Minimalism and Speakers’ Intuitions

Minimalism proposes a semantics that does not account for speakers’ intuitions about the truth conditions of a range of sentences or utterances. Thus, a challenge for this view is to offer an explanation of how its assignment of semantic contents to these sentences is grounded in their use. Such an account was mainly offered by Soames, but also suggested by Cappelen and Lepore. The article criticizes this explanation by presenting four kinds of counterexamples to it, and arrives at the conclusion that minimalism has not successfully answered the above-mentioned challenge.

El minimalismo y las intuiciones de los hablantes

MATÍAS GARIAZZO
Universidad de la República - Uruguay
mgariazzo1@gmail.com

Article received: July 12 2010; accepted: October 28 2010.

---

Abstract

Minimalism proposes a semantics that does not account for speakers' intuitions about the truth conditions of a range of sentences or utterances. Thus, a challenge for this view is to offer an explanation of how its assignment of semantic contents to these sentences is grounded in their use. Such an account was mainly offered by Soames, but also suggested by Cappelen and Lepore. The article criticizes this explanation by presenting four kinds of counterexamples to it, and arrives at the conclusion that minimalism has not successfully answered the above-mentioned challenge.

Keywords: H. Cappelen, E. Lepore, S. Soames, intuitions, minimalism.

---

Resumen

El minimalismo propone una semántica que no da cuenta de las intuiciones de los hablantes acerca de las condiciones de verdad de una gama de proposiciones o enunciados. Por lo tanto, uno de los retos que enfrenta dicha posición es el de brindar una explicación de cómo su atribución de contenidos semánticos a estas proposiciones se fundamenta en el uso. Dicha explicación fue ofrecida principalmente por Soames pero también sugerida por Cappelen y Lepore. El artículo desarrolla una crítica de esta explicación mediante cuatro tipos de contraejemplos y llega a la conclusión de que el minimalismo no logra responder exitosamente al reto que se le plantea.

Palabras clave: H. Cappelen, E. Lepore, S. Soames, intuiciones, minimalismo.

---

The view that philosophers like Cappelen-Lepore (2005a) and Emma Borg (2007) call "minimalism" constitutes an approach to the semantics/pragmatics distinction. An important feature of this view is that it deliberately proposes a semantics that does not account for speakers' intuitions about the truth conditions of a range of sentences or utterances.¹-²Therefore, an important challenge that this view faces is offering an explanation of how its attribution of semantic contents to these sentences is grounded in their use, so that it would not be unwarranted. The main explanation provided so far is that offered by Soames (2002 55-109) but also suggested by Cappelen and Lepore (2005a 2000).³-⁴ The purpose of this essay is to criticize this explanation by presenting four kinds of counterexamples to it. The essay has three sections. In the first one, I briefly characterize minimalism and present the mentioned explanation. In the second one, I criticize this explanation by means of four counterexamples. Finally, in the third section, I make some general remarks about the relevance of the intuitions referred to for a theory of natural language.

I

According to the view that Cappelen and Lepore call "minimalism", every declarative sentence in context semantically expresses a proposition (that is, a content that specifies truth conditions) that is a function of the contents of the constituent parts of the sentence and its syntactic structure. The influence of context in determining this proposition is reduced to the provision of contents to a certain limited number of expressions present in the syntactic surface, typically expressions that are obviously indexical and words or aspects of them that indicate tense (Cappelen and Lepore call the set of expressions that minimalism considers context sensitive "the Basic Set" [2005a 2-4]).⁵ Thus, minimalism accepts the principle of compositionality and restricts in a particular way the intervention of context in the determination of sentences' truth conditions. In the several cases where this view assigns truth conditions that do not account for speakers' intuitions, the strategy used to explain the existence of these intuitions usually consists in distinguishing between the semantic content of the uttered sentence and the content or contents of its utterance, and claiming that speakers do not discriminate between them. Most of the contents of the utterance -the semantic content is usually considered one of these contents- are considered to be only pragmatically communicated.

On the other hand, according to the theories of those that Cappelen and Lepore call "radical contextualists", no sentence in context or utterance semantically expresses a proposition, at most only a propositional scheme.⁶ The truth conditions of a sentence's utterance could be considered to correspond to those of a certain pragmatically communicated proposition usually called "what is said", which would be directly revealed by the speakers' intuitions about the utterance's truth conditions (Recanati 13-20). An intermediate stance is adopted by those whom the above-mentioned authors call "moderate contextualists". According to their theories, although most or all sentences in context semantically express a proposition, there are more expressions that require the intervention of context in order to determine this proposition than those that minimalism acknowledges, such as quantifier expressions, adverbs like "ready" and "sufficient", or verbs like "rain".⁷ We have to add to this classification those theories known as relativistic, which consider that the truth value of certain sentences is obtained by relating their content not only to a world but also to at least one other extra parameter. All these views are different from minimalism in their methodological approach towards speakers' intuitions about truth conditions: while minimalist theories are not concerned with accounting for these intuitions when they assign semantic contents to sentences in context, the other theories consider that these intuitions are at least very important evidence for any truth conditions that could be reasonably assigned.⁸

Soames (2002 55-109) tries to detect the regularity in the use of an expression that reveals the competence conditions for its use and thus also its semantic content. According to him, the proposition semantically expressed by a non-ambiguous sentence free of expressions included in the Basic Set, is the one that speakers communicate and even assert by uttering this sentence in any normal context. The notions of communication and assertion are pragmatic notions by means of which the actual use of an expression would be contemplated: that which a speaker communicates or asserts depends on his intention and on the conversational presuppositions present in the context of utterance. The distinction made by Soames (id . 83-85) between merely communicating, on the one hand, and also asserting, on the other, lies in the nature of the speaker's commitment to the truth of the communicated proposition and on the character of the inference that the audience has to make in order to reach the proposition based on the utterance: in both cases the speaker assumes a certain commitment to defend the truth of the communicated proposition, but in the case of an assertion the commitment is stronger and the inference that the audience has to make is more obvious. In turn, a normal context is one in which a sentence is used literally (that is, among other things, not metaphorically and without sarcasm or irony) and with its conventional linguistic meaning.⁹ On the other hand, a sentence that contains Basic Set expressions and/or ambiguous expressions would express the proposition that is asserted in every normal context in which these expressions receive the same semantic treatment. Thus, the proposition semantically expressed by a sentence would be the minimal information communicated by its utterance in any context and, consequently, it is reasonable to assume that its communication would be required by the speakers' semantic competence.

Given that according to Soames, every time a proposition is asserted all its more obvious entailments are asserted too, the previous criterion for individuating semantic contents does not deliver a unique proposition for each sentence. For this reason, Soames (2002 105-107) supplements his criterion with the requirement that the proposition semantically expressed must explain why its obvious entailments are also always asserted, but not vice versa. Thus, the final account of how the proposition semantically expressed by a sentence is individuated is roughly the following: a proposition is semantically expressed by a sentence if and only if (i) it is asserted in all normal contexts that satisfy the stated requirements concerning Basic Set and ambiguous expressions, and (ii) its assertion explains why any other proposition that satisfies (i) is asserted, but not vice versa.¹⁰

II

As we shall see, with respect to certain sentences whose treatment typically distinguishes minimalism from moderate contextualism, it is clear that the above stated criterion does not allow us to individuate the content that minimalism assigns to them nor any other content. Before considering these cases, it is worth pointing out that, for the sake of argument, the objection I will make in what follows grants minimalism certain points that could be disputed. Soames usually explains the assertion of a minimal proposition claiming that it is entailed by another asserted proposition that the speaker has -in Soames' terms- the primary intention of asserting. It is this proposition that would be mainly responsible for the intuitions we have about the truth conditions of the sentence's utterance in a certain context. But, in the case of many sentences, it could be questioned whether speakers do communicate such a proposition. Moreover, the notion of assertion (or that of communication) according to which every relatively obvious entailment of a primarily asserted (or only communicated) proposition is also asserted (or only communicated) could also be disputed. The cases that I will present below show that even granting these points, the account just presented is unsatisfactory.

The first case consists in a propositional attitude report and was first presented by Soames (2005 361-366), who offers an answer to it that he later abandons together with the minimalist view. In his book, Soames focuses his attention on the semantics of proper names and indexicals, defending the theses of direct reference and semantic innocence. According to the former thesis, a simple sentence containing proper names or indexicals expresses a singular proposition (that is, a structured proposition that contains the individuals denoted by these expressions and not conditions to be satisfied by them)¹¹, and according to the most usual version of the latter, this same proposition composes the more complex ones expressed by attitude ascribing sentences in which the previous simple sentence is embedded. In turn, minimalism accepts, at the same time, the principle of compositionality (which entails that a sentence's context sensitivity is always traceable to an element of its syntax) and considers that in a propositional attitude report of the form "A v that p" there is no context sensitive expression (understanding context in such a way that it includes the linguistic context in which an expression occurs) outside the embedded sentence and the initial nominal expression.¹² Thus, from the minimalist view and these theses it follows that co-denoting names and indexicals are substitutable salva veritate in propositional attitude contexts, and that a propositional attitude report which contains such expressions does not have different truth conditions in two contexts in which they preserve their designations. In other words, from these assumptions it follows that in all contexts the semantic content of a propositional attitude report in which names or indexicals occur in the embedded sentence consists of the same relation and relata (the subject of the ascription and certain singular proposition). Thus, the acceptance of minimalism excludes the adoption of a referentialist theory according to which something like the way in which a subject has a certain attitude toward a singular proposition could be relevant to the truth conditions of a report whose embedded sentence contains names or indexicals. According to Soames, this criterion delivers these minimal (and referentialist) contents for this kind of reports.¹³

From such a perspective, Soames considers an utterance of (1) in the following context.

(1) Mary has just learnt that Peter Hempel is Carl Hempel.

In a Princeton philosophy department party, Mary learns that the famous philosopher of science called "Carl Hempel" is a certain man who is at the party and other guests call "Peter Hempel". One of the guests realizes this after a conversation with Mary, and then utters (1) addressing another guest. Two conversational presuppositions between them are that Carl Hempel has two names, "Carl Hempel" and "Peter Hempel", and that Mary is a rational person who does not commit obvious contradictions.

The proposition that would be the semantic content of (1) -namely the proposition that Mary has just learnt that Carl Hempel is identical to himself- is obviously false: Mary always knew that Carl Hempel (=Peter Hempel) is identical to himself, and the fact that she was always willing to assent to an utterance that "Carl Hempel is Carl Hempel" is evidence of this. And, in turn, the members of the conversation know this and, if asked, they would deny that the minimal proposition just mentioned was communicated in the previous context of utterance. Here, it would not only be strange to claim that this proposition has been asserted in spite of this, but also impossible to explain the assertion of the minimal proposition in the way Soames explains it in other cases in which he recognizes that the speaker does not have the primary intention of communicating it. In many cases, Soames claims, the more specific proposition(s) that the speaker asserts entail, in a way that is relatively obvious, the minimal proposition proposed as the semantic content of the uttered sentence, and so this last proposition is also asserted. In uttering (1) the speaker would have the primary intention of asserting and she would assert a proposition different from that which constitutes its semantic content, a proposition which (given the described situation) is true and ascribes to Mary the recent acquisition of a belief in a proposition with certain descriptive content.¹⁴ But in the case under consideration there is no such an entailment relation, since the conversational participants could, without being irrational, consider the minimal proposition an obvious falsehood in spite of considering true any information that they recognize as being asserted by means of the above mentioned utterance of (1).¹⁵

It seems that (1) and other similar sentences are seldom used to assert the proposition that Soames considers as its semantic content. Insofar as the conversational participants assume the rationality of the believer, only certain propositions with descriptive content would be asserted by uttering sentences like (1). Nevertheless, Soames (2002 233) claims that it is possible to assert a proposition by means of an utterance in certain cases where the speaker lacks the intention to communicate it and where neither she nor her audience are conscious of its communication; the assertion would occur in virtue of the interaction of a series of rules that govern the use of language. But this reply has two flaws. In the first place, it is not clear what it is to assert or communicate something that the speaker does not have the intention to communicate¹⁶ and is not transmitted to her audience. And, in the second place, insofar as Soames tries to circumscribe the rules that govern language practice using the notion of assertion, resorting to these rules in order to explain the assertion of the minimal proposition in cases that prima facie constitute counterexamples to his account seems to be an illegitimate strategy.¹⁷

The described phenomenon is not reduced to sentences like the previous one. Consider the usual utterances of a sentence like (2).

(2) The waiter is efficient.

Given the expressions that minimalism considers context sensitive, the description "the waiter" would not be so considered, and given minimalism's acceptance of compositionality all context sensitivity would be syntactically traceable. Thus, as Clapp points out (265-266), given a Russellian analysis of descriptions we would be led to consider that they quantify over the whole universe, in which case (2) would be obviously false in all contexts due to the existence of more than one waiter. It seems clear that speakers hardly ever have -in Soames' terms- the primary intention of asserting by means of an utterance of (2) the obviously false proposition that there is one and only one existent waiter in the universe and he is efficient, but would have instead the primary intention of asserting (presumably) true propositions about particular contextually salient waiters. And again, Soames cannot say here that the minimal proposition is entailed by the more specific proposition(s) that speakers have the primary intention of asserting, since it could happen that the second ones are true while the first one false.¹⁸ Then, he cannot claim that the first one is always asserted.

Sentences that, like (3) and (4), contain quantifier expressions, give rise to the same problem.

(3)There is beer.

(4)There is no beer.

Given the expressions that minimalism considers context sensitive and its acceptance of compositionality, these sentences would be context insensitive. Since there is no contextual restriction on the domain of quantification, they would semantically express, respectively, the proposition that there is beer in the universe and the proposition that there is no beer in the universe. Thus, in all contexts (3) would be true and (4) would be false. According to Soames' view, speakers usually assert by uttering (3) more specific propositions that state that there is beer in a given particular place. Insofar as each of these propositions entails the minimal proposition that there is beer in the universe, we could explain the assertion of the latter proposition by resorting to the assertion of one of the former ones. However, we cannot apply this strategy to the negation of (3), namely (4). It seems clear that speakers do not have -in Soames' terms- the primary intention to assert by uttering (4) the obviously false proposition that there is no beer in the universe. Instead, by uttering (4), a speaker would have the primary intention of asserting a proposition that she considers true, presumably one that states that there is no beer in a certain particular place. Now, in this case there is no entailment relation between any of these more specific propositions and the minimal proposition that justifies the supposition that the last one is always asserted: the simple fact that the minimal proposition is obviously false and the ones that speakers primarily assert are or could be true, shows that the second ones do not entail the first one.

Finally, the same problem arises from sentences like (5) and (6).

(5)John is ready.

(6)John is not ready.

According to minimalism, expressions like "is ready", "is sufficient" and "is tall", do not belong to the set of context sensitive expressions, and consequently -given the acceptance of compositionality- the previous sentences' truth conditions would be contextually invariant. Cappelen and Lepore (2005a 155-175) claim that these expressions respectively express, in all contexts, the properties of being ready simpliciter, being sufficient simpliciter and being tall simpliciter, delegating to the metaphysician the task of ascertaining what these properties exactly are. I shall make some remarks about this thesis advanced by Cappelen and Lepore later on. But before that, I will question the proposal -suggested by several remarks made by Soames (2002)- that these expressions express certain properties that are unspecific in regard to the thing for which someone or something is ready, is sufficient, or is tall. According to this proposal, we could paraphrase the expressions under consideration in the following way: "is ready for something", "is sufficient for something" and "is tall with respect to some comparative class".¹⁹ Now, once this proposal is adopted, (5) and (6) give rise to the same inconvenient that we saw arises from (3) and (4). Given that it is certain that there would always be something for which John is ready, in every context (5) would be true and (6) false.²⁰ For reasons analogous to those presented with regard to (3) and (4), it could be claimed that the minimal proposition assigned to (5) is asserted by its utterances, but we cannot say the same thing about the minimal proposition assigned to (6). And of course, analogous considerations could be made about the application of this proposal to sentences that contain the expressions "is sufficient", "is tall", and others alike.

As I pointed out, Cappelen and Lepore claim that expressions like "is ready", "is sufficient" and "is tall" express simpliciter properties, whose nature is not something that philosophers of language must uncover, but something that should be uncovered by metaphysicians. As Harry Jackman (326) observes, this proposal is intelligible within an externalist conception about the content of an expression. According to such a conception, a speaker can refer to something -and also have the intention to do so- without knowing nearly anything about it, as has been claimed to happen with certain expressions -such as proper names and natural classes' names- that Saul Kripke called rigid designators (1998). Then, it could be thought that just as we can assert a singular proposition without knowing hardly anything about an individual that composes such a proposition, so too can we assert propositions that contain these simpliciter properties, despite ignoring their nature.

But the analogy between the expressions that are typically considered rigid and these predicates cannot be established. The arguments in favor of the externalist treatment of expressions such as names of individuals and natural classes make use of two elements that are absent in the case of the above-mentioned predicates. In the first place, the uncontroversial existence of the entities proposed as these names' contents. In support of the existence of natural classes, we can mention the possibility of increasing our knowledge about them shown by the existence of experts. On the contrary, it does not seem possible to know what exactly the previously mentioned simpliciter properties are. In the second place, as Kripke (71-105) shows, there are modal intuitions about the sentences in which names occur that give support to their externalist treatment. On the contrary, if we assume that the predicates under consideration express simpliciter properties, we will have no intuitions about the truth values in different possible circumstances of sentences like (5) and (6). Given the absence of these two elements in the case of these predicates, it is not possible to see how we could ground in their use the assignment of simpliciter properties as their semantic contents.

As I pointed out, the previous counterexamples to Soames' account belong to the collection of cases whose treatment typically distinguishes minimalism from moderate contextualism. Thus, these counterexamples constitute an argument against minimalism, insofar as Soames' account is the main and more developed one of this view available so far. I did not and will not say anything about the question of whether this explanation individuates semantic contents for other sentences; probably certain moderate contextualists consider that it does, and it is clear that radical contextualists deny that it does for all or most of these other sentences. However, as I will point out in the next section, it is reasonable to infer from the objection just set forth a very general conclusion about the approach that a theory should have towards speakers' intuitions about truth conditions.

III

The minimalist explanation we have considered tries to ground in linguistic practice the assignment of minimal semantic contents in a way that is compatible with the systematic divergence between these contents and the truth conditions that speakers intuitively assign. In order to achieve successful communication, speakers would not need to consciously discriminate the semantically expressed proposition amongst all the asserted propositions, and the intuitions that we have about truth conditions could be due to -and in the case of certain sentences they would always be due to- the assertion of some other(s) proposition(s) different from the one which constitutes the semantic content of the uttered sentence. But the cases just presented show the difficulty of disconnecting, at the same time, the semantic contents assigned to certain sentences from speakers' intuitions about their truth conditions (or about their utterances' truth conditions), and claiming that the propositions that are supposed to be semantically expressed are communicated by their utterances. Thus, this problem faced by minimalism shows the convenience of not systematically divorcing whatever truth conditions we attribute to sentences in context from the truth conditions intuitively assigned by speakers.

This convenience is recognized by all non-minimalist views. Of course, there is disagreement among non-minimalist theories regarding what intuitions speakers have and what treatment they deserve. Radical contextualists consider that insofar as it is not possible to device a truth conditional semantics but only a truth conditional pragmatics for natural languages, these intuitions directly reveal the only truth conditions that can be reasonably assigned to sentences in context. On the other hand, while some moderate contextualists (Stanley 200) seem to think that all these intuitions can be directly accounted for by semantics, others (Ezcurdia) consider that the availability of a truth conditional semantics is revealed by certain systematicity that our intuitions about truth conditions have, which is compatible with the fact that such a semantics does not respect these intuitions in certain circumstances. The analysis of these alternatives, which is necessary for assessing the relevance of speakers' intuitions for a theory of natural language, goes beyond the scope of this essay.

---

¹ Although there is agreement in considering that the minimalist semantic treatment of certain sentences does not account for speakers' intuitions, different kinds of contextualists disagree on whether this treatment in some way respects these intuitions in the case of the rest of the sentences.

² As Recanati observes (14-16), the only way to test speakers' intuitions about the truth conditions of a given sentence or utterance, is to test the intuitions about its truth value in different circumstances of evaluation.

³ Soames (2005 361-366) and Cappelen (Cappelen & Hawthorne) subsequently give up the minimalist view.

⁴ There is at least another explanation that is convincingly criticized by Clapp (256, 260), namely the one due to Borg (2004 84). This account basically consists in claiming that a clause like "'The cat is on the mat' is true if and only if THE CAT IS ON THE MAT", where the sentence in capital letters corresponds to a sentence in the language of thought, indicates which are the truth makers of the sentence "The cat is on the mat". The content of both sentences (that which belongs to the English language and that which belongs to the language of thought) would be the same, and thus we could explain how the attribution of minimal contents to English sentences is grounded in psychological facts. As Clapp shows, this supposed explanation reproduces at the level of the language of thought (granting for the sake of argument that this is not a problematic notion) the problems that we find at the level of natural language.

⁵ Cappelen and Lepore (2005a 1-4) add to this list expressions like "enemy" or "neighbor", and Borg (2007 355-357) rejects the inclusion of any indexical that requires the intervention of pragmatic factors in order to determine its reference.

⁶ Thus, according to radical contextualism, semantics, strictly speaking, deals with the conventional meaning of expression-types, relying on which speakers pragmatically communicate contents that specify truth conditions. According to this view, there is usually no contextual intervention that, because it is specifically required by the linguistic meaning of the uttered sentence, allows us to express or communicate a proposition that could be considered the semantic content of this sentence. Then, for radical contextualists the primary bearers of propositional content are utterances -that is certain intentional acts- not sentences. As Cappelen and Lepore point out (2005b 50), this does not prevent radical contextualists from considering sentences as derived bearers of content.

Minimalism finds important motivation in the fact that it offers a very simple way of preserving compositionality, a preservation that allows for both the development of a formal truth conditional semantics for a natural language and an appealing account of the speakers' ability to produce and comprehend a potentially infinite number of sentences based on a finite number of resources. But insofar as minimalism lacks a satisfactory account of how its proposed attribution of semantic contents to certain sentences is grounded in their use, it can be accused of being a view devised in an ad hoc way in order to obtain the mentioned outcomes. Now, the proponents of minimalism have attempted to offer such an explanation, the most developed of which I will present in what follows.

⁷ Cappelen and Lepore (2005a 8) point out three strategies that moderate contextualists could use in order to explain the context sensitivity of an expression e outside the basic set. The first one consists in assigning to e different contents in different contexts. The second one consists in postulating an expression in the logical form of the sentences that contain e whose content changes across contexts. The last one consists in postulating a constituent of the propositions expressed by the sentences that contain e that is not syntactically articulated (in spite of its presence being required by the content of e) and that varies across contexts.

⁸ Recanati (83-97) adopts another classification of the theories according to which a theory is called "minimalist" insofar as it recognizes the intervention of context in the determination of a sentence's truth conditions only when that intervention is necessary to express a proposition, and a theory is called "contextualist" insofar as it does not conceive in this way the mechanisms of context sensitivity. According to this criterion of classification, certain theories that Cappelen and Lepore call "moderate contextualist" are called "minimalist", namely those that postulate the existence of indexical expressions hidden in deep syntax. I adopt the classification of Cappelen-Lepore and Borg because it discriminates between two distinct attitudes towards speakers' intuitions in which I am interested in this essay. The line of argument used by authors like Stanley (minimalists according to Recanati's classification and moderate contextualists according to Cappelen and Lepore's) to criticize the theories that postulate unarticulated constituents does not consist in claiming that speakers' intuitions about truth conditions are semantically irrelevant, but in trying to show that there are no cases in which we must postulate unarticulated constituents in order to semantically account for these intuitions; there would always be syntactic evidence that shows that the supposedly unarticulated constituents are in fact articulated. These authors usually present their syntactic proofs together with a defense -by means of context shifting and incompleteness arguments, which are based on speakers' intuitions- of the claim that we are in the presence of a context sensitive sentence. As Cappelen and Lepore point out (2005a 18-21, 72-73), if the binding proofs of the existence of a domain variable that cohabits with every noun phrase were correct, they would only prove the existence of the variable, not that its value varies across contexts.

⁹ Cappelen and Lepore (2005a) do not use the notion of normal context.

¹⁰ Soames (2002 105-107) in fact proposes two criteria that according to him circumscribe the same semantic contents for the same sentences. The criterion outlined in the main text is the one that receives most of Soame's attention, presumably because, unlike the other, it would allow him to explain how his proposal of minimal semantic contents is grounded in linguistic use. The alternative criterion, insofar as it is formulated in terms of the notion of semantic convention, cannot give us such an explanation.

¹¹ It is worth pointing out that Soames (2002 110-130) recognizes the existence of names with descriptive content, and so the previously mentioned thesis would concern only those names that he considers "purely referential".

¹² It could be claimed that English attitude reports of the form "A v that p" contain a Basic Set expression in its superficial syntax, namely "that". But it is dubious that this expression is the one which obviously figures as a demonstrative in other English sentences; it seems that it does not even belong to the same grammatical category. Evidence for this is that most of the translations of these sentences to other languages do not exhibit an expression that is syntactically identical to an obviously indexical one. In part for this reason, minimalist theories do not usually consider the expression "that" which occurs in these last English sentences as a demonstrative.

¹³ Soames claims that (2002 131-146), since there is no description associated with a name or indexical in all normal contexts in which these expressions are used with the same designations, the proposition asserted in all these contexts by an utterance of a sentence that contains such expressions is singular.

¹⁴ The proposition primarily asserted could be, for example, the proposition that Mary has just learnt that the man at the party known as "Peter Hempel" is the famous philosopher of science Carl Hempel (Soames 2002 225-227).

¹⁵ Analogous considerations could be made about negations of propositional attitude reports like "Mary does not believe that Peter Hempel is Carl Hempel".

¹⁶ According to Soames (2002 83-84), there is a difference between the information that is communicated (in which the asserted one is included) and that which is imparted (in which we find the communicated information): while the transmission of the first one to the audience always depends on the speaker's intention and the conversational presuppositions, the transmission of the second one does not necessary depend on these things.

¹⁷ Soames could claim that these rules could be demarcated based on those cases where we do not have to appeal to them in order to explain the assertion of the minimal proposition, avoiding in this way the circularity charge. Thus, we could explain in a non-circular way how the semantic contents of the embedded sentence and the main clause in (1), on the one hand, and the compositional rules of the language, on the other, are delimited based on linguistic use, thus justifying the assignment of the minimal content to (1). Of course, developing such an explanation is not an easy task. In particular, it could be questioned that the sentence embedded in (1), namely "Peter Hempel is Carl Hempel", is normally used to assert that Carl Hempel is identical to himself.

¹⁸ In turn, according to what would be a minimalist Fregean analysis of descriptions, (2) would usually express a proposition that obviously lacks a truth value, and so it would not be entailed by those (presumably) true ones that speakers have -in Soames' terms- the primary intention of asserting.

¹⁹ Borg (2007 350-351) explicitly proposes a quantificational analysis of these expressions.

²⁰ I have considered that the negation that occurs in (6) -and also that which occurs in (4)- has wide scope, and in consequence that it can be paraphrased by (6'):
(6') There is no X, such that John is ready for X.
But there is another possible construal of the negation that figures in (6) -and the same could be said about the negation which figures in (4) (Borg 2007 351). According to this construal, this negation has narrow scope, and so (6) could be paraphrased by (6''):
(6'') There is an X, such that John is not ready for X.
If we consider that the negation in (6) has narrow scope, it could be reasonable to claim that the minimal proposition assigned to (6) is asserted, since it would be entailed by the more specific propositions that speakers (primarily) assert by uttering this sentence. But the problem also arises with sentences like "It is not the case that John is ready" and "It is false that John is ready", where a construal in the line of (6'') does not seem equally plausible. And surely, in every context speakers judge that these last sentences and (6) have the same truth value. Moreover, since it is implausible that the expressions "It is the case that" and "It is false that" have narrow scope when prefixed to most sentences, claiming that this is so in the particular case of the sentences where expressions like "is ready" occur seems to be ad hoc.

---

Bibliography

Borg, E. "Minimalism versus Contextualism in Semantics". Context-Sensitivity and Semantic Minimalism: New Essays on Semantics and Pragmatics, Preyer, G. & Peter, G. (eds.). Oxford: Oxford University Press, 2007. 339-359.

Borg, E. Minimalism Semantics. Oxford: Oxford University Press, 2004.

Cappelen, H. & Hawthorne, J. Relativism and Monadic Truth. Oxford: Oxford University Press, 2009.

Cappelen, H. & Lepore, E. Insensitive Semantics: A Defense of Semantic Minimalism and Speech Act Pluralism. Oxford: Blackwell, 2005a.

Cappelen, H. & Lepore, E. "Radical and Moderate Pragmatics: Does Meaning Determine Truth Conditions?". Semantics vs . Pragmatics, Szabo, Z. G. (ed.). Oxford: Clarendon Press, 2005b. 45-71.

Clapp, L. "Minimal (Disagreement about) Semantics". Context-Sensitivity and Semantic Minimalism: New Essays on Semantics and Pragmatics, Preyer, G. & Peter, G. (eds.). Oxford: Oxford University Press, 2007. 251-277.

Ezcurdia, M. "Motivating Moderate Contextualism", Revista Internacional de Filosofía 31/2 (2008): 55-108.

Jackman, H. "Minimalism and Psychological Reality". Context-Sensitivity and Semantic Minimalism: New Essays on Semantics and Pragmatics, Preyer, G. & Peter, G. (eds.). Oxford: Oxford University Press, 2007. 320-338.

Kripke, S. Naming and Necessity [1972]. Cambridge: Harvard University Press, 1998.

Recanati, F. Literal Meaning. Cambridge: Cambridge University Press, 2004.

Soames, S. Beyond Rigidity. The Unfinished Semantic Agenda of Naming and Necessity. Oxford: Oxford University Press, 2002.

Soames, S. "Naming and Asserting". Semantics vs. Pragmatics, Szabo, Z. G. (ed.). Oxford: Oxford University Press, 2005. 356-382.

Stanley, J. "Context and Logical Form", Linguistics and Philosophy 23 (2000): 391-434.