Skip to navigation – Site map

HomeIssues21PapersAxiological connotation and the s...

Papers

Axiological connotation and the semantic-pragmatic distinction

Anouch Bourmayan

Abstracts

In this paper, I investigate how axiological connotation articulates with the semantic-pragmatic distinction. First, I consider the semantic-pragmatic distinction in terms of encoded vs. non encoded meaning: I argue that one should distinguish between linguistically encoded axiological connotation and culturally encoded axiological connotation. Second, I consider the semantic-pragmatic distinction in terms of truth-conditional vs. non truth-conditional content. I show that both linguistically and culturally encoded axiological connotation should be analyzed as presupposed content, a kind of content that is not strictly speaking truth-conditional, in the sense that it is not part of the proposition expressed. Finally, in the last section, I embrace a strong conception of pragmatics, according to which there are top-down pragmatic processes, that is, contextual processes involved in the determining of truth-conditional content that are not linguistically mandated but take place for purely communicative reasons. I contend that pragmatic modulation, namely pragmatic modification of a word’s linguistic meaning, may apply to culturally encoded axiological connotation. But I argue that the situation is trickier for linguistically encoded axiological connotation: the initially encoded axiological connotation may only be supplemented with another distinct axiological connotation that comes and superimposes itself to the former one.

Top of page

Full text

Introduction

1In this paper, I focus on the notion of “axiological connotation” and the way it articulates with the semantic-pragmatic distinction.

  • 1 My translation (as for other quotations from Kerbrat-Orecchioni and other French authors).

2As noted by Kerbrat-Orecchioni [1977: 11-12], the term connotation initially comes from logic. In that field, the connotation of a concept designates its definition (or what one would call today the concept’s comprehension or intension), by contrast with the concept’s denotation, corresponding to the items exemplifying this definition (i.e. the concept’s extension). But the term connotation is understood differently in linguistics. In that field, the denotation of a word encompasses all the information that make it possible to select the referent of the word. The “remaining” pieces of information, namely the meaning components that are expressed by the word but are not relevant to identify the word’s referent, correspond to the connotation. As Kerbrat-Orecchioni [1977: 17] writes: “the contents of connotation can be diverse […] But they share the property of being indifferent to the strict identification of a referent”1. To illustrate the phenomenon, she borrows from Barthes [1964] the example of the name Panzani. The word itself refers to a pasta company that is French. But it also “sounds” Italian and thus evokes “italianity”, that is, some Italian culinary art or more generally Italian art of living. This latter meaning component is not relevant to identify Panzani’s denotation, but it is nevertheless conveyed by the word: it is part of the word’s connotation.

  • 2 For an overview of the field of axiology – or more generally value theory – in philosophy, see amon (...)

3Here, I will specifically focus on the axiological connotation attached to words. By axiological, I mean information pertaining to the value that the speaker assigns to the referent of the word2. The value may be positive or negative, and it may be more or less so. According to Kerbrat-Orecchioni, axiological connotation is a kind of enunciative connotation, for it brings information “not on the referent of the message, but on its speaker” [1977: 104]. Indeed, by assigning a specific value to such or such item, the speaker gives information on their own beliefs and preferences. Beside axiological connotation, Kerbrat-Orecchioni argues that there are two other kinds of enunciative connotations: stylistic connotation and affective connotation. Stylistic connotation conveys information about the social or geographical origin of the speaker; to take an example in English (which Kerbrat-Orecchioni does not do), lift and elevator have the same denotation but distinct stylistic connotations, since the former belongs to British English and the latter to American English. As to affective connotation, it pertains to the emotional involvement of the speaker.

  • 3 The term also displays another specialized meaning that is prominent in linguistics and with which (...)

4Should one clearly distinguish between axiological and affective connotation, as Kerbrat-Orecchioni does? She notices that many authors conflate the two notions, but argues that “one may very well have an emotional attitude with regard to an object on which one does not bear any value judgment, and conversely, consider without emotion an item that one judges” [1977: 110]. Although we can be sensitive to this argument, it can be hard to disentangle the two sorts of connotations for every given word. For example, if we compare wop and Italian, the two words clearly display distinct connotations. Or more precisely, the first term has a negative connotation, while the second term is neutral. But is the connotation conveyed by wop affective or axiological? One may argue that it is first and foremost axiological, as for any slur. But at the same time, wop also seems to convey dislike or even hatred. Furthermore, in the psychological literature, which I will also take into consideration in this paper, one traditionally does not distinguish between affective connotation and axiological connotation. One more generally talks about valence3, defined by the American Psychological Association’s online dictionary as

the subjective value of an event, object, person, or other entity in the life space of the individual. An entity that attracts the individual has positive valence, whereas one that repels has negative valence (https://dictionary.apa.org/​valence, accessed August 31, 2023).

For these two reasons, I will talk about axiological connotation in a broad sense, also encompassing affective value.

5In some linguistic approaches, connotation has been almost completely dismissed as an unsuitable object of study. This in particular is the case in formal linguistics, due to the influence of the philosopher and logician Gottlob Frege. In his seminal 1892 text, Frege famously distinguishes between sense and reference: the latter corresponds to the portion of reality that is selected by the linguistic expression, the former to the mode of presentation that makes it possible to identify the reference. According to Frege, an expression may have a sense and no reference: for example, the proper name Odysseus has a sense, but it is doubtful whether it has a reference since it might only be a fictional character. Also, two expressions may have the same reference and different senses: for instance, the morning star and the evening star both refer to the planet Venus, but they involve distinct modes of presentation of their referent, i.e. distinct senses. However, a given expression has one and only one sense, whatever the time and the person using the expression. This, according to Frege, contrasts with the idea one attaches to a linguistic expression. Interestingly, Frege defines an idea as

an internal image arising from memories of sense impressions which I have had and acts, both internal and external, which I have performed. Such an idea is often saturated with feeling [1960: 59 [1892: 26]].

Thus, ideas in the Fregean terminology may be equated with connotation, since they correspond to additional cognitive material that is not directly relevant to grasp the reference of the expression. Furthermore, Frege insists on the sensitive and affective dimension of ideas, so the parallel between axiological connotation and ideas appears particularly relevant.

6According to Frege, an idea is vague and imprecise: “the clarity of its separate parts varies and oscillates” [1892: 26]. An idea is subjective: “on account of the uncertain connexion of ideas with words, a difference may hold for one person, which another does not find” [1892: 27]. “A painter, a horseman, and a zoologist will probably connect different ideas with the name ‘Bucephalus’” [1892: 26]. An idea is unstable across time: “[T]he same sense is not always connected, even in the same man, with the same idea” [1892: 26]. As a consequence, “in the case of an idea one must, strictly speaking, add to whom it belongs and at what time” [1892: 26]. Also, ideas turn out difficult to communicate:

[t]o the possible differences here belong also the colouring and shading which poetic eloquence seeks to give to the sense. Such colouring and shading are not objective, and must be evoked by each hearer or reader according to the hints of the poet or the speaker. Without some affinity in human ideas art would certainly be impossible; but it can never be exactly determined how far the intentions of the poet are realized” [1892: 27].

7Because of their unclarity, instability, subjectivity, and uncommunicability, ideas should not be taken into account in the study of language, according to Frege. However, this is questionable for at least two reasons. First, one should not go to extremes as Frege does. True, a speaker may project some affective value on a given word depending on their own background and experience in a way that differs from some other speaker. But may words not also convey an affective or axiological attitude that is fully stable and encoded? Second, if we consider axiological connotation as a whole, and not only inasmuch as it is encoded in lexical meaning, why should we dismiss it from the study of communication? Grice has shown that implicatures are worth studying, even if they are implicit and inferred. Likewise, connotations – or in Frege’s terminology, ideas – may turn out to play a significant role in communication, even if they seem unstable and elusive.

8In this paper, I want to explore these various issues, which somehow can all be subsumed under the following question: how does axiological connotation articulate with the semantic-pragmatic distinction? In Section 1., I will consider the semantic-pragmatic distinction in the sense of linguistic meaning being encoded or not. I will identify classes of words that do encode axiological connotation as part of their linguistic meaning, review some work from the field of psychology and raise the following issue: do not all (open-class) words somehow encode an axiological value? Thus, I will distinguish between linguistically encoded axiological connotation and culturally encoded axiological connotation. In Section 2, I will consider the semantic-pragmatic distinction in terms of truth-conditional vs. non truth-conditional content. I will draw on recent work showing that linguistically encoded axiological connotation corresponds to presupposed content, and I will argue that culturally encoded axiological connotation should be analyzed along the same lines. Finally, in Section 3., I will consider the phenomenon of pragmatic modulation, corresponding to a strong conception of pragmatics endorsed in so-called “radical contextualism”: I will argue that culturally encoded axiological connotation can be pragmatically modulated, but I will show that the matter is trickier in the case of linguistically encoded axiological connotation.

1. Axiological connotation and the distinction between conventional and contextual meaning

1.1. A first possible interpretation of the semantics-pragmatics divide: conventional vs. contextual meaning

9In a paper outlining the different interpretations of the semantics-pragmatics divide, Recanati [2006] first mentions the Carnapian view on which semantics concerns the word-world relation, and pragmatics the use of the words. On that view, the conventional, encoded meaning of words is sufficient to study the word-world relation, that is, the denotation of words and of larger linguistic expressions. Pragmatics focuses on the past uses of words that may explain their current conventional meanings, or on current uses of expressions that may give rise to new, derived interpretations, distinct from conventional meaning, as in the case of indirect speech acts (see Searle [1975/1991]) or conversational implicatures (see Grice [1975/1989]). But in this Carnapian approach, denotation can be studied in abstraction of use, for denotation is fully determined by the conventional meaning of words.

10Recanati challenges this approach on the following ground: the word-world relation is not necessarily fully determined by conventional meaning. Indeed, some words like I, now and here are indexical, they conventionally indicate that one should consider certain contextual parameters (the identity of the speaker for I, the time of utterance for now, the place of utterance for here) to grasp their reference. Thus, denotation cannot be studied in abstraction from use. As Recanati puts it, words denote things with respect to contexts of use. Does this imply changing our understanding of the word semantics? Not necessarily, if semantics is understood as the study of conventional, i.e., linguistically encoded meaning. What needs to be updated is the view that truth-conditional content is fully determined by semantics. As to pragmatics, it can be understood as the study of contextual meaning, resulting from how human language is used in social interactions or from what we do with words. In that sense, pragmatics may concern truth-conditional content.

11In the following subsection, I consider whether axiological connotation should be analyzed as semantic or pragmatic content, that is, as encoded or contextual meaning.

1.2. Axiological connotation as linguistically encoded content

12Is axiological connotation encoded in the meaning of words? Is it at least encoded in the meaning of some words? Here, I put forward some examples suggesting that axiological connotation is indeed part of the conventional meaning of some lexical items.

13Over the last decade or so, linguists and philosophers of language have shown a growing interest for so-called thick terms. Cepollaro, Soria Ruiz & Stojanovic [2021] illustrate the phenomenon with the following example:

  • 4 As an anonymous reviewer rightly suggests, one may want to add that Lila does so intentionally, sin (...)

(1a) Lila is cruel.
(1b) Lila inflicts unnecessary suffering4.
(1c) Things or people that inflict unnecessary suffering are bad in virtue of doing so.

14The word cruel featuring in (1a) can be analyzed as a thick term, for it involves two levels of content: a descriptive one, whose contribution is expressed in (1b), and an evaluative one, corresponding to the content in (1c). Other examples of thick terms are words like generous, courageous, brutal or disgusting: the first two words convey a positive evaluative content and the last two words a negative one, beside their respective descriptive content. As emphasized by Cepollaro & Stojanovic [2016], the evaluative component of thick terms might pertain to various domains: moral, like in lewd, brutal, chaste or generous, but also aesthetic, like in balanced and graceful, or even sensory like in painful. How precisely this evaluative content should be analyzed is a controversial topic, to which I will return in Section 2. But for now, let me simply emphasize that such content can typically be considered as axiological connotation, since it corresponds to evaluative content that comes in addition to the core descriptive import of the word.

15Is the axiological connotation of thick terms linguistically encoded? I see two criteria to establish whether a semantic feature is linguistically encoded or not. First, one might wonder whether it features by default, when the word is considered out of context. Let us call this first criterion the default presence criterion. Words like painful, brutal or disgusting undoubtedly sound negative and courageous, generous or graceful undoubtedly positive when considered out of context. However, one might argue that when considering a word allegedly “out of context”, one in fact reconstructs a specific context, which might overcome the default, encoded value of the word. Thus, as a second, maybe more reliable criterion, one may consider whether the semantic feature at play features in every use of the word. Let us call this second criterion the persistence criterion. This actually is the criterion endorsed by Cepollaro & Stojanovic [2016]. If one can find literal uses of the word that do not convey any evaluation at all, this shows that the axiological feature is not really encoded.

16As an example of how this criterion applies to thick terms, Cepollaro & Stojanovic [2016] examine the word athletic: they argue that in some occurrences athletic simply means “related to sports” whereas in other instances the word seems to imply that being related to sports is a good thing. To illustrate this point, they put forward the following quotations from the Corpus of Contemporary American English (COCA):

  • 5 Here and elsewhere (in the other quotations), I am responsible for the use of bold characters.

(2) In the Roman context, the heavy athletic disciplines were wrestling, boxing, and the pankration, a brutal blend of wrestling and boxing5.

(3) I wasn’t a particularly athletic or popular child.

(4) Everybody in cirque is athletic and handsome or beautiful.

17In (2) athletic is axiologically neutral, while in (3) and (4) it bears a positive connotation. Cepollaro & Stojanovic conclude that the positive connotation of athletic is not linguistically encoded in the literal meaning of the word: athletic is not a genuine thick term.

18This example is particularly interesting, for it paradoxically shows that the criterion of persistence should be used with caution. Indeed, another analysis is available: when it pertains to activities, like in (2), athletic conveys no axiological connotation, but when applied to people, it does convey a positive evaluation, like in (3) and (4). In other words, athletic is polysemous and can convey a sense that is axiologically marked and a sense that is axiologically neutral. Other examples found on the Internet seem to confirm this hypothesis:

(5) Boys living with their mothers scored significantly higher in scholastic, athletic and physical domain. (https://www.linguee.fr/​francais-anglais/​search?source=auto&query=athletic+domains, accessed July 24, 2023)

(6) The Medical Payments coverage under our policy does not apply to participants in athletic events. (https://www.facebook.com/​notes/​629592691062220/​?paipv=0&eav=Afb_QzYSOo__tCLkZCj8-s9N1R6Z4IeyOW5xGKXEtHQUsLuUTLrXRtHwFYRcFWgYhi4, accessed July 24, 2023)

(7) I enjoy many sports and travelling, and am very athletic and outgoing. (https://www.linguee.fr/​anglais-francais/​traduction/​i+am+very+athletic.html, accessed July 24, 2023)

(8) Artistic, curious and athletic, Garry is a multi-talented person as impressive in music, martial arts and skiing as he is in tennis. (https://www.linguee.fr/​francais-anglais/​search?source=auto&query=athletic+person, accessed July 24, 2023)

  • 6 As an anonymous reviewer interestingly suggests, another alternative explanation is also available: (...)

19Only in (7) and (8), where athletic applies to people, does the word convey a positive axiological connotation. Note that the translation of athletic in French also seems to support the analysis of athletic in terms of polysemy: when applied to activities, athletic can only be translated into sportif, whereas its translation into athlétique becomes possible when used for people. This actually does not fundamentally challenge the persistence criterion. But it shows that one should use it carefully and possibly apply it to the distinct meanings of a word rather than to the word as a whole to decide whether an evaluative content is lexically encoded or not6.

  • 7 In Bourmayan [2022], I argue that in minimal pairs like brave vs. foolhardy, cautious vs. shy or sp (...)

20Bourmayan [2022] specifically focuses on thick terms involved in minimal pairs like brave vs. foolhardy, cautious vs. shy or sparing vs. stingy, that are referentially similar but convey distinct axiological connotations. These minimal pairs are particularly relevant to identify linguistically encoded axiological content. Indeed, the existence of a word that is referentially similar but does not convey the same connotative content strongly suggests that the axiological connotative content at play is linguistic in nature. For example, since foolhardy is referentially very close to brave, it seems unlikely that its negative connotation arises for purely cultural reasons. Otherwise, brave should similarly be perceived as negative. If brave and foolhardy describe similar behaviors, and only the latter term conveys a negative connotation, this much more likely is a linguistic feature, encoded in the literal meaning of the word7.

  • 8 Opinions on whether an expression like rug rat is only colloquial or also derogatory might slightly (...)

21Interestingly, such minimal pairs, made up of two words that are referential synonyms with one of them exhibiting a negative axiological connotation that the other word is devoid of, could appear among words displaying what Kerbrat-Orecchioni calls “stylistic connotation”, e.g. words like mutt (vs. dog), rug rat, brat (vs. child) or pig (vs. policeman). If one follows Kerbrat-Orecchioni’s [1977] classification outlined above, one might hesitate between the categories of stylistic and axiological connotation to analyze these words. Indeed, they all belong to colloquial language: by using one of them instead of its neutral counterpart, the speaker indirectly claims that they belong to lower classes. But do such words only give us information on the speaker’s sociological positioning, or do they also tell us about how the referent of the word is perceived by the speaker? It does seem that words like mutt, rug rat, brat and pig convey some negative axiological connotation with respect to dogs, children and policemen respectively8. This might not be true for every word belonging to slang. For example, threads (vs. clothes) or mate (vs. friend) do not sound particularly pejorative. But at least for some words, the lowering of the speakers’ social belonging appears intertwined with the lowering of the person or entity referred to. And again, since such words convey an axiologically negative connotation while referentially equivalent words do not, this negative feature cannot result from the way the referent is culturally perceived; it has to be linguistically encoded.

1.3. Axiological connotation as culturally encoded content

22Linguistically encoded axiological connotation concerns a limited number of words. For such terms, the negative feature is not – or not only – induced by the way the referent of the word is culturally perceived; it is first and foremost a linguistic, i.e. formal parameter. However, the referential content of words also matters with respect to axiological connotation, for it helps shaping the way one axiologically perceives a given word. We have seen that some words like foolhardy, stingy or mutt display an axiological connotation that can be considered as linguistically encoded. But do not other words, and potentially all open class words, also convey axiological connotation corresponding to the value that is culturally assigned to the referent? And if such value is broadly shared in the cultural community and hence strongly attached to the referential content of the word, why not talk about encoded axiological connotation? Thus, should we not also consider culturally encoded axiological connotation, beside linguistically encoded axiological connotation?

23Kerbrat-Orrechioni [1977: 109] interestingly calls for “a procedure allowing one to assign to each lexical unit an index corresponding to the average affectivity rate that it potentially involves”. In psychology, such a procedure has actually been explored through the study of valence, which, as we have seen, is an overarching notion encompassing both affectivity and axiology understood in a strict sense. Indeed, in psychology, valence, arousal and dominance are considered to be the primary dimensions of meaning. They are traditionally understood as respectively corresponding to the positiveness or negativeness of the stimulus (valence), to the intensity of emotion provoked by the stimulus (arousal) and to the degree of control exerted by the stimulus (dominance). Bradley & Lang [1999], Warriner et al. [2013] or Mohammad et al. [2018] for example have explored to which extent these three dimensions are attached to words of the English lexicon. Their corpora differ in nature and in size (the largest one is the corpus used by Mohammad et al. [2013]); their methods also somehow diverge (Bradley & Lang [1999] and Warriner et al. [2013] use rating scales for each word, whereas Mohammad et al. [2013] resort to best-worst scaling, i.e., comparison between a given number of words); and their precise results are not fully identical either. But their overall conclusion is the same: participants’ judgements largely converge on the valence of lexical items. Indeed, according to Warriner et al. [2013], judgements about valence give rise to more consistency than ratings of arousal and dominance.

24This axiological value – or connotation – is not purely linguistic. It is clearly strongly influenced by the referential content of the word. For example, as emphasized by Stojanovic [forthcoming], terms like AIDS or leukemia relating to illness or death are among the most negatively valenced ones. In addition to the results of these studies, it can be noted that the axiological value of words may vary depending on cultural changes. Indeed, words like liberal and welfare have become progressively negatively connoted, and this may not be unrelated to changes in political trends and cultural values. Thus, when a word gets attached a positive or negative value depending on its referential content and the way the latter is actually perceived within society, and when this value is widely endorsed across speakers, one can talk about culturally encoded axiological connotation.

25Interestingly, this gives rise to cases in which speakers try to manipulate the perception of a given reality by changing the terminology used to refer to it. For example, the position of cleaner is not highly valued in our societies and one might assume that the corresponding word cleaner is assigned a rather negative or at least not neatly positive valence. Thus, to re-enhance the image attached to the position, the expression cleaning technician is sometimes used instead. Analyzing such attempts to manipulate cultural representations by creating a new expression is trickier than it seems. The word cleaner is initially linguistically neutral, but culturally negatively charged. Thus, the creation of the expression cleaning technician, which de facto refers to the same activity as cleaner, is supposed to shed a more positive light on the profession. How? By the compositional characteristics of the new expression. Indeed, cleaning technician puts forward the notion of competence or technical skill through the presence of the word technician. This supposedly should help re-enhance the cultural perception of the job of cleaner. But is the change in cultural axiological connotation effective? Nothing is less certain. However, such a manipulation of the terminology might have led to a change in linguistic axiological connotation. Indeed, the expression cleaning technician cannot but be considered in articulation – or should we say in contrast – with the word cleaner: choosing to speak of cleaning technician instead of cleaner is a linguistic sign that you want to present the profession in a favorable light. In other words, cleaning technician displays a linguistically encoded positive axiological connotation – unless one considers that it has now become the default, neutral expression, and that cleaner instead is linguistically marked, i.e. that it displays a negative linguistically encoded axiological connotation.

1.4. Axiological connotation, a pervasive phenomenon? A cognitive argument

  • 9 Interestingly, Warriner et al. show that more polarized – i.e. strongly positive or negative – word (...)

26If one admits that there is culturally encoded axiological connotation beside linguistically encoded axiological connotation, then axiological connotation is no longer a phenomenon restricted to a small number of words, but rather a pervasive phenomenon, which may concern any open class word. Of course, in this case, encoding should be considered a gradual phenomenon rather than an all-or-none matter: an axiological connotation will be all the more encoded in the word as its evaluation is consistent among speakers within a linguistic and cultural community9. But is there independent evidence or support for this assumption?

27I believe that there is, or at least that this story is consistent with other data in psychology. As recalled by Barrett & Bar [2009: 1328], affect has been given more or less significance with respect to cognition over the centuries. “For centuries, philosophers have believed that every moment of waking life is to some degree pleasant or unpleasant with some degree of arousal”, so that “affect is a basic ingredient of mental life”. But during the 20th century, affect started being largely ignored. This was true in the behaviorist era in psychology, where “scientists no longer regarded affect as central to perception and thought”. But this also mainly continued at the time of the cognitive revolution starting in the late 1950s, when people were considered mostly as rational actors and affect was assumed to be an epiphenomenon occurring “after perception and in reaction to it”. However, in the 1950s a marginal movement called the New Look proposed to shift the perspective on affect: it argued that affective reaction to a stimulus may precede conscious perception rather than result from it, and may even influence the actual content of the percept. Although the work carried out by proponents of the New Look appears quite controversial and has been highly criticized as suffering from many methodological flaws (see, e.g., Firestone & Scholl [2016]), it certainly paved the way for important studies on the relationship between affective responses and cognitive processes. Among them, Zajonc [1980, 2000] in particular put forward the “affective primacy hypothesis”, according to which affective information may be retrieved with minimal stimulus input, before extensive cognitive processing has occurred. In that view, affective information would be obtained more rapidly and reliably than referential, descriptive information.

28Does that also hold for the processing of lexical information? This seems too early to say, although some psycholinguistic studies already point in that direction (see, e.g., Bargh et al. [1989]). But if true, if the retrieval of affective information indeed precedes the identification of referential content, the hypothesis that virtually any open class word of the lexicon conveys some evaluative information sounds all the more plausible. For in that view, the axiological connotation of a word no longer corresponds to extra, incidental information. Rather, it is a parameter of great significance in lexical processing, which is considered first, whereas referential content is only retrieved in a second step.

2. Axiological connotation and the distinction between truth-conditional and non truth-conditional content

2.1. A second interpretation of the semantics-pragmatics divide: truth-conditional vs. non truth-conditional meaning

29In the preceding section, I explored the semantics-pragmatics divide in terms of conventional vs. non-conventional meaning. However, another way to construe this distinction is in terms of truth-conditional vs. non truth-conditional meaning. As Recanati [2006: 461] puts it, under this reading, “something is considered as semantic to the extent that it concerns the truth-conditions (or more generally, the descriptive content) of the utterance”. The notion of truth-conditional content is often taken for granted, but let us recall that it corresponds to the content that supposedly matters to determine whether the sentence is true or false. For example, everyone agrees that the truth-conditional content of a sentence like (9) involves a contextually determined location, even if there is some debate about how much context is involved in the retrieving of the location constituent (see Recanati [2005]). In other words, one needs to identify the implicit location communicated by a speaker uttering (9) in order to identify the truth-conditional content of the corresponding utterance:

(9) It’s raining.

  • 10 Whether or not metaphorical meaning is part of truth-conditional content is an ongoing debate to wh (...)

30Indeed, truth-conditional content is not an abstract, theoretical notion; it corresponds to a level of content that is genuinely communicated (hence the fact that one also sometimes refers to it as the proposition expressed). Thus, the distinction between the two senses of semantics and pragmatics matters, for as we have already seen above and unlike what Carnap seemed to believe, the distinction between conventional and non conventional content does not fully overlap with the distinction between truth-conditional and non truth-conditional content: some part of the truth-conditional content of utterances – for example the reference of indexicals – corresponds to non conventional meaning, i.e. meaning provided contextually and not by mere decoding of linguistic material. The location constituent in (9) is another example of the phenomenon10.

31When there is doubt about the belonging of a given constituent to truth-conditional content, it is possible to resort to some tests that will help clarify the status of the constituent. The tests most commonly used for that purpose involve logical operators: if a constituent of meaning falls under the scope of a logical operator (such as negation or conditional), this means that the constituent belongs to the truth-conditional content of the utterance (Recanati [1989: 321, 1991: 112] and Carston [2002: 191] dub this methodological tool the Scope Principle). For example, if we consider again the descriptive content of (1a) given in (1b), we can check its truth-conditional status by applying a negation to (1a), such as in (10b):

(1a) Lila is cruel.
(1b) Lila inflicts unnecessary suffering.

(10a) Lila is not cruel.
(10b) Lila does not inflict unnecessary suffering.

If (10a) is true, (10b) in turn has to be true as well. In other words, when negation applies to (1a), the content in (1b) falls into the scope of negation. This shows that the descriptive content in (1b) indeed is part of the truth-conditional content of (1a).

32But what if a constituent of meaning does not pass the negation test or any other test implementing the Scope Principle? There are two main possibilities. First, the constituent may correspond to presupposed content. Second, it may be part of an implicature. Let us consider well-known examples of the two kinds of content. Russell [1905] famously put forward the sentence in (11a), which presupposes the content expressed in (11b):

(11a) The present king of France is bald.
(11b) There is someone who is currently king of France.

When the sentence is negated as in (12a), the propositional content in (11b) strikingly still persists despite the presence of negative markers:

(12a) No, the present king of France is not bald.
(12b) There is someone who is currently king of France.

33Thus, presupposed content does not pass the negation test instantiating the Scope Criterion. Russell’s example involves a presupposition of existence (its propositional content corresponds to the statement that there exists someone corresponding to the description conveyed by the definite noun phrase), but projection, namely the property of not falling into the scope of logical operators, is more generally characteristic of every kind of presupposition.

34If one considers Grice’s [1989: 32 [1975: 51]] famous example of the conversational implicature (13b) conveyed by B’s response to A in (13a), when B’s answer is negated like in (14), the implicature in (13b) likewise remains out of the scope of the negation. Even if B’s utterance in (13a) is false (i.e. (14) is true), (13b) still might be true.

(13a) A: Smith doesn’t seem to have a girlfriend these days.
B: He has been paying a lot of visits to New York lately.

(13b) Smith has a girlfriend in New York.

(14) No, Smith has not been paying a lot of visits to New York lately.

But unlike in the case of presuppositions, (13b) is not necessarily true. Let us say that its status is undecidable: it might be true and it might be false.

  • 11 For an extensive comparison between presuppositions and implicatures, see Moeschler [2018].

35Thus, we have identified two kinds of contents that do not fall into the scope of negation and hence do not belong to truth-conditional content: presuppositions and implicatures. When the initial utterance is negated, the formers remain true, and the latter might be true as well as false. Indeed, presuppositions correspond to background information, i.e. information that is supposed to be taken for granted by both the speaker and the hearer, regardless of whether the sentence is true or not. By contrast, implicatures correspond to information inferred on the basis of truth conditional content as well as contextual clues; thus, they partially depend on the truth of the utterance11. Is axiological connotation part of truth-conditional content? If not, does it correspond to a presupposition or to an implicature?

2.2. Is axiological connotation part of truth-conditional content?

36In Section 1., I showed that the evaluative content of thick terms can typically be considered as axiological connotation. But what precisely is the status of such semantic content? Following Gibbard [1992], Jeshion [2013], Richard [2008] and Väyrynen [2009, 2013], Cepollaro & Stojanovic [2016] argue that the evaluative content of thick terms is presupposed, for it projects, that is, it resists when the predicate is embedded under negation, questions, conditionals or modals. They give the following example:

(15a) Madonna’s show is lewd.
(15b) Madonna’s show is sexually explicit.
(15c) Things that are sexually explicit are bad, in some relevant respect, because of being so.

(16a) Madonna’s show isn’t lewd.
(16b) Is Madonna’s show lewd?
(16c) If Madonna show is lewd, I wouldn’t like to go.
(16d) Madonna’s show might be lewd.

  • 12 Of course this excludes metalinguistic uses of lewd, where what is targeted by the operator precise (...)
  • 13 Although such data appear as knock-down arguments in favor of a presuppositional account, it should (...)

The axiological connotation of the thick term lewd can be made explicit with (15c). In other words, (15c) is an evaluative content (a negative one) communicated by an utterance like (15a). But interestingly, the content in (15c) resists when (15a) is embedded under operators like negation (16a), interrogation (16b), conditionals (16c) and modals (16d): the sentences in (16a)-(16d) convey the content in (15c) just as (15a) does. By contrast, the descriptive content of (15a) made explicit in (15b) does not resist when embedded under logical operators: it is no longer communicated by (16a)-(16d)12. This is evidence that the descriptive content of (15a) in (15b) is part of the truth-conditional content of (15a), while its evaluative content in (15c) is not. And since (15c) necessarily remains true in the contexts of (16a)-(16d), this also shows that the evaluative content of (15a) corresponds to presupposed content rather than to an implicature. It operates as common knowledge allegedly endorsed by both the speaker and the hearer, regardless of the very truth-conditional content that is under discussion13.

  • 14 As we have just seen in footnote 12., a speaker may distance himself from this value system by usin (...)

37Notice that when studying the evaluative content of (15a), we have focused on (15c), which corresponds to what one may call “a general evaluative content”. This content is fully propositional and makes explicit a value system conveyed by the term lewd, namely that any sexually explicit sign or behavior is to be judged negatively14. This content does not take into account the specific entity the adjective applies to; or rather, it encompasses it in a more general statement. However, one might emphasize that uttering (15a) also boils down to communicating the “specific evaluative content” expressed in (15d):

(15d) Madonna’s show is bad, because it is sexually explicit.

  • 15 (15d) is not an implicature. It would be if it were inferred on the basis of truth-conditional cont (...)

38How does (15d) originates? (15d) is a logical implication of both the descriptive, truth-conditional content of (15a), namely (15b), and the general evaluative presupposed content of (15a), (15c)15:

(15a) Madonna’s show is lewd. [initial sentence]
(15b) Madonna’s show is sexually explicit. [truth-conditional descriptive content]
(15c) Things that are sexually explicit are bad, in some relevant respect, because of being so. [presupposed general evaluative content]
(15d) Madonna’s show is bad, because it is sexually explicit. [logically implied specific evaluative content]

We have seen that when negated like in (16a), the general evaluative content in (15c) resists. By contrast, interestingly, (15d) does fall into the scope of negation in (16a): the specific negative content about Madonna’s show conveyed by the word lewd in (15a) disappears. Should we straightforwardly conclude that the evaluative content of lewd is truth-conditional? When asserting that Madonna’s show is not lewd, one still presupposes that sexually explicit things are bad, but one no longer contends that the show is bad because of its sexual explicitness. What is at issue is not the value system lexically conveyed by lewd, but the objective characteristics of Madonna’s show regarding its sexual explicitness: (15a) says that the show is sufficiently sexually explicit to be considered bad, (16a) says it is not. Would there not be the initial general evaluative consideration corresponding to (15c) presupposed by (15a) because of the presence of lewd, (15a) would not entail the specific evaluative comment corresponding to (15d) about Madonna’s show. And whereas this entailment disappears in (16a) because of the descriptive content (15b) being negated, the general evaluative content expressed in (15c) remains. Thus, in the rest of this paper, I will continue talking about presupposed evaluative content. But it should remain clear that this concerns the general evaluative content – or axiological connotation – attached to words, and not the more specific evaluative content that might be entailed by the use of these words in a specific context.

39In the preceding section, I distinguished two kinds of encoded axiological content: linguistically encoded axiological content and culturally encoded axiological content. Now, for the sake of the argument, I would like to check with paradigmatic examples of the two categories whether both kinds of axiological content are indeed presuppositional. First, let us consider the expression rug rat which, as we have seen above, displays a linguistically encoded axiological connotation, articulated in (19) and communicated for instance by a sentence like (17). Is the evaluative content in (19) also conveyed by the sentences in (18a)-(18d), which respectively result from applying the negation, the interrogation, a conditional and a modal operator to (17)?

(17) Samy is a rug rat.

(18a) Samy isn’t a rug rat.
(18b) Is Samy a rug rat?
(18c) If Samy is a rug rat, he is young.
(18d) Samy might be a rug rat.

(19) Children are nasty, unpleasant, annoying.

  • 16 Again, this excludes metalinguistic uses of rug rat, where what is targeted by the operator precise (...)

It seems to be the case: the sentences in (18a)-(18d) all convey the axiological content in (19), despite the presence of operators16. But if the semantic content of an expression like rug rat projects, this strongly suggests that linguistically encoded axiological connotation is presuppositional in nature.

40Let us now consider a word like welfare which, as I have pointed out, has become progressively negatively connoted. Let us assume that this negative value is well entrenched in the linguistic community. If so, a sentence like (20) will convey the semantic content in (22), roughly corresponding to the axiological connotation of welfare state. But does this content projects? Is it still communicated when the sentence in (20) is targeted by operators like negation (21a), interrogation (21b), a conditional (21c) or a modal (21d)?

(20) This program is typical of the welfare state.

(21a) This program isn’t typical of the welfare state.
(21b) Is this program typical of the welfare state?
(21c) If this program is typical of the welfare state, it might be debatable.
(21d) This program might be typical of the welfare state.

(22) A system whereby the state plays a key role in the protection and promotion of the economic and social well-being of citizens, especially those in financial or social need, is negative for society.

Again, the evaluative content of welfare state does project since it is still communicated when (21a)-(21d) are respectively uttered.

41To sum up, both linguistically encoded axiological connotation and culturally encoded axiological connotation remain true when the predicate they attach to falls into the scope of a logical operator. This strongly suggests that they do not contribute to truth-conditional content. Rather, they correspond to presupposed content, i.e., content presented as already established and endorsed by participants, and hence not under discussion.

2.3. Is axiological connotation a “lower-level explicature”?

42In this subsection, I consider a level of content that has been dubbed by relevance theorists higher-order or higher-level explicature. I compare it to axiological connotation: does the parallel hold between the two kinds of content? And if so, to which extent?

43In the glossary of Thoughts and Utterances [2002], Carston defines higher-level explicature in the following way:

a particular kind of explicature [...] which involves embedding the propositional form of the utterance or one of its constituent propositional forms under a higher-level description such as a speech-act description, a propositional attitude description or some other comment on the embedded proposition.

What relevance theorists call explicature is a proposition that is inferentially developed from the encoded content of the utterance (relevance theorists talk about the logical form) and actually communicated. The term explicature by default refers to the basic level explicature, corresponding to what we have called so far truth-conditional content or proposition expressed. But relevance theorists argue that the basic explicature does not exhaust the full propositional content actually communicated on the basis of a given encoded representation. According to them, full propositional content involves further levels of content embedding the basic explicature such as speech act descriptions or propositional attitude descriptions. Thus, a given utterance may communicate several explicatures embedded in one another: the basic explicature and higher-level explicatures. Wilson & Sperber [2004] put forward the following example:

(23a) Peter: Will you pay back the money by Tuesday?
(23b) Mary: I will pay it back by then.

Mary’s utterance conveys the basic explicature in (24a), the higher-level explicature in (24b) embedding (24a) under a speech-act description, and the higher-level explicature in (24c) embedding (24a) under a propositional-attitude description.

(24a) Mary will pay back the money by Tuesday.
(24b) Mary is promising to pay back the money by Tuesday.
(24c) Mary believes she will pay back the money by Tuesday.

44This way of analyzing implicit speech acts or propositional attitudes as being somehow on a par with truth-conditional content, as being mere expansions of it recovered by a similar mixture of decoding and inference, is quite unusual in the field of pragmatics. However, it rightly emphasizes that such elements of meaning have nothing to do with conversational implicatures, that they do not correspond to autonomous propositions distinct from truth-conditional content. Rather, they directly apply to truth-conditional content. And it is precisely on the basis of such fully developed propositional contents that hearers in turn will be able to derive the relevant implicatures: if a hearer misses the right speech act or propositional attitude conveyed by a speaker with a given utterance, they might simply be unable to identify the right conversational implicatures.

45But if every “basic explicature”, every “proposition expressed” can be embedded under higher-level descriptions such as a speech act description or a propositional attitude description, we have also seen in Section 1.4. that axiological connotation may be a more pervasive phenomenon than what is generally assumed, pertaining to other words than those that are usually dubbed as expressives. Thus, is axiological connotation in some way or other a replication of higher-level descriptions at the lexical level? Can axiological connotation be viewed as a kind of lower-level explicature? The parallel may not be entirely satisfactory, but it still might be worth exploring the similarities and differences of the two phenomena in order to better circumscribe axiological connotation.

46First, higher-level descriptions and axiological connotation differ in that they do not apply to the same kind of content. Higher-level descriptions bear on full propositions and in turn give rise to new – extended – propositions. For example, the speech-act description Mary is promising to applies to the propositional content in (24a), which corresponds to the basic explicature conveyed by (23b), and gives rise to the new full proposition – or higher-level explicature – in (24b).

(23b) Mary: I will pay it back by then.

(24a) Mary will pay back the money by Tuesday.
(24b) Mary is promising to pay back the money by Tuesday.

47In contrast, axiological connotation applies to a content that is not fully propositional. Whether it ultimately gives back a full proposition is somehow debatable. So far I have proceeded as if it did, but this was mostly for the sake of convenience, in order to make explicit and help clarifying the content at play. Indeed, in the case of linguistically encoded axiological connotation, when another term is available that is axiologically neutral, like dog for mutt or child for rug rat, the propositional content of the word’s axiological connotation can be phrased as applying to the neutral term. For example, the axiological content of rug rat used in (17) can be articulated such as in (19), which involves the word child:

(17) Samy is a rug rat.

(19) Children are nasty, unpleasant, annoying.

48In the case of culturally encoded axiological connotation, or when no corresponding neutral term is available, the axiological connotation’s propositional content must be articulated as targeting the word itself. Thus, the axiological connotation of AIDS, used in (25a), can be phrased like in (25b):

(25a) John has AIDS.
(25b) AIDS is terrifying.

49However, explicating axiological content in a full propositional form does not mean that this content is necessarily communicated as such. This is compatible as well with a view in which axiological content is non-conceptual.

50Also, another significant difference between the two phenomena, somehow linked to the former point, is that higher-level explicatures correspond to a development of the proposition expressed or basic explicature, whereas axiological connotation gives rise to a new proposition, distinct from the proposition expressed. For example, (24b), the higher-level explicature of (23b), is an expanded version of (24a), the proposition expressed in (23b). By contrast, (19), the axiological connotation of rug rat, corresponds to a proposition distinct from (17). And the same is true for (25). The propositional content of AIDS’s axiological connotation, given in (25b), is not a development of the proposition in (25a); it is a separate, autonomous proposition.

51Another, at first view striking difference between higher-level explicatures and axiological connotation is their more or less subjective nature. Higher-level explicatures are always related to the speaker, as shown for example in (24b) and (24c), two higher-level explicatures of (23):

(23b) Mary: I will pay it back by then.

(24b) Mary is promising to pay back the money by Tuesday.
(24c) Mary believes she will pay back the money by Tuesday.

52By contrast, axiological connotation is usually phrased in general terms, as appears in (19), the axiological connotation of rug rat:

(17) Samy is a rug rat.

(19) Children are nasty, unpleasant, annoying.

53So can we claim that higher-level explicatures relate the proposition expressed to the subjectivity of the speaker, whereas axiological connotation presents an evaluative statement as objective and generally endorsed? If we consider the propositional content of a given assertion such as (23b), why not say that Mary is not only asserting that she believes the content of (23b), but also presenting it as a fact that is objectively true and should be endorsed by anyone? Conversely, when someone uses the term rug rat, she might be claiming that children are annoying. But why could that assessment not be restricted to the speaker alone? Why would the speaker of (17) not be simply claiming that they believe children are annoying? Anyone subsequently using the term rug rat would show that they have a similar view on children, whereas other speakers using instead the word child would show that they does not endorse this negative evaluation.

54Finally, I showed that axiological connotation can be analyzed as presupposed content. But is this also true for higher-level descriptions? Let us consider again the sentence in (17). When uttered, (17) may convey the higher-level explicature in (26), where the speech act description has been made explicit:

(17) Samy is a rug rat.

(26) The speaker believes that Samy is a rug rat.

55But let us consider again (18a)-(18d), corresponding to (17) embedded under different operators:

(18a) Samy isn’t a rug rat.
(18b) Is Samy a rug rat?
(18c) If Samy is a rug rat, he is young.
(18d) Samy might be a rug rat.

56In these contexts, the proposition in (26) does not project, i.e., it does not stay true. Thus, a higher-level explicature like (26) cannot be analyzed as presupposed content. Also, it cannot be an implicature. Indeed, when (17) is negated like in (18a), (26) can no longer be true. But as we have seen in Section 2.1., this would not be the case if (26) was an implicature: (18a) would still be compatible with some contexts in which (26) is true. Thus, should a higher-level explicature like (26) be analyzed as truth-conditional content? The matter is trickier than it seems. For implicit speech-act descriptions or propositional attitude descriptions embed the proposition expressed rather than they belong to it. Indeed, it seems that the operators in (18) directly modify higher-level descriptions. For example, the negation in (18a) gives rise to a new higher-level explicature corresponding to (27a), and the interrogation in (18b) to the higher-level explicature in (27b), which both differ from (26) regarding the speech act involved:

(27a) The speaker disbelieves that Samy is a rug rat.
(27b) The speaker wonders whether Samy is a rug rat.

57Thus, obviously, higher-level explicatures cannot be treated as standard truth-conditional content either.

58Indeed, higher-level explicatures and axiological connotation also seem to display some similarities. Obviously, both kinds of content can be said to have an expressive dimension, in the sense that they both convey information on attitudes, feelings or emotions of the speaker. But do the two kinds of content display the three properties identified as characteristic of expressives by Wharton [2016], who follows Potts [2007]? First, Wharton [2016] argues that expressives correspond to non truth-conditional content and contribute to speaker meaning in a manner that is somehow independent from the utterance in which they appear. This is true for axiological connotation, which corresponds to non truth-conditional, presupposed content (see Section 2.2.). And as I have just shown, this is true as well for higher-level explicatures, which are not presuppositions but cannot either be considered as part of the proposition expressed.

59Second, Wharton [2016] emphasizes the “descriptive ineffability” of expressives, a point that is central in Pott’s [2007] account of expressives. On this view, the meaning of expressives is hard to pin down in conceptual terms. In Pott’s [2007: 166] own terms, “speakers are never fully satisfied when they paraphrase expressive content using descriptive, i.e., nonexpressive, terms”. Is that true for higher-level explicatures and axiological connotation respectively? Although the difficulty might not be as striking as in the case of an expressive like damn, I believe that it does concerns higher-level explicatures and axiological connotation as well. In the preceding sections of this paper, I tried to make explicit the axiological connotation of different expressions, like welfare state or AIDS. However, I am fully aware that the result is not fully satisfying, in the sense that other people would potentially explicate such contents in a – more or less – different way. As to higher-level explicatures, it seems that there is some leeway as well in how they should be phrased. For example, regarding (23b), should one say that Mary asserts, claims, confirms, contends, argues... that she will pay back the money? All of these descriptions could be right, but none of them is perfectly satisfying.

60Finally, Wharton [2016: 21] argues that “expressive meaning is often conveyed by non-linguistic means” and “even linguistic expressives […] have a flavour of the non linguistic about them”. On that view, linguistic expressives would function like facial expressions, tone of voice or other physical signals that help a “speaker” communicate specific emotions or attitudes. Is this true for higher-level explicatures and axiological connotation? I believe that it is, for both kinds of content might be made more accessible to the hearer via signals like facial expressions or the tone of voice used by the speaker.

61Thus, higher-level explicatures and axiological connotation do share significant features with expressives as defined by Potts [2007] or Wharton [2016]. But interestingly, according to the latter, who follows Blakemore [2013: 3544], expressives “play a role in the communication of a speaker’s emotions by corresponding to a procedure for retrieving representations [of emotional states] rather than a constituent of a propositional representation”; thus, since interjections express attitudes, someone hearing (28a) might embed the proposition [I am here] under an attitudinal description, giving rise to the higher-level explicature in (28b):

(28a) Aha! You’re here.
(28b) The speaker is surprised that I am here.

62The higher-level explicature, although expressed in a propositional form, has still been retrieved through a procedure, and it seems probable that axiological connotations attached to lexemes are also conveyed through a procedure making available representations of emotional states.

3. Axiological connotation and top-down pragmatic processes

3.1. Top-down processes in the determining of truth-conditional content

63In the previous section, I considered the semantics-pragmatics divide in terms of truth-conditional vs. non truth-conditional content. In particular, I argued that truth-conditional meaning cannot be equated with conventional meaning. Indeed, some contextual phenomena are at play in the determining of truth-conditional content. But what kind of contextual phenomena? How is the context intruding into propositional content? As proposed by Recanati [2004], there are two ways of conceiving of context: a narrow one, in which the main role of context is to determine the semantic values of indexicals; and a larger one, in which context may provide any information relevant to determining the speakers’s intentions. On the first point of view, the only pragmatic process that may affect the proposition expressed is saturation, a pragmatic process that Recanati describes as being bottom-up because it is linguistically required. Saturation corresponds to the filling of a given contextual slot, which, if unfilled, would leave the utterance semantically incomplete; it mainly involves the determination of the reference of indexicals (I, now, here...) and pronouns (she, it, this...). But as argued by proponents of truth-conditional pragmatics like Recanati, truth-conditional content is not exhausted by decoding and bottom-up pragmatics processes. Other pragmatic processes are involved, that are not linguistically mandated but take place for purely pragmatic reasons. Recanati calls them top-down pragmatic processes.

64Among top-down pragmatic processes, Recanati [2004] distinguishes two main classes of phenomena: unarticulated constituents and pragmatic modulation. Recanati [2006] illustrates the case of unarticulated constituents by the following example: if someone asks me about lunch time whether I am hungry, and I answer (28a), the hearer might rightly infer that my answer is “no”. But to retrieve the conversational implicature in (28c), they need to identify the right truth-conditional content conveyed by (28a), namely (28b):

(28a) I’ve had a very large breakfast.
(28b) I’ve had a very large breakfast [today].
(28c) No, I’m not hungry.

(28a) thus involves an unarticulated constituent, namely the temporal location of the breakfast event on the day of utterance. This constituent, added for pragmatic reasons and by pragmatic means, does not correspond to any specific linguistic constituent in the utterance, but is part of its truth-conditions.

65The second type of top-down pragmatic process is what Recanati [2004] calls pragmatic modulation. Unlike the previous phenomenon which involves the addition of an extra component, modulation directly operates on the meaning of a linguistic expression. As Recanati [2004: 131] puts it,

[i]n context the meaning of words is adjusted or ‘modulated’ so as to fit what is being talked about. Sense modulation is essential to speech, because we use a (more or less) fixed stock of lexemes to talk about an indefinite variety of things, situations and experiences. Through the interaction between the context-independent meanings of our words and the particulars of the situation talked about, contextualised, modulated senses emerge, appropriate to the situation at hand.

66Recanati identifies different types of pragmatic modulation. Indeed, the meaning of the expression may be narrowed down, like in (29a) and (29c) where rabbit alternatively means rabbit fur and rabbit meat:

(29a) She wears rabbit.
(29b) She eats rabbit.

67The meaning of the expression may also undergo broadening, such as in (30) where swallow is given a metaphorical extended sense compared to its literal use:

(30) The ATM swallowed my credit card.

68Also, pragmatic modulation may consist in a process of predicate transfer (an instance of metonymy), such as in (31), where the property of “being parked out back” is literally assigned to the owner of the car rather than to the car itself:

(31) I am parked out back.

To retrieve the right truth-conditions of (31), the hearer will have to pragmatically reattribute the property expressed by the verbal predicate to the car owned by the speaker.

69Can top-down pragmatic processes also somehow affect axiological connotation? And if so, along which lines?

3.2. What kind of pragmatic modulation could axiological connotation undergo?

70Let us first consider the phenomenon a priori. If axiological connotation could be affected by context, this would mean that a given word or expression displays an encoded axiological connotation ac, but under the influence of context, ac would be changed into a new, distinct axiological connotation ac’. This process, if it occurs, has nothing to do with unarticulated constituents, for the point is not to “add” a new meaning component. Rather, it is to modulate or adjust the value attached to the referent of the word. Thus, the process could be subsumed under the general phenomenon of pragmatic modulation, in the sense of a top-down pragmatic process modifying the semantic information attached to the word. But a kind of pragmatic modulation that is neither narrowing, nor broadening, nor predicate transfer.

  • 17 True, axiological connotation somehow depends on the speaker’s and hearer’s subjectivity, as was ar (...)

71Along what lines precisely could pragmatic modulation of axiological connotation occur? Two distinct dimensions can be considered. First, the valence of the connotation: ac could be positive and changed into a negative connotation ac’; or conversely, ac could be negative and changed into a positive value ac’. Second, the force or height of the value: the degree of positivity or negativity assigned to the referent of a word could be reinforced or conversely lowered. Of course, the two parameters are somehow linked, since lowering a positive connotation ac could possibly turn it into a negative connotation ac’. And conversely, strengthening a negative connotation ac could change it into a new, positive connotation ac’ (I discuss more specific examples in Section 3.3.)17.

72Has such a kind of pragmatic modulation already been identified? Certainly not as far as axiological connotation is concerned. But it is reminiscent of another type of pragmatic modulation, pointed out among others by relevance theorists like Wilson & Carston [2007], namely hyperbole. Indeed, when one says about someone that she is a saint or a giant, the degree of kindness or the height assigned to the individual is extremely high, and it has to be pragmatically reinterpreted as somehow lower. Conversely, the figure of understatement, less frequently mentioned, consists in literally assigning to a given referent a property with a certain degree that has to be pragmatically modulated as being in fact higher. For instance, when one says about some achievement that it is “not bad”, one may in fact pragmatically communicate that it is “really good”. Thus, in both cases, a given degree communicated literally has to be reinterpreted into a higher or lower degree depending on context. However, hyperbole and understatement differ from pragmatic modulation of axiological connotation in a crucial point: they concern truth-conditional content – or at-issue content – and not presupposed content.

73Could pragmatic modulation actually target presupposed content such as axiological connotation?

3.3. What kind of axiological connotation could pragmatic modulation target?

74In Section 1., I showed that some words display an encoded axiological connotation, and I even considered the hypothesis that this is potentially true for any open-class word. But according to contextualists, pragmatic modulation is a pervasive phenomenon that potentially affects any word or linguistic expression. So one might in turn wonder why pragmatic modulation should be limited to at-issue content. If contextual adjustment may be required in every context to determine the propositional contribution of an utterance, why would it not also be needed for presupposed content? Why could the encoded axiological connotation not be sometimes pragmatically adjusted, in a way that is both triggered and performed by context? Just like encoded at-issue content is a starting point that calls for pragmatic adjustment in the process of communicating and retrieving the actual truth-conditional content, encoded axiological connotation may be supplemented by context in the communication and identification of the evaluative content actually attached to a word.

75But can any kind of encoded axiological connotation actually undergo pragmatic modulation? In Sections 1.2. and 1.3. respectively, I have distinguished linguistically encoded axiological connotation and culturally encoded axiological connotation. The latter appears to be the most obvious candidate for pragmatic modulation. Indeed, different authors have expressed such an intuition. For example, when mentioning “an index corresponding to the average affectivity rate that [each lexical unit] potentially involves” – a phenomenon that I have argued corresponds to what I call culturally encoded axiological content – Kerbrat-Orecchioni [1977: 109] suggests that one should “investigate in each speech act how the context operates to increase (eg. the word “cheminée” in Du Bellay’s poem “Heureux qui comme Ulysse”) or reduce (eg. the word “liberté” in a textbook of jurisprudence) this average rate”. Indeed, “Heureux qui comme Ulysse” is a poem in which the author Du Bellay, during a disappointing trip in Rome, expresses nostalgia for the French village where he grew up. In this specific context, the chimney of his childhood home becomes a highly valued element and one may assume that it receives a more positive axiological connotation than in default contexts. By contrast, the somehow neutral and dispassionate context created by a textbook of jurisprudence is likely to decrease the – positive – emotional commitment usually attached to the word liberté. Even a word like AIDS, which we have seen is perceived by speakers as strongly negative, may in some contexts sound more positive. For example, in (32), what is put forward in the denotation of the word is not the devastating impact of the disease on one’s health but rather the fact that it constitutes a tremendous research field giving rise to significant scientific progress:

(32) The objectives of the conference were: 1) to present the most recent research findings and trends in HIV/AIDS and technology […] (https://www.ncbi.nlm.nih.gov/​pmc/​articles/​PMC4269289/​, accessed July 27, 2023)

Thus, in that specific context, the word is likely to be perceived maybe not fully positively, but at least less negatively than usual.

76Regarding linguistically encoded axiological connotation, the question seems slightly more intricate. True, one can find instances of terms displaying a linguistically encoded negative axiological connotation that are used in a positive way. For example, in the utterance in (32), found on Reverso (https://context.reverso.net/​traduction/​anglais-francais/​rug+rat, accessed August 31, 2023), the expression rug rat receives its default negative axiology:

(32) A: Not time to fetch the rug rat?
B: It is […]. Don’t call him that, okay?

77By contrast, in (33), rug rat clearly features in a context that lowers its negative connotation and rather turns it into a hypocoristic term, expressing a positive affective value:

(33) I’ve always wanted to be a dad. I just can’t wait to have a little rug rat running around. I used to want five or six kids, but maybe I’ve become too self-absorbed over the years. I think two would be perfect. (https://quotefancy.com/​quote/​1168585/​Josh-Duhamel-I-ve-always-wanted-to-be-a-dad-I-just-can-t-wait-to-have-a-little-rug-rat, accessed July 27, 2023)

78The very phenomenon is described in the following online comment about the notion of dysphemism:

Dysphemisms involve the use of negative instead of positive expressions, although not all are intended to be rude. They can be used in an affectionate and light-hearted manner too, for example a parent calling her child ‘rugrat’” (https://www.educationworld.in/​exchanging-unpleasantries/​#:%7E:text=%27%20Dysphemisms%20involve%20the%20use%20of,her%20husband%20%27old%20man%27, accessed July 27, 2023).

  • 18 Traugott & Dasher [2009: 55] describe the process of melioration or amelioration as “the tendency t (...)
  • 19 The process we are mentioning for rug rat is purely pragmatic and occurs synchronically. By contras (...)

Interestingly, the following general comment made by Traugott & Dasher [2002: 56] about (a)melioration18 might shed some light on the phenomenon at play: “the earlier and later meanings19 share a sense of emotional concern by a stronger party toward a weaker party”. Likewise, Camp [2013: 332] emphasizes that “not all uses of slurs do express an occurrent, negatively charged attitude”. She illustrates the phenomenon with the following example featuring the word spic:

(34) I’m glad we have so many spics at our school: they always bring the best food to our fund-raising functions.

According to the Cambridge dictionary, spic “is an extremely offensive word for someone who comes from Latin America, or whose family comes from Latin America”. However, as argued by Camp, “[i]n (34), the speaker is not using the slur to give vent to a feeling of contempt, but in praise”.

79Does this mean that linguistically encoded axiological connotation can undergo pragmatic modulation like culturally encoded axiological connotation? This is far from clear. Regarding the “positive” uses of rug rat, the expression is precisely chosen because of its initial derogatory connotation. To put it otherwise, the hypocoristic value that rug rat acquires when uttered by caring and benevolent parents derives from the very fact that the expression initially sounds pejorative. This is a distinct phenomenon from what happens when culturally encoded axiological connotation is pragmatically modulated. In the latter case, as outlined in Section 3.2. and illustrated at the beginning of this section, the initial encoded axiological connotation ac is turned into a new, pragmatically adjusted axiological connotation ac’. In the case of affectionate uses of rug rat, ac is not changed into ac’: ac’ gets superimposed on ac. The initial pejorative or derogatory value of the expression persists: it is still perceived, even if the expression is used positively. Interestingly, this view is defended by Belleri [2020: 19] about affectionate uses of slurs more generally. According to her, when a slur is “accompanied by feelings of affection and admiration”, like in Camp’s example in (34), and the slur is used “in ordinary, literal and non-ironical communicative circumstances, amongst competent, sincere speakers, neither of whom belongs to the slur’s target group”, the expression remains offensive (see Belleri [2020]’s Section 3. for detailed arguments). This boils down to saying that the axiological connotation linguistically encoded in the word or expression necessarily persists, i.e. gets communicated, even in contexts where the affectionate or laudatory intentions of the speaker are obvious and the expression is loaded with this additional positive connotation. Can other instances of linguistically encoded axiological connotation undergo a process of pragmatic modulation? For example, is it possible to find contexts in which words like foolhardy or stingy are used in a positive way? I have found no such example, but I expect that if so, the words again will not get rid of their initial negative value; just as in the case of rug rat, the pragmatically acquired positive connotation ac’ will add to the linguistically encoded negative connotation ac rather than substitute for it or be a modulation of it. Further analyses, both in terms of corpus studies and in terms of experimental work, could help shed light on this phenomenon.

Conclusion

80In this paper, I investigated how axiological connotation articulates with the semantic-pragmatic distinction. Since the words semantic and pragmatic are extremely polysemous, I examined in turn their different uses, in a way that has progressively helped me better circumscribe the very phenomenon of axiological connotation.

81First, I considered the semantic-pragmatic distinction in terms of encoded vs. non encoded meaning. I argued that one should distinguish between two kinds of encoded axiological connotation, namely linguistically encoded axiological connotation and culturally encoded axiological connotation. Furthermore, I reviewed work in psychology suggesting that culturally encoded axiological connotation may not be restricted to a small subset of words, but may correspond to a more pervasive phenomenon, that potentially concerns any open class word, or at least many more words than those that are usually dubbed as expressives.

82I argued that this view seems supported by experimental work that has been led in the field of psychology. First, it has been shown in several experiments that participant’s judgements largely converge on the valence of lexical items. Second, it has been contended by some psychologists that the retrieving of affective information actually precedes the identification of referential content. But if this is also true for words, this suggests that axiological connotation does not correspond to incidental information but to central, significant information, which in turn makes all the more plausible the fact that any open class word conveys axiological connotation.

83In the second section, I considered the semantic-pragmatic distinction in terms of truth-conditional vs. non truth-conditional content. In light of recent work on thick terms, I showed that both linguistically and culturally encoded axiological connotation should be analyzed as presupposed content, a kind of content that is not strictly speaking truth-conditional, in the sense that it is not part of the proposition expressed. Also, I considered to which extent axiological connotation parallels the phenomenon of higher-level explicatures. Although the differences are real (in particular, higher-level explicatures operate at the level of the whole proposition, whereas axiological connotation operates at the level of the lexicon), I emphasized that the two kinds of content correspond to expressive meaning and may be analyzed as providing procedures for retrieving representations of emotional states.

84Finally, in the last section, I embraced a strong conception of pragmatics, according to which there are top-down pragmatic processes, that is, contextual processes involved in the determining of truth-conditional content that are not linguistically mandated but take place for purely communicative reasons. I contended that pragmatic modulation, namely pragmatic modification of a word’s linguistic meaning, may indeed apply to culturally encoded axiological connotation, a phenomenon that is all the more striking that pragmatic modulation then applies to presupposed content. But I argued that the situation is trickier for linguistically encoded axiological connotation: the initially encoded axiological connotation cannot be pragmatically modulated or adjusted – it may only be supplemented with another distinct axiological connotation that comes and superimposes itself to the former one. But the core linguistically encoded axiological connotation cannot be erased.

85This study of course calls for further investigation of the phenomenon of axiological connotation. In particular, experimental work might be extremely helpful to investigate the significance of axiological connotation in the processing of lexical meaning. Indeed, it might be worth reassessing a long-standing tradition that has taken for granted that referential meaning is central in semantics, and rather re-enhancing the cognitive role of axiological connotation. Is referential meaning or axiological connotation detected first? Is referential meaning or axiological meaning given more weight to determine to which extent two words are semantically close? New experimental investigations might help address these questions. Also, experimental work may help establish whether pragmatic modulation indeed can target culturally encoded axiological connotation but not linguistically encoded axiological connotation. Finally, how do several axiological connotations attached to distinct lexical items interact with each other within a given utterance? Along which parameters? This work on lexical items calls for further analyses regarding the way lexical axiological connotation fits in a broader compositional picture of language.

Top of page

Bibliography

Bargh John A., Litt J. E., Pratto Felicia & Spielman L. A., “On the preconscious evaluation of social stimuli”, in Bennett Adrienne F. & MacConkey Kevin M. (Eds.), Cognition in individual and social contexts, Proceedings of the XXIV International Congress of Psychology, Vol. 3, 357-370.

Barrett Lisa F. & Bar Moshe, 2009, “See It with Feeling: Affective Predictions in the Human Brain”, Royal Society Philosophical B, Vol. 364, n°1513, 1325-1334.

Barthes Roland, 1964, “Rhétorique de l’image”, Communication n°4, 41-42.

Belleri Delia, 2020, “Slurs: departures from genuine uses and derogation”, Studies in logic, grammar and rhetoric, Vol. 62, n°75, 9-24.

Blakemore Diane, 2011, “On the descriptive ineffability of expressive meaning”, Journal of Pragmatics 43, 3537-3550.

Blakemore Diane, 2013, “The expressive meaning of racial epithets: towards a non-unitary account of expressive meaning”, University of Brighton College of Arts and Humanities, Lectures in Language and Linguistics.

Bourmayan Anouch, 2022, “Les adjectifs axiologiquement opposés : le cas de prudent vs. timoré”, in Behe Louise, Carel Marion, Dall’ Cortivo Lebler Cristiane, Denuc Corentin & Gomes Lauro, Énonciation et Argumentation, Humanidades & Inovação, Vol. 9, n°4, 74-88.

Bradley Margaret M. & Lang Peter J., 1999, Affective norms for English words (ANEW): Instruction manual and affective ratings, Technical Report C-1, The Center for Research in Psychophysiology, University of Florida.

Camp Elisabeth, 2013, “Slurring perspectives”, Analytic Philosophy, Vol. 54, n°3, 330-349.

Carston Robyn, 2002, Thoughts and Utterances: The Pragmatics of Explicit Communication, Oxford: Blackwell.

Cepollaro Bianca & Stojanovic Isidora, 2016, “Hybrid Evaluatives: In Defense of a Presuppositional Account”, Grazer Philosophiesche Studien 93, 458-488.

Firestone Chaz & Scholl Brian J., 2016, “Cognition does not affect perception: Evaluating the evidence for “top-down” effects”, Behavioral and Brain Sciences 39, 1-77.

Frege Gottlob, 1892, “Über Sinn und Bedeutung”, Zeitschrift für Philosophie und philosophische Kritik NF 100.1, 25–50, translated by Max Black as “On sense and reference”, in Geach Peter & Black Max (Eds), Translations from the Philosophical Writings of Gottlob Frege, 1960 [1952], Oxford: Blackwell, 56-78.

Grice Paul H., 1975, “Logic and Conversation”, in Cole P. & Morgan J. L. (Eds.), Syntax and Semantics, Vol.3, Speech Acts. New York: Academic Press, 41-58. Reprinted in 1989, Studies in the Way of Words, Cambridge, Mass: Harvard University Press, 22-40.

Gutzmann Daniel, 2015, Use-conditional meaning. Studies in Semantics and Pragmatics (Oxford Studies in Semantics and Pragmatics), Oxford: Oxford University Press.

Hart Samuel L., 1971, “Axiology – Theory of Values”, Philosophy and Phenomenological Research Vol. 32, n°1, 29-41.

Jeshion Robyn, 2021, “Varieties of pejoratives”, in Khoo Justin & Sterken Rachel (Eds.), The Routledge Handbook of Social and Political Philosophy of Language, New York: Routledge, 211-231.

Kerbrat-Orecchioni Catherine, 1977, La connotation, Lyon: Presses Universitaires de Lyon.

Kyle Brent, 2013, “How Are Thick Terms Evaluative?”, Philosophers’ Imprint 13(1), 1–20.

Mohammad Saif M., 2018, “Obtaining reliable human ratings of valence, arousal, and dominance for 20,000 English words”, Proceedings of the 56th Annual Conference of the Association for Computational Linguistics, 174-184.

Moeschler Jacques, 2018, “Présupposition et implicature. Où passe la frontière ?”, in Biglari Amir & Bonhomme Marc (Eds.), La Présupposition entre théorisation et mise en discours, Paris: Classiques Garnier, 57-82.

Orsi Francesco, 2015, Value Theory, New York: Bloomsbury Academic.

Potts Christopher, 2007, “The expressive dimension”, Theoretical Linguistics 33(2), 165-197.

Recanati François, 2006, “Pragmatics and Semantics”, in Horn Larry & Ward Gregory, Handbook of Pragmatics, Oxford: Blackwell, 442-462.

Recanati François, 2005, “It is raining (somewhere)”, Linguistics and Philosophy 30(1), 123-146.

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

Recanati François, 1989, “The pragmatics of what is said”, Mind and Language 4, 295-329. Reprinted in Steven Davis (Ed.), 1991, Pragmatics: A Reader, Oxford: Oxford University Press, 97-120.

Richard Mark, 2008, When truth gives out, Oxford: Oxford University Press.

Rouillé Louis, 2019, Disagreeing about fiction, Phd thesis, Université Paris sciences et lettres.

Schroeder Mark, 2021, “Value Theory, The Stanford Encyclopedia of Philosophy, Edward N. Zalta (Ed.), URL: https://plato.stanford.edu/archives/fall2021/entries/value-theory/.

Searle John, 1975, “Indirect speech acts”, in Cole P. & Morgan J. L. (Eds.), Syntax and Semantics, Vol. 3, Speech Acts, New York: Academic Press, 59-82. Reprinted in David S. (Ed.), Pragmatics: A Reader, Oxford: Oxford University Press, 265-277.

Soria Ruiz Andrés, Cepollaro Bianca & Stojanovic Isidora, 2021 “The Semantics and Pragmatics of Value Judgments”, in Stalmaszczyk Piotr (Ed.), Cambridge Handbook of the Philosophy of Language, Cambridge: Cambridge University Press, 434-449.

Stojanovic Isidora, forthcoming, “Evaluativity”, in Lepore Ernie & Stojnic Una (Eds.), Oxford Handbook of Contemporary Philosophy of Language, Oxford: Oxford University Press.

Traugott Elizabeth & Dasher Richard B., 2009, Regularity in semantic change, Cambridge: Cambridge University Press.

Väyrynen Pekka, 2009, “Objectionable Thick Concepts in Denials”, Philosophical Perspectives 23: 439-469.

Väyrynen Pekka, 2013, The Lewd, the Rude, and the Nasty, Oxford: Oxford University Press.

Warriner Amy Beth, Kuperman Victor & Brysbaert Marc, 2013, “Norms of valence, arousal, and dominance for 13,915 English lemmas”, Behavior Research Methods 45(4), 1191-1207.

Wilson Deirdre & Carston Robyn, 2007, “A unitary approach to lexical pragmatics: relevance, inference and ad hoc concepts”, in Burton-Roberts Noel (Ed.), Pragmatics, Basingstoke Hampshire/New York: Palgrave-Macmillan; 230-260.

Wilson Deirdre & Sperber Dan, 2004, “Relevance Theory”, in Horn Laurence R. & Ward Gregory (Eds.), The Handbook of Pragmatics, Oxford: Blackwell, 607-632.

Wharton Tim, “That bloody so-and-so has retired: Expressives revisited”, Lingua 175-176, 20-35.

Zajonc Robert B., 1980, “Feeling and thinking: Preferences need no inferences”, American Psychologist 35(2), 151-175.

Zajonc Robert B., 2000, “Feeling and thinking: Closing the debate over the independence of affect”, in Forgas Joseph P. (Ed.), Feeling and thinking: The role of affect in social cognition, Cambridge: Cambridge University Press, 31-58.

Top of page

Notes

1 My translation (as for other quotations from Kerbrat-Orecchioni and other French authors).

2 For an overview of the field of axiology – or more generally value theory – in philosophy, see among others Hart [1971], Orsi [2015] and Schroeder [2021].

3 The term also displays another specialized meaning that is prominent in linguistics and with which the psychological sense should not be confused: it designates the number of participants that a verb or more generally a predicate selects.

4 As an anonymous reviewer rightly suggests, one may want to add that Lila does so intentionally, since someone who accidentally and unaware that she is doing so inflicts unnecessary suffering is not thereby cruel.

5 Here and elsewhere (in the other quotations), I am responsible for the use of bold characters.

6 As an anonymous reviewer interestingly suggests, another alternative explanation is also available: perhaps athletic is not polysemous – i.e. it has one stored linguistic entry – but when this entry is deployed to talk about people we get a positive connotation.

7 In Bourmayan [2022], I argue that in minimal pairs like brave vs. foolhardy, cautious vs. shy or sparing vs. stingy, the “positive” value of brave, cautious or sparing is not necessarily linguistically encoded. I put forward an alternative analysis where only the “negative” words display a linguistically encoded axiological value: the apparently positive connotation of their lexical counterpart by contrast arises pragmatically. But whether brave, cautious and sparing are linguistically neutral or positive does not matter here, for my point specifically targets “negative” words. And the reasoning still holds: since foolhardy, shy and stingy display a negative connotation but are referentially close to other terms that do not convey the same evaluative content (be they axiologically neutral or positive), this suggests that the negative feature is linguistically encoded and not only cultural. Also, as emphasized by an anonymous reviewer, the existence of minimal triplets like [slim/thin/skinny] including a positive word, a neutral word and a negative word suggests that the positive value of some words at least must be linguistically encoded: the positive word cannot be reinterpreted as pragmatically positive but linguistically neutral, since it already displays a linguistically neutral counterpart.

8 Opinions on whether an expression like rug rat is only colloquial or also derogatory might slightly diverge among speakers. For example, the Cambridge English Dictionary only highlights the informal dimension of the word (https://dictionary.cambridge.org/fr/dictionnaire/anglais/rugrat, accessed August 31, 2023). However, other sources do mention rug rat’s derogatory aspect, among which Wiktionary (https://en.wiktionary.org/wiki/rugrat, accessed August 31, 2023) and WordReference (https://www.wordreference.com/enfr/rug%20rat, accessed August 31, 2023).

9 Interestingly, Warriner et al. show that more polarized – i.e. strongly positive or negative – words (like for example abuse or jolly) elicit more consistent ratings than do more neutral ones (like vial or triplicate).

10 Whether or not metaphorical meaning is part of truth-conditional content is an ongoing debate to which I will briefly come back in Section 3. Likewise, the articulation between fictional content and truth-conditions is a controversial issue, but it will not be addressed here (see, for example, Rouillé [2019] on the topic).

11 For an extensive comparison between presuppositions and implicatures, see Moeschler [2018].

12 Of course this excludes metalinguistic uses of lewd, where what is targeted by the operator precisely is not the truth-conditional content but non at-issue content. For example, in (1) and (2), (14c) does fall into the scope of the operator – respectively the negation and the interrogation – whereas (14b) resists:
(1) Madonna’s show is not lewd, it is just sexually explicit.
(2) Is Madonna’s show lewd? No, it is just sexually explicit.

13 Although such data appear as knock-down arguments in favor of a presuppositional account, it should be mentioned that other analyses of the evaluative content of thick terms have been put forward. In particular, Kyle [2013] has argued for an account in terms of entailment (the evaluative content would correspond to a contribution to the truth-conditional content of the utterance) and Väyrynen [2013] for an account in terms of conversational implicatures. See Cepollaro & Stojanovic [2016] for answers to the queries raised by Kyle [2013] and Väyrynen [2013] about the presuppositional view.

14 As we have just seen in footnote 12., a speaker may distance himself from this value system by using lewd metalinguistically and calling into question in some way or other the negative judgement attached to it:
(1) Madonna’s show is not lewd, it is just sexually explicit.
(2) Is Madonna’s show lewd? No, it is just sexually explicit.e.g.
(3) I don’t like the term lewd because it conveys the debatable idea that any sexually explicit thing is bad.

15 (15d) is not an implicature. It would be if it were inferred on the basis of truth-conditional content and contextual information, that is, information that can be true or false independently of the utterance. But (15c) is not contextual information. It corresponds to presupposed information, namely information that has to be true as long as the sentence has been uttered. If (15d) were an implicature, it could be cancelled. For example, to go back to Grice’s example in (13), one can assert without difficulty: “Smith has been paying a lot of visits to New York lately, but he is single”, hence cancelling the implicature that Smith might have a girlfriend in New York. Yet, one cannot say coherently: “Madonna’s show is lewd, but it is not bad because of being sexually explicit”, so (15d) is not an implicature. But on the view that (15d) is a logical implication of the descriptive, truth-conditional content of (15a) expressed in (15b) and the general evaluative presupposed content of (15a) expressed in (15c), these data make full sense: by uttering (15a), the speaker asserts that (15b) is true and presupposes that (15c) is true, so any logical implication of these two propositions – including (15d) – has to be true as well. Hence the fact that one cannot utter (15a) and cancel the proposition in (15d). (See Moeschler [2018] on the fact that implicatures can be cancelled, unlike presuppositions).

16 Again, this excludes metalinguistic uses of rug rat, where what is targeted by the operator precisely is not the truth-conditional content but non at-issue content. For a similar – and more detailed – reasoning, see footnote 12.

17 True, axiological connotation somehow depends on the speaker’s and hearer’s subjectivity, as was argued by Frege [1892]. But we have seen in Section 1.3. that there is nonetheless a strong convergence among speakers regarding culturally encoded axiological connotation, not to speak about linguistically encoded axiological value. At a pragmatic level, the same is true. Some pragmatic variation in the personal context of the speaker (let us call it “subjectivity”) may not be graspable by the hearer, but we are focusing here on pragmatic changes that occur in a cognitively shared context and that are hence communicated.

18 Traugott & Dasher [2009: 55] describe the process of melioration or amelioration as “the tendency to semanticize more positive connotations”.

19 The process we are mentioning for rug rat is purely pragmatic and occurs synchronically. By contrast, in that quotation, Traugott & Dasher are describing a change that occurs diachronically and affects the encoded, conventional meaning of the term, hence the expression of “earlier and later meanings” instead of “semantic (or conventional) and pragmatic meanings”. But Traugott & Dasher’s remark is still fully relevant for the phenomenon at play, for diachronic semantic changes in conventional meaning actually result from changes that have first occurred pragmatically and synchronically. Thus, what is relevant for the ultimate conventional phenomenon is relevant as well for the original pragmatic process.

Top of page

References

Electronic reference

Anouch Bourmayan, Axiological connotation and the semantic-pragmatic distinctionLexis [Online], 21 | 2023, Online since 13 November 2023, connection on 02 December 2023. URL: http://journals.openedition.org/lexis/6979; DOI: https://doi.org/10.4000/lexis.6979

Top of page

About the author

Anouch Bourmayan

Sorbonne Université, France
anouch.bourmayan@sorbonne-universite.fr

Top of page

Copyright

CC-BY-SA-4.0

The text only may be used under licence CC BY-SA 4.0. All other elements (illustrations, imported files) are “All rights reserved”, unless otherwise stated.

Top of page
Search OpenEdition Search

You will be redirected to OpenEdition Search