the tensed theory of time williamlane craig
TRANSCRIPT
-
8/20/2019 The Tensed Theory of Time WilliamLane Craig
1/298
THE TENSED THEORY OF TIME
-
8/20/2019 The Tensed Theory of Time WilliamLane Craig
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
-
8/20/2019 The Tensed Theory of Time WilliamLane Craig
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.
-
8/20/2019 The Tensed Theory of Time WilliamLane Craig
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
-
8/20/2019 The Tensed Theory of Time WilliamLane Craig
5/298
To
ALVIN PLANTINGA
who by his work and his life
has pointed the way
-
8/20/2019 The Tensed Theory of Time WilliamLane Craig
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
-
8/20/2019 The Tensed Theory of Time WilliamLane Craig
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
-
8/20/2019 The Tensed Theory of Time WilliamLane Craig
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
-
8/20/2019 The Tensed Theory of Time WilliamLane Craig
9/298
PART I
ARGUMENTS
FOR
AN A-THEORY OF TIME
-
8/20/2019 The Tensed Theory of Time WilliamLane Craig
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
-
8/20/2019 The Tensed Theory of Time WilliamLane Craig
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.
-
8/20/2019 The Tensed Theory of Time WilliamLane Craig
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.
-
8/20/2019 The Tensed Theory of Time WilliamLane Craig
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
-
8/20/2019 The Tensed Theory of Time WilliamLane Craig
14/298
INTRODUCTION: LANGUAGE, TENSE, AND ONTOLOGY
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.
-
8/20/2019 The Tensed Theory of Time WilliamLane Craig
15/298
-
8/20/2019 The Tensed Theory of Time WilliamLane Craig
16/298
INTRODUCTION: LANGUAGE, TENSE, AND ONTOLOGY
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.
-
8/20/2019 The Tensed Theory of Time WilliamLane Craig
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.
-
8/20/2019 The Tensed Theory of Time WilliamLane Craig
18/298
INTRODUCTION: LANGUAGE, TENSE, AND ONTOLOGY
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
-
8/20/2019 The Tensed Theory of Time WilliamLane Craig
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.
-
8/20/2019 The Tensed Theory of Time WilliamLane Craig
20/298
INTRODUCTION: LANGUAGE, TENSE, AND ONTOLOGY
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."
-
8/20/2019 The Tensed Theory of Time WilliamLane Craig
21/298
-
8/20/2019 The Tensed Theory of Time WilliamLane Craig
22/298
INTRODUCTION: LANGUAGE, TENSE, AND ONTOLOGY
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.
-
8/20/2019 The Tensed Theory of Time WilliamLane Craig
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.
-
8/20/2019 The Tensed Theory of Time WilliamLane Craig
24/298
INTRODUCTION: LANGUAGE, TENSE, AND ONTOLOGY
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
-
8/20/2019 The Tensed Theory of Time WilliamLane Craig
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.
-
8/20/2019 The Tensed Theory of Time WilliamLane Craig
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.
-
8/20/2019 The Tensed Theory of Time WilliamLane Craig
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.
-
8/20/2019 The Tensed Theory of Time WilliamLane Craig
28/298
INTRODUCTION: LANGUAGE, TENSE, AND ONTOLOGY
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,
-
8/20/2019 The Tensed Theory of Time WilliamLane Craig
29/298