A FREGEAN ANSWER TO THE PROBLEM OF THE ESSENTIAL INDEXICAL
According to John Perry, indexical utterances expressing indexical thoughts are essential in the sense that they cannot be translated into eternal sentences in a way that preserves Fregean senses. The aim of this paper is to show how Fregean paraphrases of indexical utterances are possible and why they must be accepted as preserving the sense of the original indexical utterances.
Key words: indexicals, essential indexicals, Fregean thoughts
For Frege, proper names abbreviate definite descriptions. The name ‘Aristotle’ abbreviates definite descriptions like ‘the teacher of Alexander’ and ‘the greatest student of Plato’. These descriptions express senses, that is, modes of presentation (Arten des Gegebenseins) of the proper name's owner. These modes of presentation belong to the complete meanings of singular sentences, which Frege calls thoughts (Gedanken). Fregean thoughts are complete and unchangeable, as are their truth-values. Moreover, this view seems to imply that since indexicals are singular terms and appear as complete parts of complete sentences, they must also have senses. Sentences containing indexicals must furthermore express complete senses, or what we may call indexical thoughts. This perspective is appealing, at least if we reject Frege’s dispensable Platonist understanding of the nature of thought.
Externalists like John Perry think that the semantic content of an indexical should belong to the domain of its reference. They reject the Fregean view that an indexical’s semantic content (meaning) must be its reference’s sense or mode of presentation. The indexical’s sense should be restricted to something like its contextually invariant semantic role, expressible by the rule: ‘the personal pronoun “I” refers to the person who utters this word at the time of its utterance’.
From this externalist perspective, Perry worked out some interesting examples of indexical utterances that seem to challenge the Fregean view. These examples focus on cases of who, when and where. I will consider in detail only his best-known example. Perry recalls that once while shopping in a supermarket he noticed a sugar trail on the floor and decided to follow it back to the responsible person. Therefore, he steered his shopping cart through a maze of supermarket shelves until he finally realized that the sugar was leaking from his own shopping cart. There are two thoughts to be considered here. The first is:
1. Someone is making a mess.
After he discovered that he was the one making the mess, his thought changed to:
2. I am making a mess.
Perry notes that unlike sentence (1), sentence (2) is accompanied by a sudden change of behavior, which shows that the belief expressed in (2) isn’t the same as the belief expressed in (1). Now, suppose we replace sentence (2) with sentence (3): ‘Perry is making a mess’. This does not express precisely the same sense, since this replacement cannot be made in all contexts. It is possible, for example, that Perry is suffering from dementia and has forgotten his own name... In this case, he would not be able to recognize this replacement as legitimate. Of course (2) could be replaced by (4): ‘Perry is making a mess, and I am Perry’, but in this case we would again have the indexical.
Perry offers several similar examples, such as the statement ‘Today is July 4, 1972’. Consider this sentence. As he writes, one cannot replace ‘today’ with ‘Independence Day’. For suppose that the utterance is spoken on July 3. In this case, one cannot replace ‘today’ with the expression ‘Independence Day’ without changing the truth-value of the sentence from false to true, since by definition the sentence ‘Independence Day is July 4’ is necessarily true.
The upshot of his arguments is that we cannot replace an indexical proposition with any non-indexical proposition expressed by an eternal sentence in a way that exactly preserves the same semantic content. According to Perry, this result has serious consequences for the Fregean view of indexicals. If Frege is right and the senses of indexicals are their modes of presentation of objects in indexical utterances, then we should be able to find eternal sentences containing the descriptions expressing the precise semantic content of the indexical. But this is impossible, since we cannot find descriptive translations of the semantic content of the indexical that can explain the behaviour and preserve the same truth-value in all contexts. Consequently, there are no senses of indexicals able to complete a Fregean thought.
The usual Fregean response
Philosophers who accept the Fregean view usually have a standard response to Perry’s objection. They accept Perry’s conclusion that it is impossible to find a completing sense as a descriptive translation of the indexical utterance that preserves its precise original semantic content or meaning. Nevertheless, they think that this conclusion is insufficient to rebut the claim that indexical sentences express Fregean senses, for even if we cannot replace indexicals with definite descriptions with the same completing sense, this does not imply that the semantic content of an indexical isn’t a Fregean sense – a mode of presentation of the given reference. It is surely possible that the indexical utterance denotes by means of a mode of presentation without our being able to translate it into the form of a definite description.
Perry’s answer to this way of getting around the problem could be that if we admit that propositions (indexical Fregean thoughts) can be limited to the speaker’s spatio-temporal perspective, they turn out to have limited accessibility. But in this case, we would have to concede that the world is home to a myriad of incommensurable private perspectives, so that there is no common world… However, it seems that if this charge were true, then it would apply not only against the Fregean view, but even against Perry’s own conclusion that ways of believing are not translatable, since it also seems to imply that they are accessible only to speakers and cannot be communicatively shared.
Reconsidering the Fregean view
Before we surrender to Perry’s view, it would be good to reconsider Frege’s remarks about indexical utterances and see how much can be done to rescue contextual elements with completing sentences. Thus, he writes that the thought expressed by the utterance
(5) This tree is covered with green leaves.
must include, in order to preserve its truth-value, the time of the utterance as belonging to its expression. For if this sentence alone were the expression of the thought, the same thought would turn out to be false in the coming winter, after the tree had lost its leaves. If the time of the utterance belongs to the expression of the thought, the place of utterance should also belong to it. So, the more natural proposal seems to be to construct an eternal sentence that makes explicit the spatio-temporal location, such as the following:
(6) The tree located before the front door of Frege’s home in Jena at 9:20 o’clock a.m. on May 23, 1914 is covered with green leaves.
Indeed, the eternal sentence (6) expresses a Fregean thought eternally true or false. However, there is a problem with paraphrases like (6), for there are an infinite number of different ways to make them. Consider the following two:
(7) The tree located three meters in front of Frege when he stood at the front door of his home in Jena at 9:20 o’clock a.m. on May 23, 1914 was covered with green leaves.
(8) The only tree located twenty-two meters south of the final east corner of Weigelstraße in Jena at 9:20 o’clock a.m. on May 23, 1914 was covered with green leaves.
Although these sentences denote the same tree, each of them refers to it by means of a different sense, a different mode of presentation, a different thought. However, what we need is not to multiply the senses, but to develop a strategy to find the thought that preserves the sense of the original indexical thought, if there is one. Moreover, paraphrases like these completely lose the speaker’s original egocentric perspective. Frege wrote that this perspective could be retained as part of the thought not only for the utterance ‘I feel cold’, but also for ‘This rose is beautiful’; and for him the persons who utter these sentences must belong to the contextual elements expressing their thoughts.
Assuming that in preserving the speaker’s egocentric perspective we preserve the mode of presentation of the indexical thought, we can ask if there is a way to preserve such a perspective in the eternal sentence. I think there is a way to do this that is somewhat like taking a plant from a garden, and not just the plant, but the whole clump of earth enclosing its roots, and transplanting it into a vase in a public room. In literal terms: instead of attempting to retain the egocentric perspective expressing it in the form of an entire third person eternal sentence – which is impossible – we should place the whole indexical sentence within an eternal sentence that contains the original context of utterance. Moreover, I suggest we do this in such a way that the original sentence loses its assertive force, in this way sacrificing its original indexical-referential function. Here is how such a relocation of the indexical content could be applied to sentence (5):
(9) At 9:20 o’clock in the morning of May 23, 1914, standing at the front door of his home in Jena, Frege points to the green tree in front of him and says that the tree is covered with green leaves.
A first comment about (9) is that it is an eternal sentence within which the content of sentence (6) is completely expressed without being referentially bound by means of indexicals to the original state of affairs. Here the original indexical sentence is replaced by the non-indexical complementary ‘that-sentence’ reporting what Frege says. Clearly, this latter sentence not only lacks indexicality, it also lacks an indexical reference, since it has what Frege would call an indirect reference: it refers to its own sense. (We could say, in a non-Platonist manner, that it refers to qualitatively identical spoken thoughts that any speaker of (9) would think that Frege has thought).
A second comment concerns the verb ‘to point’ in (9). Since we are reporting that Frege is pointing, we are interested in a description of Frege’s act of pointing in an already linguistically given spatio-temporal location, and not in the original indexical reference proper to his act.
A further issue is that the truth-value of the subordinate clause belonging to (9) is dependent on the conceptual contents asserted in the remaining part of this eternal sentence, since these contents representatively repeat the truth-conditions surrounding the indexical utterance.
A final and obvious remark is that in paraphrases like (9) the thought expressed by the eternal sentence is always regarded as true. Even if the indexical sentence were false, the whole sentence would still be true. So, suppose that Frege incorrectly said, ‘This tree is covered with yellow leaves’. In this case, the whole eternal sentence would be:
(10) At 9:20 o’clock in the morning on May 23, 1914, standing at the front door of his home in Jena, Frege points to a green tree in front of him and says that the tree is covered with yellow leaves.
Sentence (10) remains true, even if the thought expressed by the subordinate clause encapsulated in the eternal sentence is in itself false, since this clause correctly states what Frege said, notwithstanding the falsity of what he said. This is confirmed by the truth-conditions for its negation asserted in the first part of the eternal sentence.
Reconsidering Perry’s examples
We can relocate the indexicals of all of Perry’s examples in similar ways. Sentence (2), ‘I am making a mess’, can be translated as:
(11) At 10 a.m. on March 26, 1968 in the confectionary supplies section of the Fleuty supermarket in the city of Berkeley, CA, Perry notices a sugar trail stretching outward from his shopping cart and perceives that it is he himself who has made a mess.
Here again we have an eternal sentence that includes in itself a sentence expressing a Fregean thought that is considered true. Moreover, the included indexical sentence will be regarded as true, insofar as it agrees with its truth conditions, as asserted in the eternal sentence. And, finally, the eternal sentence seems to explain Perry’s behavior, at least for us, at the time when he noticed that he was making a mess.
The same procedure can be applied to Perry’s other examples, such as:
(12) This is the trail leading up to Mount Tallac.
(13) Now I need to go to the meeting.
(14) Today is the 4th of July.
(15) Yesterday was a beautiful day.
(16) I wrote the Treatise.
The paraphrases could be respectively:
(17) In the afternoon of December 18, 1932, after staying at Gilmore Lake and hiking in the mountains near Lake Tahoe, the author of The Hikers’ Guide to Desolation Wilderness finds the Mt. Tallac trail and thinks that it must be the Mt. Tallac trail.
(18) At noon on August 2, 1972, the time of his meeting in the Department of Philosophy, Perry is sitting outside the Department of Philosophy of UC Berkeley and states that at this moment he has to go to the meeting.
(19) On the morning of the 3rd of July 1972, in Berkeley, CA, Perry states that today is the 4th of July.
(20) On February 13, 1793, waking up after sleeping for twenty years, and intending to refer to the day he fell asleep (believing that he has slept for only one night), Rip Van Winkle says that the day before was a beautiful day.
(21) On the afternoon of October 3, 1775, Heimson, who is a bit crazy and imagines himself to be David Hume, says to himself in his study in Edinburgh: ‘I wrote the Treatise of Human Nature’.
Although all the thoughts expressed by sentences (17) to (21) are true, only the thoughts expressed by the subordinate clauses in the eternal sentences in (17) and (18) are in themselves true, while the thoughts expressed by the subordinate clauses in the eternal sentences (19) to (21) are in themselves false. This is implied by the truth conditions affirmed in the remaining parts of these eternal sentences.
First objection: change in content
Now I will address some objections. The first and most obvious is that the proposed eternal sentences clearly do not translate the original indexical utterances: They do not express the same thought. For example, suppose that Perry has forgotten his name and does not know that his name is ‘Perry’... In this context, he would not be able to see that sentence (11) is true, even if when uttering it he may believe that sentence (2) is true. Moreover, he does not need to know that the day of his utterance is March 26, 1968, that he is in the Fleuty supermarket, etc., for he does not need to know such things in order to know the truth of what he perceives. Furthermore, concerning statement (19), Perry surely does not know that today is the 3rd of July, since he intends to tell the truth when he states that today is the 4th of July. Only we, trusting in what is said by statement (19), know that the day referred to is the 3rd and that his belief is false. Indeed, we don’t have a translation, for the senses of the eternal sentences in our paraphrases considerably and inevitably exceed the content of the original indexical utterances.
From this objection, one could conclude that Perry is right, but I do not think he is. Although the procedure of relocating the indexical sense does not give us a translation of an indexical utterance into an eternal sentence, it allows us to preserve the original indexical sense without returning us to the real indexical circumstances, even if more information about the context is added. This is the case even if the indexical utterance reported in the subordinate clause is in itself false, as in examples (10), (19), (20) and (21). A problem would arise only if something relevant in the original sense were missing in the eternal sentence. But this is not the case, and this is why there is nothing wrong with the relocation of the indexical senses.
The proposed relocation preserves the Fregean modes of presentation of indexical utterances, although it unavoidably adds some additional information. The reason for this is that such utterances must be spatio-temporally independent of the speaker’s contextually bounded perspective, being made from a third-person viewpoint. The indexical utterances must be context-bound, while the eternal sentences reporting them must be interpersonally accessible independently of the context. This requires additional linguistic information, repeating the spatio-temporal location and the relevant elements belonging to the context (a stock of information that can always be expanded). This is why they might have a content that is to a greater or lesser extent not at the speaker's disposal as he utters the indexical sentence. However, although the semantic content of the replacing eternal sentence – the complex Fregean sense or thought expressed by it – is wider, it is not different in the sense of having excluded something relevant from the original mode of presentation.
Evidence for the view I am suggesting is that such relocations would typically and naturally be accepted by a cognitively competent speaker if they were adequately explained to him. If someone explained to Frege why eternal sentence (9) can replace (5), and someone explained to Perry why sentence (11) can replace (2), giving adequate justification for these claims, they would agree. Consider example (15) of Rip Van Winkle’s utterance. According to Washington Irving’s story, after discovering the incredible fact that Rip Van Winkle had slept for twenty years, the townspeople agree with the truth of the eternal sentence (20) as replacing what is represented by the indexical sentence (15). In the end, even Van Winkle agrees that this must be its right relocation, able to explain the behaviour of his fellow townspeople and the changes in his own appearance and subsequent behaviour.
In fact, we often remember other people’s indexical utterances by framing them in eternal sentences. Perry’s indexical utterance could be relocated by his friends in a looser way as: ‘Once in a supermarket Perry saw a trail of sugar and followed it, only to discover that it was he himself who was making a mess’. And in a similar way, one could say: ‘In Lars von Trier’s film Melancholia, Justine’s mother ends her speech at the wedding reception with the sentence: “Enjoy it while it lasts.”’ Indexical relocation is a very common procedure indeed.
Second objection: failure to preserve indexical distinctiveness
At this point one could raise a new objection against the proposed relocations of senses. One could claim that something essential to the original content – to the sense – has been lost. The relocations are all made in an impersonal mode, and in this way, the concrete first person access to the content of the utterance has been irretrievably lost.
Our answer is that something was indeed irretrievably lost, but it was irrelevant. It is not the speaker's egocentric perspective, since the relocation allows us to reconstruct his perspective. What is lost is the phenomenal situation experienced by the speaker at the time when he had his indexical thought. We cannot relocate the way Perry sees, hears and feels when he thinks ‘I am making a mess’. This phenomenal element is irretrievably lost when we relocate the indexical thought into an eternal sentence. Only the speaker himself can have access to this phenomenal situation at the time of the utterance or possibly later, with the help of his memory, when he says, for example, ‘Some time ago I made a mess in the supermarket’. Indeed, the phenomenal element cannot be preserved when we refer to indexical utterances by means of a third-person indexical utterance, for example, when someone says, referring to Perry, ‘Some days ago he made a mess in a supermarket’. Now, my claim is that this missing phenomenal element is irrelevant for the communication. It belongs to what Frege called representations (Vorstellungen) and even illuminations (Beleuchtungen) – phenomena that are themselves non-shareable or only shareable insofar as they may be recognized as belonging to our common human nature – and not to what he called ‘thoughts’. The reason why these phenomenal elements cannot be preserved in any non-personal translation into eternal sentences is that they do not belong to what Frege would call the realm of thoughts. They lack the conventional achievability, while the Fregean senses must have some kind of conventionally grounded access, which is a necessary condition for their universal communicability.
Our conclusion is that by showing how indexical senses can be preserved in eternal sentences, we show that their existence only seems to be impossible when we let ourselves be allured by Perry’s views.
 Gottlob Frege, ‘Über Sinn und Bedeutung’, Zeitschrift für Philosophie und philosophische Kritik, NF 100 (1882), note 2.
 See Michael Dummett, The Interpretation of Frege’s Philosophy (London: Duckworth 1981), pp. 85-86. See also J. R. Searle, Intentionality (Cambridge: Cambridge University Press, 1983), p. 228.
 Gottlob Frege, ‘Der Gedanke: eine logische Untersuchung’, Beiträge zur Philosophie des deutschen Idealismus I, 2 (1918), p. 76.
 Gottlob Frege, Nachgelassene Schriften, H. Hermes, F. Kambartel, F. Kaulbach (eds.) (Hamburg: Felix Meiner, 1969), p. 146.
 One could maintain, drawing on Hector-Neri Castañeda, that the quasi-indicator ‘he himself’ goes beyond the thought referred to... According to Castañeda, in a sentence like: (i) ‘The editor of Soul believes that he himself is a millionaire’, the ‘he’ would refer directly to the real self of this editor, since the ‘he’ cannot be replaced here by the description ‘the editor of Soul’ without a change in meaning. But why couldn’t we replace ‘he himself’ with the definite description ‘the own male person’? If we do this in (i) we get: (ii) ‘The own male person, who is the editor of Soul, believes himself to be a millionaire’. This seems to show that even if the ‘he’ compromises us with its reference to a self, this referred-to self does not need to go beyond the limits of its thought. Indeed, how could the ‘he’ of a fictional reference like the one in (iii) ‘The prince knew that he was in love with Rapunzel’ directly refer to a real self? Nevertheless, we fully understand the sense of ‘he’ in this utterance. And we would understand in the same way also in the case in which the prince really existed while Rapunzel were a mere product of his phantasy. See H-N. Castañeda, ‘He: A Study in the Logic of Self-Consciousness’, in Ratio 8 (1966), 130-157.