Sentences About Time Quentin Smith

Sentences about the temporal determinations of events are of two sorts, tenseless and tensed. The former ascribe the relations of simultaneity, earlier or later (B-relations), and the latter arguably ascribe the properties of futurity, presentness or pastness (A-properties). If tensed sentences do ascribe these properties, then we are entitled to say that the two sorts of sentences reveal the two aspects of time, the B-aspect and the A-aspect.

However, it is customary among most analytic philosophers today to claim that tensed sentences about events do not really ascribe A-properties. Tensed sentences or their tokens, it is argued, are logically equivalent to, or have the same meaning as, tenseless sentences about events, and thus possess the same reference as the tenseless sentences, viz., to events with B-relations. It would follow that time has only one aspect, the B-aspect.

I believe this view is mistaken, and that tensed sentences do refer to events with A-properties, and consequently that time does possess an A-aspect. I will first criticize the token-reflexive theory of tensed sentences, and then examine the idea that tensed sentences or their tokens are analysable into tenseless sentences describing dates.

The most sophisticated token-reflexive theory has been developed by J. J. C. Smart; he claims that

when we say that the boat “was upstream, is level, will be downstream”, we are saying that occasions on which the boat is upstream are earlier than this utterance, that the occasion on which it is level is simultaneous with this utterance, and that occasions on which it is downstream are later than this utterance. That is, a language could be devised in which temporal copulae did not exist, but in which we used the words ‘earlier than’, ‘later than’, or ‘simultaneous with’ in combination with a non-temporal copula and the expression ‘this utterance’.[1]

These tenseless sentences do not ascribe A-properties, but report B- relations between the event described and the utterance.

Richard Gale has objected to Smart's analysis on the grounds that “ ‘This’ means ‘what I am now pointing at’. Anything which I am now pointing at is, by definition, present.”[2] Consequently, the analysans in the examples ascribe A-properties no less than do the sentences with the tensed copulae. I think this response to Smart is unsuccessful for the reason that it assumes without warrant that ‘this’ must be used in a tensed sense. Even if a reference to the present belongs to ‘this’ as it is ordinarily used (Webster's Dictionary arguably reflects ordinary usage when it defines ‘this’ as meaning “the person, thing or idea present”[3]), Smart need not be confined to ordinary usage in the sentences he is using to analyse ordinary discourse. Indeed, he explicitly avows that he is not using ‘this utterance’ in its ordinary tensed sense, but in a tenseless way that is

directly self-referential. We hear a token of the form ‘this utterance’ and simply understand that this token is the one referred to. We can at a later date say what the utterance referred to was: we can enumerate sufficient of its characteristics to identify it.[4]

I believe the real problem with an analysis of this sort is that the token-reflexive sentences violate the logic of the tensed sentences of which they are purported to be the analysans. Tensed sentences can be used in a consistent way to refer to the times at which there are no linguistic utterances or at which there might not have been any such utterances. A defender of the token-reflexive theory might object that a sentence like ‘The era before the Earth's formation is past’ means that the era before the Earth's formation (is) earlier than this utterance (the parentheses around the ‘is’ signify that it is tenseless), and that this token-reflexive sentence succeeds in referring consistently to a time at which there are no linguistic utterances no less than does the tensed sentence. But further reflection shows that the tensed sentence implies something that cannot be captured by any token-reflexive analysis, namely that if the preEarth era is past at some time t, (the time of utterance of this tensed sentence), then this era is present at some earlier time t₀. But that the preEarth era is present at t, cannot mean that the preEarth era (is) simultaneous with some utterance at t₀, since at this time there are no utterances.

It might be said that like Gale I am making unwarranted assumptions about language-usage and defending the A-theory on the basis of these assumptions. However, the cases are not parallel. Gale's assumptions are unwarranted since he supposes that ‘this’ must be used in its ordinary sense in the analysans of sentences, and it is not logically necessary that words in the analysans of sentences as they are ordinarily used themselves be used in the ordinary way. My claim, on the other hand, is about the analysandum of Smart's analyses, about the ordinary usage of the sentences being analysed. It is logically necessary that if some sentence S₁ is to be an analysans of another sentence S₂, then S₁ must have the same rules of usage as S₂. If it is a rule that S₂ can be used to refer to times devoid of linguistic utterances in a way that S₁ cannot be, S₁ cannot be an analysans of S₂. And that this difference obtains between tensed and token-reflexive sentences is a fact for which there is abundant evidence; it is a commonly accepted rule of usage that tensed sentences like ‘There will be no survivors after a nuclear war’ can be used to refer to times that if or when present are present without there being any utterances present.

It should be noted that this criticism of the token-reflexive analysis of tensed sentences need not be formulated in a way that seems to presuppose that tensed sentences ascribe A-determinations, for the point can be stated solely in terms of entailment relations between sentences, e.g. that

It need not be supposed that these sentences ascribe A-properties in order for it to be known that (2) cannot be analyzed as

(3) The preEarth era (is) simultaneous with some utterance at some earlier time t₀.

For given the rule that (2) can be used to make a true statement irrespective of whether or not there are language-users at t₀, and that (3) can be so used only if there are language-users then, and given the scientific fact that no language-users are known to exist then, it follows that (2) is true and (3) is false and hence that they are logically nonequivalent.

Rather than being presupposed by the criticism of the token-reflexive theory, the idea that tensed sentences ascribe A-properties is a result of this criticism, for it is introduced as an explanation of why (2) and (3) are nonequivalent --- namely that tensed sentences but not token-reflexive sentences ascribe utterance-independent temporal determinations, A-properties. That this explanation is not only allowed but also required by the nonequivalence of tensed sentences and token-reflexive sentences is demonstrated at the end of Section II.

It may be objected at this point that this explanation is not even allowed by my argument, since (2), although not logically equivalent to (3), is nevertheless logically equivalent to another tenseless sentence, namely

and therefore does not ascribe an A-property but the same B-relation ascribed by (4). The logical equivalence of (2) and (4), the objector argues, is due to the presence in (2) of an identification of a time, t₀, at which the preEarth era is present. To be present at a time t₀, it is pointed out, is toto caelo different from being present simpliciter, for to be present at t₀ is just to occur (tenselessly) at t₀. To be present simpliciter is not necessarily always to be present, but to be present at t₀ is necessarily always to be present at t₀, for once an event or series of events is present at t₀, it can never be the case that it is not present at t₀. And that is just to say that it is tenselessly true that the event or series of events is present at t₀; if this is tenselessly true, then it is logically equivalent to the assertion that the event or series of events occurs (tenselessly) at t₀. Similarly, to be past at a time t₁ is just to occur (tenselessly) earlier than t₁, so that (1) is logically equivalent to

Instead of addressing the issue of the validity of this objection, I shall reformulate my argument so that the objection no longer applies.[5] Let us substitute ascriptions of A-predicates simpliciter for the ascriptions of A-predicates at a time, and introduce sentences about propositions. With these changes, the first step in the argument is that

(7) is a past tense ascription of a truth-value of truth to the proposition expressed by ‘the preEarth era is present’. (7) accordingly is logically equivalent to

(8) A proposition was true, and this proposition is that expressed by ‘the preEarth era is present’.

The next step is to show that the token-reflexive analysis of (8) is false, whereas (8) is true. (8) consists of two conjuncts, and each requires a tenseless analysis. The first conjunct presents some difficulty, for if it were analysed according to Smart’s dictum that ‘was’ means ‘(is) earlier than this utterance’, it would read

but since tenselessly true propositions are true at all times, the last phrase ‘earlier than this utterance’ can be dropped as either misleading (implying that the proposition (is) true only earlier than the utterance) or part of a trivially implied phrase (the complete phrase being ‘earlier than this utterance, simultaneously with this utterance, and later than this utterance’). The first conjunct of (8) is thus analysed as

The second conjunct consists of a sentence in quotation marks and a preceding phrase. Let us assume for the sake of simplifying the analysis that the ‘is’ in the preceding phrase is a tenseless ‘is’ of identity (otherwise we would have cumbersomely to analyse ‘and this proposition is that expressed by’ as ‘and this proposition (is) that expressed simultaneously with this utterance by’). Given this assumption, the only part of the second conjunct that needs to be tenselessly analysed is the quoted sentence, which is straightforwardly analysed as ‘the preEarth era (is) simultaneous with this utterance’, so that the complete tenseless analysans of (8) is

(11) A proposition (is) true, and this proposition is that expressed by ‘the preEarth era (is) simultaneous with this utterance’.

(11) is probably false, since there probably were not utterances occuring in the preEarth era. However (8) is probably true, since scientific evidence suggests that it was true that a preEarth era is present. Consequently, (8) and (11) are logically nonequivalent, and (8) is not susceptible to a tenseless token-reflexive analysis. Note that if ‘A proposition was true’ is left unanalysed, so that the only difference between (8) and (11) is the sentence in quotation marks, it is still the case that (11) is false and (8) is true.

Two objections may be brought against this reformulation of the argument. The first is that

(12) (6) is logically equivalent to (7), and since (8) is logically equivalent to (7), (8) is also logically equivalent to (6).

(13) The token-reflexive analysans of (6) is ‘the preEarth era (is) earlier than this utterance’, which like (6) is true.

(14) By (12), it follows that ‘the preEarth era (is) earlier than this utterance’ is also an analysans of (8).

(15) Therefore, the crucial step in the argument, the claim that ‘the token-reflexive analysis of (8) is false’, is false.

This argument, although prima facie plausible, breaks down under scrutiny. 1 will not attack (13), since that would amount to begging a question at issue; rather I will show that (14) is false. (14) is based on the false assumption that a sufficient condition for an analysans of one sentence S, to also be an analysans of another sentence S₂ is that S₁ and S₂ be equivalent, that they entail each other. If this assumption were true, then an analysans of ‘The box has a shape’ would also be an analysans of ‘The box has a size’, since these two sentences entail each other; but their analysans are not identical, since the analysis of the predicate ‘has a shape’ is not identical with the analysis of the predicate ‘has a size’. It is not a sufficient but merely a necessary condition for two sentences having the same analysans that they be equivalent. A necessary and sufficient condition is that they be logically identical, i.e., that they attribute the same monadic or polyadic property or properties to the same logical subject(s). This definition is somewhat vague, but some brief examples will provide it with the minimal precision required for my purposes. ‘The box has a shape’ is logically nonidentical with ‘The box has a size’ since the former attributes the property, having a size, to the box, whereas the latter attributes the different property, having a shape, to the box. The box is the logical subject of both sentences. ‘The box has a figure’, on the other hand, is logically identical to ‘The box has a shape’, since they both attribute the same property to the same logical subject, the box. Likewise, ‘Each car is a pollutant’ is logically identical to ‘Each automobile is a pollutant’, since these two sentences attribute the same property, being a pollutant, to the same logical subjects.

I shall now show that (6) is logically nonidentical to (8) and therefore that the analysans of (6) cannot be the analysans of (8). The logical subject of (6) is the preEarth era, but the logical subject of (8) is a proposition, viz., the proposition expressed by ‘The preEarth era is present’. Since (6) is about an era and (8) is about a proposition, they cannot have the same analysans. Therefore the argument that the token-reflexive analysans of (8) is the true sentence ‘The preEarth era (is) earlier than this utterance’ is without merit.

A second objection to my argument is one that a proponent of the tenseless theory might have been wanting to make for some time now, namely that the proposition expressed by ‘The preEarth era is present’ cannot be said to have been probably true on the basis of current scientific evidence, since the sciences represent the world in tenseless terms and thus provide no evidence that any era is, was, or will be present. My response to this is that a tensed representation of the world is an essential ingredient of the semantic content of scientific theories --- but since I have argued this in detail elsewhere.[6] I will not develop this response further here.

There is a second argument that can be used against the token-reflexive theory. G. E. Moore’s “A Defence of Common Sense”[7] and related articles can be interpreted as exposing something that belongs to the logic of ordinary discourse about physical events, namely a commitment to the percipient-independence of most of these events. I believe this involves a related commitment to the utterer-independence of most future, present and past physical events. Consider these parallel cases: It belongs to rules of ordinary usage that if a present-tensed sentence like

(17) The volcano would be erupting now, even if nobody were around to perceive it erupting

can be used to express a proposition that is probably true. Similarly, it belongs to rules of ordinary usage that if a sentence like (16) is used to express a true proposition, then a sentence like

(18) The volcano would be erupting now, even if nobody were to comment upon its erupting

can be used in the same circumstance to express a proposition that is probably true. But whereas (18) is probably true, the token-reflexive sentence that is its putative analysans,

(19) The volcano would (be) erupting simultaneously with this utterance, even if nobody were to comment upon its erupting,

Paul Fitzgerald, adopting a suggestion made by Ronald de Sousa and Calvin Normore, argues that a contradiction-free token-reflexive analysis of sentences like (18) can be constructed.[8] The relevant analysans of (18) would be

(20) The volcano would (be) erupting at the time when (in fact, in our world as it is) this utterance (is) occurring, even if nobody were to comment upon its erupting.

(20) is contradiction-free, but it fails to analyse (18) since it expresses a proposition that entails the proposition expressed by

whereas the proposition expressed by (18) does not entail the proposition expressed by (21).

It might be objected that “this utterance” in (20) does not refer to an utterance of (20) but to an utterance of (18), since (20) is an analysans of (18). Such a position has been maintained by George Schlesinger[9] (but not by Fitzgerald[10]. But even in this case there is a logical nonequivalence: the proposition expressed by (20) entails the proposition expressed by

but the proposition expressed by (18) does not carry this entailment, since the latter proposition might have been true even if it were not linguistically expressed.

This latter observation points to an important difference between tensed and token-reflexive sentences. It belongs to the rules of ordinary discourse that if factual sentences like (16) and counterfactual sentences like (18) are used in some circumstance to express true propositions, then those propositions would have been true even if they had not been expressed. For example, it would have been true that the volcano is erupting even if nobody had tokened (16) on this occasion. I ask the person who denies this to imagine this conversation:

Sp₂: Yes, but it wouldn’t have been true if you hadn’t just asked that question.

Consider how different this conversation is from the conversation that the defender of the token-reflexive theory believes to be the analysans of the above conversation:

Sp₁: Is it true that the volcano (is) erupting simultaneously with this utterance?

Sp₂: Yes, but it wouldn’t have been true if you hadn’t just asked that question.

The difference between these two conversations reflects a basic difference between the rules of usage of tensed sentences and the rules of usage of token-reflexive sentences, a difference that prevents the latter from being analysans of the former .

A third argument can be developed if we apply an argument presented by Hector-Neri Castañeda[11] to refute the token-reflexive theory of the word 'I'. According to Reichenbach,[12] ‘I’ means ‘the person who utters this token’. Castañeda points out that if this analysis were correct, the proposition expressed by ‘I am uttering nothing’ would be self-contradictory. In fact, however, if a person utters this sentence he expresses a contingently false proposition, for the proposition expressed by this sentence might have been true; it would have been true had he then not expressed it.

This argument can also be used against the token-reflexive theory of the present tense ‘am’. This theory analyses ‘I am uttering nothing’ as ‘I (am) uttering nothing simultaneously with this utterance’, which is self-contradictory. But surely the proposition expressed by ‘I am uttering nothing’ might have been true; I might not have expressed it.

It is worthwhile noting at this point that other common forms of token- reflexive analyses of tensed sentences are also subject to the three arguments I have offered. Some token-reflexive analyses do not use the word ‘this’ to refer to the utterance to which the described event bears a B-relation, but instead include a description or name of the utterance. In terms of the example ‘The boat is level’, the analysans of some utterance u₁ of this sentence would be

(23) The utterance u₁ of ‘The boat is level’ (occurs) simultaneously with the boat (being) level

where the parentheses around ‘occurs’ and ‘being’ indicate they are here used tenselessly. In this case, the analysans is not token-reflexive, but it can be said to belong to a token-reflexive analysis of U₁ in the sense that it indicates u₁ to be token-reflexive.

The three arguments also apply to what is perhaps the most familiar token-reflexive analysis, the analysis of an utterance at t₀ of ‘That boat is level’ as

‘T₀’ does not in this case stand for a calendrical date, like May 9, 1985, but is defined solely as the time at which the utterance occurs.[13] Of course, each such time can be correlated with a date, but the date with which it is correlated does not enter into the meaning of ‘t’. The sign ‘t₀’ means the time at which the utterance occurs, such that sentences like (24) are logically identical with sentences like

(25) The boat (is) level at the time at which the utterance of ‘The boat is level’ occurs.

In (25) the reference of (24) to an utterance is made explicit, revealing that (24) is also subject to the three arguments presented.[14]

I shall conclude this section by defending my three arguments against a view advanced by Schlesinger. Schlesinger seems to believe that arguments of the sort presented in the previous pages are unsuccessful since they beg the question against proponents of the token-reflexive theory. This apparently is the idea behind the following passage from his Aspects of Time;[15] he first considers an objection:

Another objection to the view that equates the phrase ‘at present’ with the phrase ‘is simultaneous with this utterance’ has been that this cannot be so because the proposition ‘At present nothing is being uttered’ is a contingent proposition, whereas ‘At the time which is simultaneous with this utterance nothing is being uttered’ is self-contradictory.

Reichenbach and Smart do seem to be committed to the view that ‘At present nothing is being uttered’ is also self-contradictory. If this strikes us as strange, it is only because we have been tacitly assuming a different analysis of ‘at present’.

In these and other passages Schlesinger seems to be assuming that there are no matters of fact concerning rules of ordinary usage that could decide between the tenser and token-reflexive theory, and hence that the tenser and detenser are reduced to presenting question-begging arguments, each presupposing his own analyses of tensed sentences. I believe this assumption is mistaken; there are matters of fact concerning rules of ordinary usage, and these matters of fact are open to empirical observation and can be used as a court of appeal to decide between the different analyses offered by the tenser and detenser. It is precisely these matters of fact to which I have been appealing in my previous arguments. I believe these matters of fact support the tenser’s position. Consider for example the proposition expressed by ‘At present nothing is being uttered’. The view that this proposition is self-contradictory “strikes us as strange” not because we are presupposing the philosophical theory of the tenser, but because it clashes with our ordinary linguistic intuitions. Is it not intuitively obvious that the proposition this sentence is ordinarily used to express is not self- contradictory? I think it is obvious, and that the token-reflexive analysis of this and other tensed sentences is at odds with the facts about ordinary usage.

The analysis of tensed sentences by tenseless sentences that includes descriptions of calendrical dates is subject to none of the three arguments developed in the last section. The reason for this is simple: tenseless date-sentences do not in every case express propositions about sentence-tokens, and date-sentences that do not express such propositions can be used as analysans of tensed sentences that do not express propositions about sentence-tokens. ‘The boat is level’, according to this view, is analysed by some such sentence as ‘The boat (is) level on July 20, 1949’. This tenseless sentence ascribes to the event of the boat’s (being) level the B-determination of (being) 1,948 years, 6 months, and over 19 days later than the event of Jesus' birth.

It might be thought that this proposed analysis of tensed sentences fails on the grounds that the date-sentences do not indicate whether the described event is present, past or future, and consequently cannot analyse the tensed sentences, which do so indicate. This is essentially the manner in which Richard Gale tries to refute the date-sentence theory. However, this argument begs the question, for it assumes precisely what the detenser denies --- that tensed sentences refer to events with A-properties.[16] To construct a non-question-begging argument, the tenser must employ premises stating rules of usage of tensed sentences other than the alleged rule that they are used to refer to present, past or future events.

A valid argument against the detenser concerns the truth-condition rules of usage of sentences of these two types. Tensed sentences like ‘The boat is level’ have temporally restricted truth-conditions: they are true if and only if they are uttered simultaneously with the event they purport to designate. However, tenseless date-sentences have temporally unrestricted truth-conditions; ‘The boat (is) level on July 20, 1949’ is true if and only if it is true on every occasion of its utterance. Since one sentence can be an analysans of another sentence only if it has the same truth-conditions as the latter, it follows that the date-sentence is not an analysans of the tensed sentence.

It is also true that anyone token of ‘The boat is level’ has different truth-conditions from any one token of the date-sentence. A token of the tensed sentence is true if and only if it occurs at a time when the boat is level, but whether a token of the date-sentence occurs simultaneously with the boat's being level or at some other time is irrelevant to its truth-value. The tenseless token is true if and only if the boat (is) level on July 20, 1949 --- regardless of when the tenseless token occurs.

I shall now provide four further but related arguments to the effect that tenseless date-sentences (or tokens) cannot be analysantia of tensed sentences (or tokens). These arguments are based on four different theories of meaning. These theories are relevant because each of them shares in common the assumption that one sentence (or token) can analyse another sentence (or token) only if it “has the same meaning” as the latter.

The first argument is really a restatement of the one just given, but in the context of an extensional truth-condition semantics for natural languages. According to Donald Davidson, “to give truth conditions is a way of giving the meaning of a sentence",[17] such that if two sentences have different truth-conditions then they have different meanings. The same applies to any two tokens. It immediately follows from this semantical theory that tensed sentences (or tokens) have different meanings from tenseless date-sentences (or tokens).

A second and closely related argument is based on the model-theoretic semantics developed by David Lewis.[18] Differences in intension, Lewis maintains, give us differences in meaning. An intension is a function which determines how the truth-value of a sentence depends on the possible world in which it is uttered, the time at which it is uttered, and six other factors that I need not mention. It is sufficient to note that tensed sentences have different functions from tenseless date-sentences, for the functions of the former but not of the latter determine the truth-value of the sentence to be dependent upon the B-relation between the utterance of the sentence and the event described by the sentence. Since functions are intensions, and differences in intension determine differences in meaning, it follows that tensed sentences have different meanings from tenseless date-sentences (and the same applies for any two tokens of sentences of these two sorts).

A third argument is based on a verificationist theory of meaning. According to this theory, two sentences or tokens have the same meaning if and only if they are confirmed or disconfirmed to the same degree by observations of the same sort. Consider ‘The boat is level’ as tokened on some occasion. An observation that confirms this tokening as true is an observation that the event of the boat being level occurs simultaneously with the event of the tokening. However, no observation of a similar sort could confirm any token of the tenseless date-sentence. To observe that the boat’s being level is simultaneous with some tokening of ‘The boat (is) level on July 20, 1949’ neither confirms nor disconfirms this token. It is not the observation of this simultaneity relation that confirms this token, but the observation of the boat’s being level on the date described by the token.

A fourth argument is based on what can be termed the equivalence theory of meaning; two sentences or tokens have the same meaning if and only if they are equivalent in the sense that they entail each other and everything entailed by one of them is entailed by the other. Each token of ‘The boat is level’ entails

(26) The boat’s (being) level (is) simultaneous with some token about the boat’s levelness

for each token of ‘The boat is level’ cannot be true unless it is simultaneous with the boat’s being level. However, (26) is not entailed by any token of ‘The boat (is) level on July 20, 1949’, for each token of this sentence could be true even if no tensed or tenseless token about the boat being level occurred simultaneously with this event.[19]

All four of these arguments reveal that tensed sentences and tokens semantically differ from date-sentences and tokens in that the former alone have truth-conditions, verification conditions and entailment relations that involve B-relations between the tokens and the events described by the tokens. These differences make manifest the fact that tensed sentences/tokens do not ascribe to events the very same properties or relations that are ascribed by date-sentences, viz., B-relations to the birth of Christ. They ascribe determinations of a different sort, determinations that require the above-described token-reflexive truth-conditions, verification conditions and entailment relations possessed by tensed sentences and tokens.

It might be thought that these token-reflexive properties of tensed sentences and tokens can be explained by the hypothesis that these linguistic items are token-reflexive in Smart’s sense, i.e., that the only temporal determinations they ascribe are B-relations between themselves and the events they report. This seems like a natural explanation, given that these are token-reflexive properties of tensed sentences and tokens. However, we know from Section I that this explanation is inconsistent with the many non-token-reflexive properties that tensed sentences and tokens also possess. If the token-reflexive properties of tensed sentences and tokens are to be explained, they must be explained by a hypothesis that also accounts for the semantical properties of these sentences and tokens that renders them insusceptible to a token-reflexive analysis of the sort proposed by Smart. I believe there is only one such hypothesis, that these sentences or tokens ascribe A-properties to events.

If tensed sentences/tokens ascribe A-properties, then they have tensed truth-conditions, verification conditions and entailment relations in addition to the above-described tenseless ones, and the tenseless ones can be deduced from the tensed ones. Let us say that a token u₁ of “The boat is level’ is presently true if and only if the levelness of the boat and u₁ both possess the A-property of presentness. This enables u₁'s tenseless token-reflexive truth-conditions to be deduced:

(32) Therefore, the levelness of the boat is observed to (be) simultaneous with u_l.

(33) Therefore, the boat’s (being) level (is) simultaneous with some token about the boat’s levelness.

It is in this manner that the three tenseless semantic differences between tensed and date-sentences/tokens is explained by the assumption that tensed sentences/tokens ascribe A-properties to events.

This assumption implies that the propositions expressed by tensed sentences/tokens refer to past, present or future events, and this implication enables the semantic differences between these sentences/tokens and the token-reflexive sentences/tokens to be explained. We found three such differences:

(34) Tensed sentences alone can express propositions that have been or will be true at times when no language-users exist.

(35) Tensed sentences alone can express propositions that if true are true regardless of whether or not they are expressed.

(36) Tensed sentences alone can express consistent propositions to the effect that I am not uttering anything.

These utterance-independent characteristics of the propositions expressed by tensed sentences are explained by the assumption that these propositions have tensed truth-conditions parallel to those of tensed sentences/tokens. We can assume for example that present tensed propositions are now true if and only if they and the events to which they refer are both present. Thus the proposition P₁, which is expressible by a token of ‘The boat is level’, is presently true if and only if

Notice that (27) and (37) do not separately or jointly entail that some token of ‘The boat is level’ present.[20] By failing to entail this, they allow for the possibility that P₁ is now true but is not now being linguistically expressed. It is tensed truth-conditions of these sorts that enable the three semantic differences between tensed and token-reflexive sentences/tokens to be explained. In terms of the sample sentences we discussed in Section I, we can argue:

(38) The proposition expressed by ‘The preEarth era is present’ is now true if and only if the era and the proposition are both present, was true if and only if the era and the proposition were simultaneously present, and will be true if and only if the era and the proposition will be simultaneously present.

(39) Therefore, the sentence ‘The preEarth era is present’ can express a proposition that was true at a time when no language-users existed.

(40) The proposition expressed by ‘The volcano is erupting’ is now true if and only if the volcano’s eruption and the proposition are both present.

(41) Therefore, the sentence ‘The volcano is erupting’ expresses a proposition that if now true is now true regardless of whether or not it is now expressed.

(42) The proposition expressed by ‘I am not uttering anything’ is now true if and only if my silence and the proposition are both present.

(43) Therefore, the sentence ‘I am not uttering anything’ expresses a proposition that can be true if it is not expressed, and that is thereby consistent.

These tensed truth-conditions of tensed propositions are not only consistent with but help explain the tenseless token-reflexive truth-conditions of the tensed sentence-tokens by which they can be expressed. For example, the tenseless truth-conditions of the token u₁of ‘The boat is level’ are deducible from premises stating the tensed propositional truth-conditions (27) and (37) and the contingent fact that the proposition p₁is now being expressed by u₁:

The assumption that tokens of tensed sentences express propositions that refer to future, present or past events and have tensed truth conditions is the only assumption that is able to explain all the semanticalproperties of tensed sentences and tokens. If we accept this assumption, then we can prove our original thesis, viz., that some items do possess A-properties and hence that time does have an A-aspect. The only additional premise we need is the uncontroversial one that there are true tokens of tensed sentences, for

[1] J.J.C. Smart, “The River of Time”, Mind 58 (1949), p. 492.

[2] Richard Gale, The Language of Time (New York, 1968), p. 199.

[3] Webster’s Seventh New Collegiate Dictionary (Springfield, 1967), p. 199.

[4] J.J.C. Smart, Philosophy and Scientific Realism (New York, 1963), p. 140.

[5] Some defenders of the reality of events with A-properties, such as George Schlesinger, would accept this objection, but others, such as Nicholas Wolterstorff, would reject it. Cf. Schlesinger's Aspects of Time (Indianapolis: Hackett, 1980), ch.3, and Wolterstorff’s “Can Ontology Do Without Events?”, in Essays on the Philosophy of Roderick M. Chisholm, ed. E. Sosa (Amsterdam, 1979), p. 193. I am inclined to reject it, but I do not believe it can be refuted without first establishing that sentences (1) and (2) refer to events with A-properties, which is precisely the point at issue.

[6] Quentin Smith, “The Mind-Independence of Temporal Becoming”, Philosophical Studies 47 (1985), pp. 109-19.

[7] G.E. Moore, “A Defence of Common Sense”, in Philosophical Papers (New York, 1959), pp. 32-59.

[8] Paul Fitzgerald, “Nowness and the Understanding of Time”, Boston Studies in the Philosophy of Science, ed. K.F. Schaffner and R.S. Cohen (Dordrecht, 1974), vol. 20, p. 266.

[9] Schlesinger, Aspects of Time, op. cit., p. 132.

[10] Fitzgerald, “Nowness and the Understanding of Time”, op. cit., pp. 267-8.

[11] Hector-Neri Castañeda, “Indicators and Quasi-Indicators”, American Philosophical Quarterly 4 (1967), pp. 85-100; see especially p. 87.

[12] Hans Reichenbach, Elements of Symbolic Logic (New York, 1947), p. 284.

[13] This theory is succinctly summarized by Brian Loar, who claimed (in a letter to me of 11th May, 1984) that “...an utterance of ‘now’ refers to that moment of time at which it is uttered”. But compare his Mind and Meaning (Cambridge, 1981), pp. 107-13 L Nathan Oaklander has committed himself to this theory and to a theory like Smart’s in his Temporal Relations and Temporal Becoming. (Washington, D.C., 1984), p. 127: “Surely, any expressions in which the indexical word ‘now’ occurs could be replaced by one in which ‘this moment’ or ‘this utterance’ occurs.” For a discussion of the relation between “t” and dates, see Milton Fisk’s “A Pragmatic Account of Tenses”, American Philosophical Quarterly 8 (1971), pp. 93-8, in which this token-reflexive theory is developed in detail.

[14] Hugh Mellor has argued in Real Time (Cambridge, 1981), Ch. 5, that sentences like ‘The boat is level’ and utterances like u₈ are not analysable into sentences like (23), but nevertheless ascribe only temporal determinations of the sort ascribed by sentences like (23), viz., B-relations between utterances and other events. This theory, although not identical with the token-reflexive theories I am criticizing, nevertheless is still subject to my three arguments, for my arguments imply that the relevant tensed sentences and utterances ascribe temporal determinations of a sort not ascribed by tenseless sentences about utterances that are B-related to other events. I have elsewhere argued that MelIor’s theory is self-contradictory. Cf. ‘Problems with the New Tenseless Theory of Time’ (mimeograph, 1985).

[15] Schlesinger, Aspects of Time, op. cit., pp. 110-11 and p.133.

[16] Cf. Richard Gale, The Language of Time, op. cit., p. 55. The circular structure of Gale's argument is analysed in chapter 1 of my Tenses, Temporal lndexicals and Time (mimeograph, 1986).

[17] Donald Davidson, “Truth and Meaning”, in Readings in the Philosophy of Language, ed. Jay F. Rosenberg and Charles Travis (Englewood Cliffs, 1971), p.456. Reprinted from Synthese 17 (1967), pp.304-23.

[18] David Lewis, “General Semantics”, in Semantics of Natural Language, ed. Donald Davidson and Gilbert Harman (Boston, 1972), pp. 169-218.

[19] Aside from these several senses in which tensed sentences/tokens “have a different meaning” than date-sentences/tokens, there is one sense in which each tensed token “has the same meaning” as some date-sentence. This is based on the extensional interchangeability theory of meaning; two linguistic items are regarded as “having the same meaning” in the sense that the proposition expressed by one of them is interchangeable salva veritate with the proposition expressed by the other. For each tensed token occurring at a date D, there is a date-sentence describing a B-relation between the event reported by the tensed token and D, such that the proposition expressed by the date-sentence has the same truth-value as the proposition expressed by the tensed token. For instance, the proposition expressed by ‘The boat (is) level on July 20, 1949’ is interchangeable without loss of truth-value with the proposition expressed on July 20, 1949 by ‘The boat is level’, and the proposition expressed by ‘The boat (is) level on July 21, 1949’ is interchangeable salva veritate with that expressed on July 21, 1949 by ‘The boat is level’. This ‘extensional interchangeability’ account of tensed tokens and date-sentences has been given its most detailed exposition in Clifford E. Williams’ “‘Now’, Extensional Interchangeability, and the Passage of Time”, The Philosophical Forum 5 (1974), pp. 405-23. These semantic facts, however, are insufficient to establish that date-sentences analyse or express the same propositions as tensed tokens, for the “extensional interchangeability” criterion of sameness of meaning is only one of the necessary conditions that must be met for one linguistic item to analyze or express the same proposition as another. The other necessary conditions, sameness of truth-conditions, sameness of verification conditions, and sameness of entailment relations, are not met by tensed tokens and date-sentences, as I have shown in the preceding arguments.

[20] It should be apparent by this time that I believe the nominalist reduction of propositions to sentence-tokens or logical constructions from sentence-tokens to be false; however, I lack the space to defend this belief here.

[21] This paper presents a semantic argument for the existence of items with A-properties; other arguments are also possible. In The Felt Meanings of the World (West Lafayette, 1986), chs. 4 and 6, I presented a phenomenological argument for their existence. The coherency of the idea that items possess A-properties is defended against considerations based on McTaggart’s paradox in my “The Infinite Regress of Temporal Attributions”, The Southern Journal of Philosophy 24 (1986).