make a true (false) statement then any past or future use of this sentence

also makes a true (false) statement. 10

Gale's bifurcation would exclude from A-statement status a range of tensed

sentences;ll but so long as it captured a class

of

tensed sentences as expressing A

statements, the discussion could proceed. Unfortunately, although Gale is,

of

course, free to defme his terms

in

whatever manner he

wi.shes,

his definitions

involve controversial assumptions such that some A-theorists might be forced to say

that there are no such things as A-statements according to these definitions. This

would be the case, for example, for any A-theorist who holds to a non-propositional

theory

of

belief, according to which there are no propositional objects

of

belief, their

role being taken by the self-ascription

of

properties; or again, for any A-theorist who

holds that the propositional content

of

a tensed sentence is tense less, the tense being

contributed by non-propositional factors. Since these views merit discussion, it

would be unwise to preclude them from the outset.

More recently, Quentin Smith, whose Language and Time is destined to replace

Gale's book as the standard work in this area, has tried to layout the defining

conditions for

A-

and B-sentences

in

the following way:

12

A sentence

is

an A-sentence iff:

i.

It contains a future, present, or past-tensed copula and/or verb; it may also contain

a temporal indexical, such as a temporal adverb or pronoun.

that John believes the tenselessly true sentence "2+2=4", but does not mean that he believes of2 and of4

that 2 added to itself

is

4. If his beliefs are not de re, then 2 and 4 do not acquire and lose the property

being referred

to

by John. Hence, 2 and 4 could exist timelessly despite the fact that John makes

reference to them at various times.

But let that pass. The salient shortcoming

of

Smith's case

is

that he fails to justify the underlying

presupposition that tenseless truths cannot refer to temporal entities. That this assumption

is

moot is

evident from the fact that sentences ascribing B-determinations to temporal events are purportedly

tenseless. How then does the demonstration that all entities and states are temporal go to show that there

are no tenseless sentences about such things? Even if the tenseless sentences entail their multiply tensed

counterparts, how does that prove that they are synonymous, that the former are not in fact tense less?

Once again, Smith

is

forced to recur to his arguments about belief contexts and semantic correlations,

which are in the case now under consideration manifestly inadequate. It seems to me, therefore, that

Smith has failed to show that there are no tenseless sentences.

lO Gale, Language ofTime, pp. 42, 51.

I I

E.g.,

mUltiply tensed sentences, sentences containing the adverb "always," tensed sentences which

are broadly logically necessary, etc. The definitions also exclude a range

of

tenseless sentences from

expressing B-statements, since only statements describing temporal relations between events are allowed.

12

Smith, Language and

Time,

pp. 6-7.


15/298


16/298


9

Analogously, the future-tense version of this same sentence is to be analyzed

as

It

will be the case that John catches the 8:15 train to London," symbolized

Fp.

Since

the operand sentence remains

in

the present tense, the present-tense operator

is

vacuous and redundant. Like modal operators, to which they bear a close

resemblance, tense operators can be iterated to form ever more complex tenses. For

example, on Prior's analysis the future-perfect sentence "John will have caught the

8: 15 train to London" is to be analyzed as It will be the case that it was the case

that John catches the

8:

15 train to London" and symbolized

FPp.

One could string

together tense operators to form endlessly complex tenses.

In

order to analyze

sentences involving temporal quantification like "Three days ago John caught the

8:

15

train

to

London," one may introduce a variable after the tense-logical operators

to indicate how many time units from the present it was or will be the case that

p,

thusly:

Pnp.

We

need not at this point delve into the cottage industry that has grown

up

around tense logic in order to appreciate the beauty and perspicacity

of

Prior's

analysis

of

verbal tense.

Temporal Indexicals

By contrast, the nature

of

temporal indexicals and

of

indexicals in general

is

much less clear. Indexicals are usually defined

as

words whose respective referents

are not given once for all in the way that the referent of a proper name or definite

description

is

given, but whose respective referents vary in a systematic way from

context

to

context.

14

There are four broad types

of

indexicals, examples

of

which

serve to

clarifY

the above definition:

15

(1)

personal indexicals such

as

"I," "you,"

"he,"

etc.;

(2) temporal indexicals such as "now," "then," "yesterday," "today,"

"tomorrow," etc.; (3) spatial indexicals such as "here," "there," "nearby," "far

away,"

etc.,

and (4) demonstratives such as "this," "that," "these," and "those."

It is

evident that the referent of such words will vary with the context. If John says, "I'm

hungry " and Jane says, ''I'm hungry " they do not refer to the same person, despite

their using the same words. To refer to the same person as John does, Jane must

employ a different personal indexical, either "You're hungry " or "He's hungry "

depending once again upon the context

of

who

is

being addressed by Jane.

Similarly, the adage that "Tomorrow never comes" plays upon the indexical fact that

when the day now referred to as "tomorrow" does arrive, it will be referred

to

by a

new word, "today." In the case

of

spatial indexicals, two persons

in

a telephone

conversation, for example, must use different indexical words to refer to the same

place: "How's everything there?" asks the caller, to which the other replies,

"Everything here

is fine." Demonstratives are distinguished from other indexicals-

14 See,

for

example, Graeme Forbes, "Indexicals," in Handbook of Philosophical Logic, ed. D. Gabbay

and

F.

Guenther, vol.

IV:

Topics

in

the Philosophy

of

Language,

Synthese Library 167 (Dordrecht:

D.

Reidel, 1989), p. 463.

15

I omit from indexical status expressions indicating what

is

actually the case, though some thinkers,

notably David Lewis, have defended actuality as an indexical notion. According to Lewis, what

is

actually the case varies from context to context, where the context

is

or includes the possible world in

which the sentence

is

purported to be true (David Lewis, On

the Plurality

of

Worlds

[Oxford: Basil

Blackwell, 1986],

p.

93; cf. idem, "Anselm and Actuality,"

Noils

4 [1970]: 185). Lewis's comparison of

"actual" with the temporal indexical "present"

is

suggestive and merits our discussion in the sequel.


17/298

10

CHAPTER

1

despite the fact that all indexicals have sometimes been misleadingly labeled

"demonstratives" 16

- by

the fact that their user employs them to point to something,

to direct attention to something. Other indexicals can be used

demonstratively-for

example, pointing to a spot on a map, one could say, "Right here "-, but then the

demonstrative force

of

the expression derives not from the indexical word, but from

the accompanying physical demonstration. B1 contrast, demonstrative words do not

require an accompanying demonstration; for example, "That remark was

unnecessary, Senator" or "This situation is perilous." Rather demonstratives have

an inherent demonstrative force, though, of course, a physical demonstration may be

necessary to communicate

one's

idea to another, as, for example, when one says, "I

choose that one out of the lot."

These examples make clear the context-dependent nature

of

the referents

of

indexical words. But very little reflection is required to see that it is not simply in

virtue of their context-dependence that indexical words differ from proper names

and definite descriptions, for the referents

of

proper names and definite descriptions

are no more "given once for all" in a context-independent way than are the referents

of indexicals. If one says, "Kennedy was assassinated by a deranged gunman," two

persons could truly believe this sentence without referring to the same man. Who is

referred to will depend upon the context of utterance to the same extent that "Now is

when the meeting starts" so depends. Rather a salient difference between these

types of referring expressions seems to have to do with the

systematic way

in which

the referents

of

indexicals vary from context to context. The referent

of

a proper

name or defmite description may differ from context to context, but, having

determined the referent in a particular context, one may still use that same name to

refer to the same individual in a different context. But in the case

of

indexicals, a

systematically correlated class

of

words exists which must be used to pick out the

same referent under appropriately different circumstances. For example, if we were

talking about Kennedy's assassination and were clear that we meant the President,

then at a later time and different place, talking to different persons, we still may use

the name "Kennedy" to refer to the same person. But when we later refer to the

moment which we once referred to as "now," we use the word "then"; to refer in

another place to the location elsewhere referred to as "here," we say "there"; to refer

to the person whom I called "you" in another encounter, I instead use the words

"he" or "she."

Of

course,

if

the context changes in such a way that we are no longer

16

Most notably by David Kaplan, "On the Logic of Demonstratives," in

Contemporary Perspectives in

the Philosophy of Language,

ed. Peter

A.

French, Theodore E. Uehling, Jr., and Howard K. Wettstein

(Minneapolis: University of Minnesota Press, 1979), pp. 401-412; so also Palle Yourgrau, "Introduction,"

in

Demonstratives,

ed. Palle Yourgrau (New York: Oxford University Press, 1990), pp. 1-8. Kaplan later

repented of this misleading terminology, favoring the term "indexical" (Idem, "Demonstratives: an Essay

on the Semantics, Logic, Metaphysics, and Epistemology

of

Demonstratives and Other Indexicals,"

in

Themes from Kaplan,

ed. Joseph Almog, John Perry, and Howard Wettstein [Oxford: Oxford University

Press, 1989],

p.

490). Unfortunately, he still misconstrues a demonstrative as an indexical requiring an

associated demonstration (Ibid., pp. 490-492). However, in his "Afterthoughts," in Themes, pp. 587-588,

he correctly states,

"".

the meaning

of

a demonstrative requires that each syntactic occurrence be

associated with a directing intention... . The need for a directing intention to determine the referent

of

a

demonstrative still allows us to distinguish the true demonstratives from the pure indexicals."

17 Very often, the precise contradictory is asserted; but this is erroneous, as is pointed out by Howard

Wettstein, "Has Semantics Rested on a Mistake?"

Journal ofPhilosophy

83 (1986): 196.


18/298


11

talking about the President's assassination, but rather the Attorney General's, then

"Kennedy" will have a different referent in those contexts. But there is no

systematically correlated term to use in the new context to pick out the same referent

picked out in the former context; one can only add additional information, for

example, a first name or a fuller description, and these will be arbitrary. Thus, it

seems to be the systematic way in which indexicals pick out different referents in

different contexts that serves to distinguish them from proper names and definite

descriptions. Kaplan expresses the point by asserting that the meaning of an

indexical word, unlike proper names or definite descriptions, provides a rule

of

use

which determines the referent in terms

of

certain aspects

of

the context.

IS

For

example, the meaning

of

"I" determines a linguistic rule in English that the referent

of the term in a given context will be the person tokening in speech or thought the

sentence

in

which "I" appears;

by

contrast the meaning

of

"you" determines a rule

that the referent

is

the individual being addressed. With respect

to

temporal

indexicals, there are linguistic rules determined by the meanings

of

the words which

stipulate that the time referred to is a moment or interval earlier than, simultaneous

with, or later than the time

of

any tokening

of

the sentence.

19

By contrast, proper

names and definite descriptions do not determine by the meaning of the words

involved any such linguistic rules to fix

their referents.

The question now arises as to what relationship exists between verbs and

temporal indexicals. Are verbs themselves temporal indexicals? It would seem so,

for a sentence like "The meeting

is

starting" seems

to

mean,

in

virtue

of

its present

tense verb, the same as "The meeting is starting now."

If

that is the case, then

it

seems that just

as

the referent

of

"now" various systematically with the context, so

the time implicitly referred to in the verb varies systematically with the context.

Every time the present-tense sentence

is

tokened, a different time

is

referred to, and

in order to refer to the same event at another time, we must use different forms

of

the verb, namely, past- or future-tense. Accordingly, verbs, too, seem to be

temporal indexicals.

What this position fails to take cognizance

of is

the necessity

of

what is called

"double indexing" with respect to sentences which contain temporal indexicals and

are governed by temporal operators. First noticed by Hans

Kamp,20

the necessity

of

18 Kaplan, "Demonstratives," pp. 490-491.

19

There are apparent, interesting exceptions to this practice; for example, "Churchill now faced the

most critical moment

of

his career." For an entertaining discussion, see Quentin Smith, "The Multiple

Uses oflndexicals," Synthese 78 (1989): 167-191. But I should say, and I think Smith would agree, that

in such circumstances the temporal indexicals are in fact tenseless and therefore not truly temporal

indexicals (Smith, Language and Time, p. 24, note 3.)

20

Hans Kamp, "Formal Properties of 'now',"

Theoria

37 (1972): 227-273; see also

A. N.

Prior,

"Now,"

Nails

2 (1968): 101-119; idem,

' ' 'Now'

Corrected and Condensed,"

Nails

2 (1968): 411-412;

Pavel Tichy, 'The Transiency of Truth," Theoria 46 (1980): 164-182. While recognizing the non

redundancy

of

"now," Tichy ostensibly opposes double indexing because it jettisons the principle that

propositions take truth values relative to a world and a time in favor of the precept that propositions take

truth values relative to a world, a time, and yet another time. On the conventional view, he asserts,

whenever a proposition is asserted, a unique world is actual and a unique moment of time is current.

Accordingly, he proposes to analyze a sentence like It will be the case that Brown is not at home now"

as expressing the proposition For some moment u, u is present and it will be the case that Brown is not at

home at

u.

But I think that Tichy is clearly confusing semantics with ontology. Double indexing is a

semantical feature of sentences that

in

no way denies that there is a unique moment of time which is


19/298

12

CHAPTER 1

double indexing arises from the fact that temporal indexicals

in

contexts governed

by tense operators are not necessarily based on the time parameter specified by the

tense operator

in

whose scope they lie, but may have a different temporal base.

21

For example, suppose during the Reagan presidency someone were to assert "The

present U.S. president was once a Democrat." How does that differ from the

assertion at the same time that "The U.S. president was once a Democrat"? The

answer is that the indexical word "present" requires one to assess the referent of

"U.S. president" relative to a different time than that indicated by the verbal tense,

whereas

in

the absence of this indexical, there

is

only one temporal parameter. This

ontologically present. Tichy's own analysis involves double indexing, since the time referred to by the

future-tense operator is not the same time as

u.

Indeed, his analysis has the virtue

of

making the indexical

time parameter's independence

of

the temporal operator perspicuous. Prior himself observed, " ... we

can

dispense with the non-redundant 'now' in favour

of

the redundant one. In other words, the non-redundant

'now' is non-redundant only in the sense that you cannot just erase it from a sentence and leave the sense

of the whole the same; you can, however, erase it and get something with the same sense by altering the

rest of the sentence somewhat" (Prior, "Now," p. 106).

21

For an outstanding account

of

double indexing and

of

temporal indexicals in general, see Nathan

Salmon, "Tense and Singular Propositions," in

Themes/rom Kaplan,

pp. 354-367. Salmon's theory

of

indexicals

is

based on a semi-Fregean propositional analysis

of

tensed sentences. On Salmon's analysis,

the semantic basis

of

a sentence is a propositional matrix, which consists

of

the referent

of

the subject

of

the sentence and the property ascribed to the referent by the sentence's predicate. The propositional

matrix becomes a piece

of

information by attaching to it a particular time at which the property inheres in

the referent. Sometimes

it

may

be

necessary to attach a particular location

as

well

to

obtain a genuine

piece

of

information or proposition. The propositional matrix serves as the information value base for a

tensed expression. An indexical expression is precisely one that takes on different value bases with

respect to different possible contexts" (Ibid.,

p.

346). As the context of utterance varies, the referent of

such expressions varies, as I explain in the text. By contrast, a tensed sentence like "Frege

is

busy," while

having different information value or propositional content at different times, has the same information

value base or propositional matrix,

viz. the complex consisting of Frege and the property of being busy.

The time is implicitly built into the information value of the predicate, which is why such non-indexical

sentences take on different truth values and different information values when uttered at different times,

even though the expression

is

not indexical. For the same reason certain non-indexical definite

descriptions like "the senior senator from California" take on different referents and different information

values at different times. The distinctive feature of

an

indexical expression like "I" or "the present senior

senator from California"

is

that it takes on different information value bases

in

different contexts. "The

predicate 'be busy,' the definite description 'the senior senator from California,' and the sentence 'The

senior senator from California

is

busy'

all

retain the same value base in all contexts. Their information

value varies with the context, but not their information value base" (Ibid.,

p.

369). For example, the

information value base of the aforementioned sentence is comprised of the senior senator and the property

of being busy, and this base remains the same at all times though attached

to

different times. But the

information value base of "The present senior senator from California

is

busy" will vary with the context

of utterance because the time

is

included in the value base itself. The semantic difference between these

similar expressions does not show up clearly in sentences in which the time attached to the value base

is

the same time

as

the time inherent in the value base of the indexical expression, but when double

indexing is required because the times differ, the difference becomes clear, e.g., "In 1913 the senior

senator from California was a child" versus "In 1913 the present senior senator from California was a

child." According

to

Salmon, the function

of

indexicals like "present" or "now"

is

"primarily

to

affect

the content base

of

its operand, eternalizing it and thereby sealing it off from the influence

of

external

occurrences

of

temporal operators" (Ibid., p. 377).

Salmon's characterization

of

indexicals depends on a Direct Reference Theory, propositions as

objects

of

belief, and a tenseless view

of

information and is thus too controversial to serve as a general

characterization of indexicals. But his focusing on the phenomenon of double indexing as distinctive of

indexicals is

an

insight which transcends his theory. One can agree that indexicals are just those

expressions which are such that they would

be

classed as indexical if Salmon's theory were correct.


20/298


13

double indexing evinces the fact that temporal indexicals have a certain

independence from the tense logical operators in whose scope they lie. Double

indexing is a very common feature

of

ordinary language: "He assured me that he

would arrive today," "We shall never forget what transpired three days ago," "What

you said now troubles me."

By contrast verbal tense and non-indexical temporal expressions do not require

double indexing. For example, contrast the two sentences

3. It will be the case tomorrow that I am sitting down.

4. It will be the case tomorrow that I am now sitting down.

The first sentence requires only single indexing, since there is no indexical within

the scope

of

the tense logical operator (the "tomorrow" serving to quantify the

operator itself) and thus naturally means that tomorrow I shall be sitting down then.

But the second sentence requires double indexing due to the indexical "now" within

the scope

of

the operator and naturally means that my sitting down today will be the

case tomorrow, that is, that tomorrow it will be true that "I was sitting down

yesterday." The absence of double indexing for merely verbally tensed sentences

reveals that the so-called "redundancy theory"

of

the "now,,22- that adding "now"

to a present-tense sentence does not change the semantic character

of

that

sentence-is

incorrect. Even

if

the information value or propositional content of It

is 3 o'clock" and

It

is now 3 o'clock" is the same, nevertheless these sentences

differ in their semantic properties in that only the latter requires double indexing

when in the scope of a tense-logical operator.

Similarly, non-indexical time indicators do not require double indexing.

Consider the sentence, "John said he would call me on the following day." The

phrase "on the following day" does not express an A-determination, but a B

determination. Perhaps in the original context John said, "Mother will arrive on

May 12, and I will call you the following day." The day referred to may now be

past, present, or future. No double indexing

is

required because the phrase does not

refer to any second time parameter. By contrast, if one reports, "John said he would

call me tomorrow," the "tomorrow" refers not to the context

of

John's utterance, but

to a second time parameter related to the reporter's present?3 Thus, non-indexical

or B-theoretical temporal expressions do not require double indexing.

Such considerations prompt Salmon to conclude:

What

is

distinctive about indexical expressions (such as

T,

'this tree,' or 'the present

U.S. president') is not merely that the extension with respect to a context

c (simpliciter)

varies with the context

c,

or even that the intension or information value with respect to

22

See A. N. Prior, Time and Modality (Oxford: Clarendon Press, 1957), pp. 9-10; idem, "Changes in

Events, pp. 8-10, 14-15; idem, "On Spurious Egocentricity," in

Papers on Time

and

Tense,

pp. 17-23.

23

Kaplan errs, therefore, in asserting that "yesterday" is an indexical and "one day ago" is an iterative

temporal operator (Kaplan, "Logic of Demonstratives,"

p.

412). Compare "It will always be the case that

John yawned yesterday" and "It will always be the case that John yawned one day ago." The expression

"one day ago" refers to the time

of

the speaker's present and is not comparable to "on the preceding day."

Thus, "One day ago it was the case that one day ago it was the case that John yawns" does not mean

"John yawned the day before yesterday," for the second token of "one day ago" within the scope of the

first requires an independent temporal reference point, unlike the iterative "on the preceding day."


21/298


22/298


15

A-Predicate Adjectives

With respect to A-predicate adjectives, Smith sets forth the interesting

contention that A-determinations like "past," "present," and "future" are not

indexical expressions just in case they are preceded by an A-copula, since in such a

case they do not require double indexing when

in

the scope oftemporal operators.

28

Contrast, for example, the sentences "The storm

is

occurring now" and "The storm

is

present.,,29

When

in

the scope

of

a temporal operator, "now" requires double

indexing, for example, It will be the case tomorrow that the storm

is

occurring

now." But "is present" remains singly indexed, for example, It will be the case

tomorrow that the storm

is

present."

In

this case "to be present" means "to have the

property of presentness," and the sentence means that tomorrow there will be a

storm. In this understanding, Smith agrees with Prior, who wrote, "In

'It

will be the

case tomorrow that

my

sitting down

is

present,' the presentness referred to

is

a

presentness that will obtain tomorrow, i.e., at the time to which we are taken by the

tensing

prefix.,,30

The use

of

"present" following the tensed copula thus differs from

the indexical

use of

the word

in

such sentences

as

It

will be the case tomorrow that

the present storm

is

occurring" or "It will be the case tomorrow that the storm is

presently occurring," which like the sentence containing "now" involve double

indexing.

It seems to me

correct that expressions consisting

of

an A-determination

preceded by a tensed copula are non-indexical, but it might be questioned whether

they are

in

fact A-expressions.

It

might be contended that "past," "present," and

"future" in such constructions are being used

in

the B-theoretical sense noted by

Goodman. For example, when it

is

said, It will be the case tomorrow that the storm

is

past," one means that the storm

is

earlier than the day specified by the operator,

and when one says, It was the case yesterday that the storm

is

future," one means

later than the time specified. In the same way, to say,

It

will be the case tomorrow

that the storm

is

present" means that it will be simultaneous with the day

in

question. Hence, these expressions are not A-determinations.

But I think that it

is

open to question whether such expressions are always used

in this B-theoretical sense. When we comfort ourselves by saying, "Tomorrow this

will all be past,"

we

do not seem to be saying that it will be earlier than

tomorrow

indeed, that fact could

be

the source

of

our distress-but that it will have the

A

determination of being past. My point

is

not to rehearse here the experiential

arguments for the A-Theory. Rather the point

is,

whatever the semantics

of

such

sentences, ordinary language users seem to employ them to express

A

determinations,

so it

is

simply incorrect to assert that expressions consisting

of

a

copula

+ an

A-determination

do

not

in

fact purport to ascribe A-determinations.

28

Smith,

Language and Time,

pp. 116-120; cf. pp. 74-77.

29

Smith's point is even more obvious

if

we change the tense of the copula

in

accord with either single

or double indexing, e.g. It will be the case tomorrow that the storm was occurring now" and It will be

the case tomorrow that the storm will be present." In ordinary language we often shift the tense of the

verb to conform to the time of reference. The artificiality of using tense operators on present-tense verbs

obscures the meaning evident in ordinary language.

30

Prior, "Now," p. 104.


23/298

16

CHAPTER 1

Perhaps the best argument against regarding the expressions in question as a

separate category

of

non-indexical A-expressions

is

to contend that they are really

examples

of

verbal tense. Some A-theorists object to treating presentness as a

property rather than as a tense logical operator and so would interpret Smith's

sentence to mean It will be the case tomorrow that the storm occurs" and the

ascription of pastness to mean It will be the case tomorrow that it was the case that

the storm occurs" and the ascription of

futurity to mean

It

was the case yesterday

that it will be the case that the storm occurs." Thus, the copula + A-determination

is

non-indexical because it really

is

verbal and, hence, a sentential tense operator. But

while such an interpretation might serve as a sound semantical analysis

of

such

sentences, the fact remains that on the linguistic-grammatical level the expressions

in question just are not verbs, but a verb plus a predicate adjective. On the linguistic

level Smith seems correct that we have here examples

of

non-indexical, non-verbal

A-expressions.

Quasi-Indexicals?

Finally, we should ask whether there

is

a fourth type of linguistic expression

used for expressing tense in English-what the late H.-N. Castafieda called ~ u a s i

indicators," or-to keep more in line with our terminology-quasi-indexicals? He

observed that in

oratio obliqua

(indirect discourse), that

is

to say,

in

a clause

subordinated to a verb expressing a propositional attitude, the indexicals used by the

original speaker in oratio recta (direct discourse) cannot be adequately captured by

de dicto or de re expressions but require quasi-indexical words. Castafieda calls

such words "quasi-" indexical because

(i)

they do not express an indexical reference

made by the speaker

of

the

oratio recta,

and

(ii)

they are used to attribute implicit

indexical references to that speaker. Consider, for example, the sentence "The

Editor of Noils believes that he himself is a millionaire." This sentence entails a

sentence such as "The Editor of Noils believes what he would express by saying,

'I

am a millionaire'." But then the first sentence cannot be equivalent in meaning to a

de dicto

sentence like "The Editor

of Noils

believes that the Editor

of Noils is

a

millionaire" because the latter sentence does not entail that the Editor

of

Noils

believes himself to be a millionaire (perhaps he has not yet learned

of

his

appointment to the editorship). But neither can the first sentence be interpreted as

ascribing a

de re

belief to the Editor, such as "The Editor

of Noils

believes

of

the

Editor of Noils that he

is

a millionaire," for this belief also fails to entail that the

Editor

of

Noils believes he himself to be a millionaire (since it does not entail that he

believes "I am the Editor of Noils"). Thus, in order to make an indexical reference

repeatable by various persons, yet preserve its indexical character, we must use

quasi-indexicals, which Castafieda proposed to mark with "*".

31 Hector-Neri Castaileda, "'He': A Study in the Logic

of

Self-Consciousness,"

Ratio

8 (1966): 130-

157; idem, "Indicators and Quasi-Indicators,"

American Philosophical Quarterly

4 (1967): 85-100; idem,

"Omniscience and Indexical Reference,"

Journal o/Philosophy

64 (1967): 203-210; idem, "On the Logic

of

Attributions

of

Self-Knowledge to Others," Journal

0/

Philosophy 65 (1968): 439-456. For a readable

secondary account, see Esa Saarinen, "Castaileda's Philosophy

of

Language," in Hector-Neri Castaneda,

ed. James

E.

Tomberlin, Profiles 6 (Dordrecht: D. Reidel, 1986), pp. 187-214.


24/298


17

Although most of Castafieda's work concerned personal indexicals, the same

point can

be

made about other indexicals, including temporal indexicals. Castafieda

provides as an example the sentence "On May

15,

1911, the German Emperor

believed that it was raining then·." What did the Kaiser believe in

aratia

recta?

Not the

de dicta

belief that it was raining on the date

in

question, for he may have

had no idea of the date; nor the de re belief concerning that day that it was a rainy

day, for that does not entail that the day was present. What he believed was what he

would have expressed by saying

It

is now raining."

Insightful

as

Castafieda's analysis

is,

it seems to me that

he

has not discovered a

new logical term called "quasi-indexicals" but simply uncovered the fact of double

indexing. The fact that such terms express, not the indexical references of the

speaker

of

the

oratio recta,

but rather the indexical references

of

the speaker

of

the

whole complex utterance,

is

due to double-indexing, not to a new logical term,

quasi-indexicals. When one says, "On May

15,

1911, the German Emperor believed

that it was raining then," the use of "then" is an ordinary indexical which

necessitates double indexing

of

this sentence in virtue

of

its lying within the scope

of the temporal operator. The fact that it

is

oratio obliqua

is

really a red herring, as

is

evident from the fact that

if

the time

of

the context

of

utterance

is

the same time

as

the time of the operator, then the word "then" becomes inappropriate in oratio

obliqua:

"The Emperor believes that it

is

raining now."

In

this case the two indices

coincide. Similarly, when the context

of

utterance does not shift persons, then "he"

is

inappropriate: "I told you I was a millionaire." It

is

not the shift from

oratio recta

to oratio obliqua as such which requires a change in the indexical, but a shift in the

context

of

utterance.

As we

noted earlier, indexicals come

in

systematically

correlated pairs or triplets, and which

of

the group

is

appropriate for retaining the

original reference varies as the context varies. Thus, in the case of temporal

indexicals, "then" must be used instead

of

"now" to refer to times past or future with

respect to the context

of

utterance (for some reason English fails to discriminate

between past-then and future-then, so any time other than the present can be referred

to

as

"then"). This change

is

required even while remaining within

oratio recta.

For example, one must

in

the context of 1995 say, "On May

15,

1911, it was raining

then" or "On May

15,

1911, it was the case that it

is

then raining," rather than "On

May 15, 1911, it was now raining." The indexical nature of "then" becomes

apparent when

we

make the verb tenseless: "On May

15,

1911, it rains then," for the

word "then" reveals that the context of utterance is not May 15, 1911.

Castafieda thinks that quasi-indexicals necessarily have antecedents

in

the

sentential context. Otherwise, a sentence like "The Emperor believed it was raining

then" would

be

ambiguous. However, pace Castaneda, the necessity of an

antecedent for temporal indexicals like "then," "past," or "future,"

is

not due to their

supposed quasi-indexical nature, but due to the fact that unlike their correlates

"now" and "present," they ascribe only an A-determination but not an A-position, so

that they need to be supplemented by another expression fixing the position.

Indexicals ascribing both an A-determination and a temporal position require no

antecedents; for example, "The Emperor believed it was raining yesterday." When

evaluated in a context

of

utterance c including the date May 16, 1911, this sentence

is

unambiguous and true.

In

other contexts "yesterday" changes its referent,

so

that


25/298

18

CHAPTER 1

the sentence may be false. I t behaves the same as "then" plus a date. What

Castafieda took to be an essential property

of

so-called quasi-indexicals

is

in fact

only a property ofindexicals which fail to ascribe an A-position.

A similar point could be made about Castafieda's claim that from a quasi

indexical in oratio obliqua one can infer an appropriate indexical originally in oratio

recta, but that from an indexical in oratio obliqua one cannot infer that in the

original oratio recta

an indexical reference was made. For example, " ... when

Gaskon says, 'Yesterday Privatus thought ... that it would be raining now (today),'

Gaskon's statement both contains his own indexical uses

of

'now' ('today') and fails

to imply that Privatus referred indexically to the time at which Gaskon made his

statement.,,32 Privatus, for all we know, could have actually said,

It

will be raining

on May 15," but Gaskon on May

15

reports this using an indexical. The reason

behind this phenomenon

is

not, however, due to the fact that quasi-indexicals are a

special term distinct from indexicals, but rather to the shift in the context of

utterance such that the time parameter

of

the indexical does not coincide with the

time of the oratio recta. But if they do coincide, then one can infer an indexical in

the original oratio recta. Suppose Gaskon, talking on the telephone to Privatus

about his expected visit, whispers to his wife, "Privatus says he's arriving today "

Then the "today" indicates both Gaskon's context

of

utterance as well as Privatus's

and expresses what Privatus would say in oratio recta. 33 What Castaneda should

have said, it seems, is that from a past- or future-tensed indexical in oratio obliqua

which refers to the same time as the time

of

the

oratio recta,

one can infer a present

tensed indexical in oratio recta. But from a present-tensed indexical in oratio

obliqua, one cannot infer an indexical expression in oratio recta, unless the oratio

obliqua

is

reported in the present tense. I am not particularly concerned to assess

the truth

of

this

claim;34

rather my point

is

that there appears to be no justification

for inventing a new singular term called "quasi-indexicals"

in

addition to ordinary

indexicals.

32 Castai'ieda, "Omniscience," p. 206.

33

Of

course, one cannot infer that Privatus actually uttered the indexical word; perhaps he said,

"I'm

arriving this afternoon" or "I'm arriving at 3:00 p.m." But to the same degree, uncertainty as to the actual

words in oratio recta also persists when we have a so-called quasi-indexical. Perhaps the Kaiser said

"This is a rainy day" or, looking at his calendar, "May 15 is a rainy day." In both these cases, the point is

that the persons believed something which could be expressed with present-tense indexicals like "now" or

"today."

Cf.

Saarinen's complaint that, unlike "he himself," "then" and "there" do not inevitably attribute

to someone the use

of

their indexical counterparts and are not therefore real quasi-indexicals (Saarinen,

"Castai'ieda's Philosophy of Language," p. 201). I agree with Castai'ieda that this merely represents a

deficit in natural language (Hector-Neri Castai'ieda, "Reply to Esa Saarinen," in

Castaneda,

pp. 349-350.

34 Castai'ieda himself confessed, "I cannot, however, muster a formal argument to show this"

(Castai'ieda, "Indicators and Quasi-Indicators," p. 96). There is a very interesting exchange of views

concerning the related issue of whether the proposition expressed by the quasi-indexical clause is the

same as that expressed by the corresponding indexical clause in oratio recta in Robert Adams and

Hector-Neri Castai'ieda, "Knowledge and Self: A Correspondence between Robert M. Adams and Hector

Neri Castai'ieda," in Agent, Language, and the Structure of the World, ed. James

E.

Tomberlin

(Indianapolis: Hackett Publishing, 1983), pp. 293-308, and Hector-Neri Castai'ieda, "Reply to John Perry:

Meaning, Belief, and Reference," in Agent, Language, and Structure, p. 327. Although Castai'ieda gives

up this claim, it seems to me that Adams's argument calls into question not so much this claim as the

claim that someone who knows that another person knows that

p

also knows that

p.


26/298

INTRODUCTION: LANGUAGE, TENSE,

AND

ONTOLOGY

19

LINGUISTIC AND ONTOLOGICAL TENSE

All of the above has been said concerning linguistic tense, such as we all know

and use it. But the A-theorist claims that tense

is

real in a more fundamental sense,

in an ontological or metaphysical sense, that the extra-linguistic world

is

itself

characterized by A-determinations like

past, present,

and

future,

that there actually

exists a "now" or a present moment

of

time. For the A-theorist tense

in

language

is

but a reflection

of

the way the world is,

of

ontological tense. This constitutes,

of

course, the great divide between A-theorists and B-theorists. The latter, while

acknowledging the undeniable presence of tense in ordinary language, hold that

reality itself is tenseless, that there are no tensed facts, no ontological past, present,

or future. As Mellor explains, for the B-theorist there are no two possible worlds W

and

W*

which are identical in their histories except for the fact that in

W

it is now

1995, while in W* it

is

now 1795: "".there

is

no tenseless difference between the

two worlds. Indeed, there are not two worlds.... There is only one world, with

things and events scattered throughout B-series time as they are throughout space,

including both the eighteenth and twentieth centuries.,,35 For the B-theorist tense

is

a feature

of

language, not of the world.

Many A-theorists maintain that tense cannot plausibly be regarded as a feature of

language and thought alone, that language furnishes us, as it were, a sort

of

window

on the world whereby we may apprehend the factual objectivity of tense. They

argue that the ineliminability or irreducibility of tense in language and

its

indispensability for human life and action make it plausible that tense

is

a feature

of

reality as well as

of

language and that therefore the A-Theory is preferable to the B

Theory.

Against this claim, B-theorists have opposed two objections: (1) The tense that

characterizes ordinary language is eliminable by translating sentences

of

ordinary

language into a tense less canonical form which preserves the meaning of the

ordinary language sentences. Linguistic A-determinations can be reduced to B

determinations by a variety

of

devices. Hence, the tense that characterizes ordinary

language furnishes no insight into the actual nature

of

time itself. This point

of

view

tends to be associated with the earlier generation

of

B-theorists, such as Russell,

Reichenbach, Quine, and Smart, but tends to be rejected by contemporary B

theorists and so may conveniently be called "The Old B-Theory of Language.,,36 (2)

While tense is an ineliminable feature

of

human language, indispensable for timely

action in human affairs, nonetheless the fact that human language cannot be divested

of

tense has no ontological implications concerning the nature

of

time. For tenseless

truth conditions can be given for all A-sentences, as well as B-sentences, so that it

becomes gratuitous to posit tensed facts in order to make A-sentences true or false.

Since nothing more than tense less facts is required in order for A-sentences to be

true or false, the fact that tense

is

not eliminable from human language and thought

15

Mellor,

Real Time, p.

56;

cf. p.

28, where

he

admits that unless the B-theorist shows that there

is

no

difference between Wand W*,

he

cannot deny that the A-series describes a real aspect

of

the world.

36 An appellation I adapt from Smith,

Language

and

Time,

sees. I. 2-3; cf. idem, "Problems with the

New Tenseless Theory of Time," Philosophical Studies 52 (1987): 371-392. I prefer my appellation

because the theories

in

question are really theories of language, not time. Reichenbach, for example, held

to an A-Theory of time, but a B-Theory of Language.


27/298

20

CHAPTER 1

does not prove that there are tensed facts in addition to tenseless facts. This point of

view characterizes many B-theorists

of

the contemporary generation such as D. H.

Mellor and has certain affinities with recent theories

of

Direct Reference and so may

conveniently be labeled "The New B-Theory

of

Language." In his most recent

work, Mellor has crucially adjusted his view, abandoning truth conditions in favor

of

so-called truth makers

of

tensed sentences, and advocates what we might call

"The Indexed B-Theory of Language."

The purpose

of

the next three chapters will be to explore the extent to which

tense in language and thought makes an A-Theory

of

time more plausible than a B

Theory. But it is imp.ortant at the outset to be clear about the probative force

of

these arguments. Some thinkers mistakenly believe that the successful reduction or

elimination

of

tense from language or the furnishing

of

tenseless truth conditions for

A-sentences

is

a positive proof that tense

is

not

ontologically real. For example,

R.

C. Coburn furnishes the following "coercive" argument against the A-Theory:

5. Ifthe

doctrine

of

passage

is

true, then there are A-facts.

6.

If

there are A-facts, then they determine the truth values of tokens

of

A-sentences.

7. It is

false that A-facts determine the truth-values

of

tokens

of A

sentences.

8. Therefore, the doctrine

of

passage is false.

3

?

The problem with this line

of

argument is that the B-theoretical claim to provide

successfully tenseless truth conditions or truth makers for A-sentences provides no

warrant for accepting (7), since it could well be the case that tensed truth conditions

and truth makers can be provided for A-sentences as wel1.

38

Thus, success in

demonstrating the central claim

of

either the Old or New or Indexed B-Theory

of

37 Robert C. Coburn, The Strangeness of the Ordinary: Problems and Issues in Contemporary

Metaphysics (Savage, Md.: Rowman & Littlefield, 1990),

p.

113. Coburn also provides the following

argument for a B-Theory:

I. Either the truth-conditions of tokens of A-sentences are timeless or they are not.

ii.

If

they are timeless, then the ostensible existence

of

A-facts

is

an illusion.

iii.

If

they are not timeless, then, provided event e occurred five years ago, any token

of the sentence "e happened five years ago" will be true regardless of the date of

its occurrence.

iv. It is false that any token of such a sentence is true regardless of the date

of

its

occurrence.

v. Therefore, the doctrine

of

passage is false.

But Coburn presupposes unjustifiably that A-sentences cannot have both tenseless and tensed truth

conditions; otherwise (ii) does not follow. Such a presupposition requires justification.

In

(iii) and (iv)

he tries to justifY the denial that tensed truth conditions can be given; but these steps just are the nerve

of

McTaggart's Paradox as Mellor exposits it (D. H. Mellor, "Tense' s Tenseless Truth Conditions," AnalYSis

46 [1986]: 171) and Coburn acknowledges his indebtedness to Mellor on this score (Coburn, Strangeness

of

the Ordinary, p. 114). Thus, the argument for the B-Theory relies on the positive demonstration that

the A-Theory is fatally defective.

38

As emphasized by Graham Priest, "Tense and Truth Conditions," Analysis 46 (1986): 162-166; also

Smith, Language and Time, chap. 4.


28/298


21

Language respectively does not furnish positive evidence for the B-Theory; it only

undercuts putative evidence for the A-Theory.39 Careful New B-theorists like

Mellor recognize this fact and therefore realize that their case against tensed facts

stands or falls with their demonstration

of

some positive defect in the A-Theory.

That is why his defense

of

McTaggart's Paradox

is

Mellor's self-confessed linchpin

in his case against tensed facts. Mellor is careful to claim no more on behalf

of

his

tenseless account in Real

Time

than that it provides one explanation

of

the truth

conditions for the truth of tensed sentence tokens, commenting, "That these simple

types

of

tensed sentences have these token-reflexive truth conditions is really quite

obvious, and

is

not seriously questioned. The serious question

is

whether this

is

all

there

is

to the facts oftense.,,4o He acknowledges that the mere existence ofa totally

tenseless account of tense differences

is

not itself enough to justify our accepting a

tenseless account

of

tensed facts. "The account's mere existence does not prove its

truth; and the existence

of

tenses is not disproved by showing how to save the

phenomena

of

tense without them."41 In order to show that providing tenseless truth

conditions "is all there is to the facts of tense," Mellor must prove that tensed facts

are impossible: "Tense

is

so striking an aspect

of

reality that only the most

compelling argument justifies denying it: namely, that the tensed view

of

time is

self-contradictory and so cannot be

true.,,42

We shall discuss Mellor's defense

of

39

Thus, one can have an A-theorist like Michael Tooley, who agrees that

if

tensed concepts are

semantically basic, then

an

A-Theory

of

time

is

correct, but who also insists that

if

tenseless concepts are

semantically basic,

it

does not follow that a B-Theory of time

is

correct (Michael Tooley, Time,

Tense,

and Causation

[Oxford: Clarendon Press, 1997], pp. 18-19). Tooley goes so far as to reject premiss (5)

of Coburn's argument. He holds that there are no tensed facts but that the body oftenseless facts which

exist as of any given time varies from time to time. For discussion see the symposium on Tooley's book

featuring comments by Storrs McCall, Nathan Oaklander, and Quentin Smith with Tooley's responses

in

Essays on Time and Related Topics, ed. L. Nathan Oaklander, Selected Papers of the Philosophy of Time

Society Proceedings, 1995-1999, pp. 2-42. Tooley's rejection of (5) presupposes that his notion of

"actual

as

of time

t" is

not a tensed concept. The difficulty,

as

Smith points out,

is

that Tooley takes this

notion

to be

an undefined primitive, so that it is far from clear what

is

meant. Smith charges that

as

Tooley employs this notion, it appears to be synonymous to "exists earlier than or simultaneous with t,"

in

which case Tooley's theory

is

not a dynamic theory of time at all. Tooley repudiates this charge,

claiming that Smith himself must assume that these are not synonymous expressions. But, of course,

Smith, unlike Tooley, accepts the reality

of

tensed facts, so that

he is

able to provide a tensed parsing

of

"actual

as

of time t." Tooley maintains that this crucial notion is not a tensed concept, since it would then

not be neutral with respect to various A-theoretical ontologies, implying as it does an ontology according

to which past and present entities alone are real and thereby excluding presentist and "full-future"

ontologies. But this allegation seems incorrect. For example, A-theorists could leave "actual"

as

an

undefined primitive and analyze "actual

as

of time t"

as

"actual by t's being present." A-theorists will

just disagree among themselves as to which events are actual once t is present. If such disagreement is

not permitted by Tooley's notion, then

it is

obviously not essential,

as

Tooley avers, to a dynamic theory

of time.

It is

extraordinarily difficult to see how anyone with Tooley's ontology of a gradually accreting

past can avoid tensed facts. For at any point

in

time there will be truths about which proper subset of the

set of

all

tenseless facts is instantiated, and which subset that is is constantly changing.

If

Tooley tries to

avoid this conclusion by saying that such subsets are only instantiated

as

of a time t, not

Simpliciter,

then

we

are back to the necessity ofa tensed parsing ofthat notion. Thus, it seems to me, Coburn's (5)

is

true.

Still, the point remains that the provision of tenseless truth conditions for tensed sentences does not

disprove the existence

of

tensed facts.

40

Mellor,

Real Time, p.

42.

41

Ibid.

p. 33.

42

Ibid., p.

5. In

his

Real Time II,

Mellor does seem to subordinate McTaggart's Paradox to his

argument against tensed facts as truth makers of tensed sentences, but even so his argument, if successful,


29/298

Documents

The Tensed Theory of Time WilliamLane Craig