Expressive APs and expletive NPs revisited: Refining the extant relevance-theoretic procedural account

EXPRESSIVE APs AND EXPLETIVE NPs REVISITED: 1 REFINING THE EXTANT RELEVANCE-THEORETIC PROCEDURAL ACCOUNT 2 3 Abstract 4 In the third stage in the development of the notion of procedural meaning, relevance 5 theorists analysed expressive devices like intonation, interjections, expressive APs, 6 expletive NPs, slurs and paralanguage. This paper revisits expressive APs and 7 expletive NPs by arguing that the emotional-state descriptions they trigger only take 8 within their scope a propositional constituent. Such shorter-ranging mental structure 9 does not fit neatly in the definition of higher-level explicature, so two possible solutions 10 are proposed in order to accommodate it within the relevance-theoretic notional 11 apparatus, the second of which could involve a refinement of the distinction between 12 lower- and higher-level explicatures. Moreover, that type of constituent-directed 13 description is shown to be also enacted by other expressives and other linguistic 14 expressions, and to be constructed in other communicative acts. Finally, since the 15 words functioning as expressive APs and expletive NPs come from various lexical 16 categories, an account of the lexical-pragmatic processes enabling their transfer is 17 given. 18 19 Keywords: expressive APs, expletive NPs, procedural semantics, explicatures, 20 proceduralisation, relevance theory 21 22 1. Introduction 23 Consider the following utterances in English and Spanish: 24 (1) The fucking lady was so lucky that she did not get stuck between the car and 25 the wall! 26 (2) Peter has got the damned flashlight. 27 (3) ¡Déjame de una vez el jodido libro! 28 ‘Lend me the fucking/blasted book at once!’ 29 (4) Llevo toda la tarde con la dichosa ponencia. 30 ‘I have been working on the bleeding presentation the whole afternoon’. 31 (5) Juan es un puñetero imbécil; se ha olvidado las entradas en su casa. 32 ‘John is a fucking idiot; he forgot the tickets at home’. 33 (6) No he visto al puto niño. 34 ‘I haven’t seen the bloody bastard’. 35 In them, the head of a noun phrase seems to be modified by the italicised present and 36 past participles ‘fucking’, ‘damned’ and ‘jodido’ (1-3), the adjectives ‘dichosa’ and 37 ‘puñetero’ (4-5), and the noun ‘puto’ (6). Despite their adnominal position, those words 38 do not function as adjectives or generate adjectival phrases. 39 Like other content words, adjectives encode concepts consisting of a logical entry 40 with information about their set of stable logical properties, a lexical entry with 41 information about the natural-language words used to express them and their 42 grammatical characteristics, and an encyclopaedic entry with their extension and/or 43 denotation –states, qualities or properties connected with notions such as shape, taste, 44 size, colour, judgement, etc. (Börjars and Burridge 2001: 62; Huddleston 1988: 108)– 45 as well as varied, idiosyncratic information connected with them (Carston 2016: 155; 46 Sperber and Wilson 1986/1995: 92-93) 1 . Clearly, none of the words above refer to 47 stable qualities or scalar properties. Neither do they have heightening or lowering 48 1 These entries are what most atomic, unstructured concepts have, but others may lack one of them. effects, nor do they fulfil a restrictive function. Impossibility to intensify or gradate the 49 participles ‘fucking’ or ‘damned’ in (1-2) and to add prefixes to them further reveal that 50 they are not adjectives 2 . Furthermore, both in English and in Spanish, moving those 51 words to predicative position or replacing them with defining relative clauses would 52 involve illicit transformations that alter what speakers intend to communicate when 53 those words appear in pre-nominal position (Börjars and Burridge 2001: 64-65; Collins 54 and Hollo 2000: 80-81; Greenbaum and Quirk 1993: 142; Haegeman and Guéron 55 1999: 56-57, 71-72; Huddleston 1988: 108-110): 56 (7) The lady who/that was fucking was so lucky that… 57 (8) Peter has got the flashlight that is/was damned. 58 (9) ¡Dame de una vez el libro que está jodido! 59 ‘Give me the book that is broken/busted at once!’ 60 The words in question, which are transferred from various lexical categories, are no 61 doubt the vehicle by means of which speakers voice something elusive and ineffable: 62 attitudes, feelings and/or emotions (Blakemore 2011; Potts 2007a, 2007b). They are 63 expressives, a label subsuming a wide array of (non-)linguistic elements like expletive 64 and non-expletive nominal epithets (‘the bastard’, ‘el puto’), slurs (‘hori’, ‘chink’), 65 primary and secondary interjections (‘wow’, ‘oh’, ‘shit’, ‘damn’), prosody or gestures, to 66 name but a few (Blakemore 2011: 3537, 2015). This paper purports to develop an 67 account of expressive APs and expletive NPs which answers the following questions: 68 (i) How are expressive APs and expletive NPs used to communicate? 69 (ii) What do they contribute to communication? 70 (iii) What made it possible for elements from diverse lexical categories to achieve 71 expressive functions? 72 2 In Spanish, in contrast, some speakers could intensify or gradate the alleged participle ‘jodido/a’ in (3) and the alleged noun ‘puto’ in (6). Despite their syntactic position, expressive APs and expletive NPs do not provide 73 descriptive, truth-conditional information about the properties of the referents of the 74 nouns they accompany, as Section 2 shows. Expressive APs and expletive NPs 75 resemble interjections, intonation and gestures in that they provide direct evidence for 76 speaker meaning, which is, however, difficult to paraphrase. Nevertheless, their 77 specialisation to express certain emotions in specific, maybe repeated circumstances 78 or by certain language users, might activate specific conceptual representations, even 79 if vague or in need of subsequent fine-tuning. This would enable expressive APs and 80 expletive NPs to communicate by also providing some indirect evidence for speaker 81 meaning. Section 3 discusses the manner whereby they communicate on the basis of 82 the showing-meaning NN distinction (Blakemore 2011, 2015; Grice 1957; Wharton 83 2001, 2002, 2003, 2009). 84 The elusiveness of what expressive APs and expletive NPs communicate suggests 85 that their contribution to communication calls for an alternative explanation. This paper 86 endorses a procedural perspective, and treats expressive APs and expletive NPs as 87 encoders of mental instructions or procedures (Blakemore 1987, 2002; Wilson and 88 Sperber 1993, 2002, 2004). Like interjections, intonation or attitudinal adverbials 89 (Blakemore 2011, 2015; Ifantidou 1992, 1993a; Wharton 2003, 2009, 2016; Wilson 90 and Wharton 2006), expressive APs and expletive NPs are claimed in Section 4 to 91 encode instructions targeting emotion-reading mechanisms (Sperber and Wilson 92 1986/1995; Wilson and Sperber 2002, 2004). However, the procedural analysis this 93 paper proposes differs from the initial one of interjections, intonation or attitudinal 94 adverbials in some respects. The attitudinal or emotional descriptions that expressive 95 APs and expletive NPs are argued to enact will be shown not to be about something 96 manifest in the speaker’s cognitive environment or about a proposition that she 3 97 expresses, but to be confined to a constituent of a proposition: namely, the referent of 98 a noun. In this respect, expressive APs and expletive NPs resemble slurs, which voice 99 attitudes towards the referents that they denote (Blakemore 2015). 100 This procedural analysis, on the one hand, lends support to a non-unitary account 101 of the broad group of expressives (Blakemore 2015). On the other hand, it might imply 102 that the relevance-theoretic notional apparatus could be refined with a view to 103 accommodating the shorter-ranging attitudinal descriptions illustrated here, as these 104 differ significantly from higher-level explicatures. These constituent-focused 105 descriptions will furthermore be shown to be the output of other linguistic expressions 106 that do not necessarily fall within the category of expressive meaning, and to be 107 constructed in the processing of various communicative acts. 108 To conclude, Section 5 discusses why items initially belonging to distinct lexical 109 categories might develop an expressive function. Although the development of such a 110 function is explained as a result of grammaticalisation and subjectification, a proposal 111 is made about the processes that might propel what this paper calls the 112 proceduralisation of those lexical items. This proposal rests on current relevance- 113 theoretic postulates on lexical pragmatics, as well as on some semantic processes 114 frequent in children’s speech. 115 116 2. Expressive APs and expletive NPs, and the proposition expressed 117 Content words encode concepts that become constituents of conceptual 118 representations (Wilson and Sperber 1993: 16). Concepts may be “[…] brought to 119 consciousness, reflected on and used in general inference” (Wilson 2016: 11). 120 3 Following a relevance-theoretic convention, reference to the speaker will be made through the feminine pronoun while reference to the hearer will be made through the masculine counterpart. Although many concepts are verbalised by means of single words, concept-word 121 correspondences are not always exhaustive: single concepts map onto different words 122 or phrases, various concepts map onto the same word or some concepts are not 123 expressible at all through words (Sperber and Wilson 1997; Wilson and Sperber 1993). 124 Decoded concepts become part of the logical form of an utterance (Carston 2016; 125 Sperber and Wilson 1986/1995), which must undergo various inferential processes to 126 become truth-evaluable: reference assignment, disambiguation, recovery of 127 unarticulated constituents and conceptual adjustment. Interaction with other concepts 128 and contextual information adjust a lexically encoded concept, so that its denotation 129 becomes more specific (narrowed) or more general (broadened): “[…] either a proper 130 subset or a superset of [its original] denotation […]”, although it may also be “[…] a 131 combination, both extending the lexical denotation and excluding some part of it” 132 (Carston 2016: 156). If the resulting concept is an occasion-specific one, it is an ad 133 hoc concept (Carston 1997, 2002, 2013, 2016; Sperber and Wilson 2008; 134 Wałaszewska 2015; Wilson 2004; Wilson and Carston 2006, 2007). The meaning of 135 most words, therefore, is sometimes thought of as semantically underspecified and in 136 need of modulating before it can provide some specific content; to put it differently, 137 most words behave as if they encoded pro-concepts that need working out (Sperber 138 and Wilson 1997). 139 The inferential operations mentioned above yield the lower-level explicature, which 140 is a truth-evaluable proposition corresponding to ‘what is said’. Adjectival modifiers 141 clearly contribute to lower-level explicatures and such contribution is relatively 142 straightforward to check: elimination or replacement with another adjective or 143 participial adjective, as well as intensification or gradation, alter the meaning of the 144 proposition expressed and, therefore, its truth conditions (10). Additionally, movement 145 from the attributive to the predicative position does not alter the meaning of the 146 resulting sentence (11): 147 (10) a. John cut down the [(very/extremely) tall/green] tree. 148 b. John attended a(n)[(very/most) interesting/boring] lecture. 149 (11) a. The green/tall tree = The tree is green/tall. 150 b. An interesting lecture = The lecture is interesting. 151 c. A worried man = The man is worried. 152 Participial adjectives may also be substituted by a defining relative clause without any 153 effect on the proposition expressed: 154 (12) a. The lost property > The property that was/has been lost. 155 b. The escaped prisoner > The prisoner who (has/had) escaped. 156 Expressive APs and expletive NPs do not encode concepts denoting conditions or 157 (scalar) properties describing nouns. In their case, adnominal position does not involve 158 an adjectival meaning or function. What they encode in utterances like (1-6), if 159 anything, does not seem to be part of the proposition expressed and, hence, does not 160 have any impact on its truth conditions. If those words were eliminated, the resulting 161 proposition would remain the same, although some indication of the speaker’s feeling 162 about or attitude to one of its constituents would obviously disappear: 163 (13) The lady was so lucky that she did not get stuck between the car and wall! 164 To some extent, expressive APs and expletive NPs behave like attitudinal adverbs 165 such as ‘unfortunately’ and ‘sadly’, which indicate the speaker’s attitude to an assertion 166 (Ifantidou 1992, 1993a): 167 (14) a. Unfortunately, John has left the party. 168 b. Sadly, the beautiful princess has died. 169 Semantically external to the proposition expressed, the adverbials they give rise to 170 modify the assertion without contributing to its truth conditions (Strawson 1973, quoted 171 in Ifantidou 1992, 1993a; Urmson 1952). Embedding an utterance containing an 172 attitudinal adverbial into a conditional structure shows that the adverbial does not affect 173 its truth conditions because it is outside the scope of the conditional operator (Ifantidou 174 1992, 1993a). Accordingly, if (14a) was slightly adapted as (15a) in order to prevent it 175 from sounding odd when embedded into a conditional structure (15b), the speaker 176 would not be saying that they will all leave in the event that (15c) was true, but in the 177 event that (15d) was true: 178 (15) a. John has unfortunately left the party. 179 b. If John has unfortunately left the party, we will all leave too. 180 c. It is unfortunate that John has left the party. 181 d. John has left the party. 182 Attitudinal adverbs, however, convey information that may be true or false in itself, 183 which reveals their conceptual status. A speaker who employs a specific attitudinal 184 adverb may be contradicted by someone who knows or discovers that what she says 185 is false (16). Their conceptual status is additionally attested by the existence of 186 synonymous manner-adverbial counterparts that determine the truth conditions of the 187 utterances where they occur (17), the possibility to be intensified (18) or integration in 188 more complex syntactic structures (19) (Ifantidou 1992: 207-208): 189 (16) It is not unfortunate! We are all happy that he left! He is so boring! 190 (17) It is unfortunately true that John left the party. 191 (18) Quite/Very unfortunately, John left the party. 192 (19) Quite unfortunately and very regrettably, John left the party. 193 Lack of intensification, impossibility to be gradated, meaning shift when placed in 194 predicative position or replaced by a corresponding defining relative clause all support 195 the conclusion that the nouns, adjectives and participles functioning as expressive APs 196 or expletive NPs have lost their original descriptive content. As with attitudinal adverbs, 197 embedding an utterance containing an expressive AP or expletive NP into a conditional 198 structure reveals that they do not fall within the scope of the conditional operator, so 199 they do not affect the truth conditions of the utterance either. If (20a) was embedded 200 into a conditional structure like (20b), the speaker would be understood to be 201 recommending to take the lady to hospital provided (20c) was true, but not in the case 202 that (20d) was true: 203 (20) a. That fucking lady has been hit by the lamp. 204 b. If that fucking lady has been hit by the lamp, someone should take her to 205 hospital. 206 c. A particular lady has been hit by the lamp. 207 d. A particular lady who was fucking has been hit by the lamp. 208 As opposed to attitudinal adverbs, the feeling or attitude that expressive APs and 209 expletive NPs express cannot be denied, so a reaction like (21b) to a (21a) would 210 sound odd, as it would not make much sense to challenge, reject or question what the 211 speaker feels (Potts 2005: 158): 212 (21) a. That fucking lady was hit by a lamp! 213 b. She is not a fucking lady. 214 Expressive APs and expletive NPs certainly display attitudes, feelings or emotions just 215 as shrieks or interjections do. They do not amount to “[…] a proposition that can be 216 cancelled, but [are] simply […] means of displaying an emotion tied to the utterance in 217 which [they are] used” (Blakemore 2015: 26). If any at all, a possible reaction to an 218 The notion of procedural meaning has undeniably been a significant contribution to 359 pragmatics (Author 2016; Sasamoto and Wilson 2016; Wilson 2016). Some linguistic 360 elements encode instructions steering the modules and mechanisms mediating 361 comprehension. Procedures are part of the causal structure of the cognitive system 362 (Carston 2016; Escandell Vidal et al. 2011; Wilson and Sperber 1993) and must be 363 analysed as dispositions (Bezuidenhout 2004): they “[…] are systematically linked to 364 states” (Wilson 2011: 10) wherein language users perform specific operations at a sub- 365 personal level. Since each operation “[…] points the hearer in a [particular] direction” 366 (Wharton 2009: 61), “[…] procedural expressions act as ‘pointers’” (Wilson 2016: 11). 367 Procedural meaning is non-translational because of its elusiveness and inaccessibility 368 to consciousness (Carston 2016: 159; Wharton 2009: 61; Wilson 2016: 11; Wilson and 369 Sperber 1993: 16). Furthermore, procedural expressions (i) do not tend to compose 370 with each other –i.e. they are non-compositional– (ii) behave rigidly –i.e. their 371 instructions usually prevail– (iii) do not have nonliteral uses, and (iv) are not 372 polysemous (Carston 2016: 159-161). 373 Procedural expressions make up a rather heterogeneous group: some contribute to 374 the proposition expressed; others constrain higher-level explicatures; others set 375 relations among propositions or guide access to assumptions; still, others assist 376 epistemic vigilance mechanisms in their decisions on reliability of information and 377 informers (Carston 2016: 158-159; Wharton 2009: 65; Wilson 2012) 10 . Expressive APs 378 and expletive NPs add to this heterogeneity. Although they enact the construction of 379 descriptions of the speaker’s attitudes, feelings or emotions in a similar way to 380 intonation, interjections or attitudinal adverbs and adverbials, the descriptions they give 381 rise to are slightly different. 382 10 Carston (2016: 158-159) groups the various relevance-theoretic contributions on procedural meaning into four (evolutionary) stages in the development of the notion. 383 4.1. The procedural semantics of expressive APs and expletive NPs 384 The procedures encoded by gestures, facial expressions, interjections and intonation 385 activate conceptual schemas amounting to descriptions of the speaker’s attitude to 386 something manifest in her cognitive environment or to a whole proposition (Wharton 387 2003, 2009, 2016; Wilson and Wharton 2006). In the latter case, they constrain higher- 388 level explicatures by guiding the formation of a description of what the speaker may 389 feel towards what she says. Attitudinal adverbs also facilitate the construction of 390 descriptions like the following (Ifantidou 1992, 1993a): 391 (23) The speaker is happy/sad/furious that p. 392 The fact that the speaker’s attitude is projected towards the whole proposition is 393 evident from their various positions within the utterance –initial (24a), medial (24b) or 394 final (24c)– and their separation from the other syntactic constituents by means of 395 pauses, thus forming an independent tone unit or group: 396 (24) a. Happily, John came to the party. 397 b. John, happily, came to the party. 398 c. John came to the party, happily. 399 Occurrence in different positions and loose insertion in the syntax of the sentence also 400 suggest that the attitudinal description triggered by interjections also affects the whole 401 proposition: 402 (25) a. Wow, John came to the party! 403 b. John came to the party, wow! 404 The attitudinal descriptions facilitated by expressive APs and expletive NPs, in 405 contrast, could not be said to affect the whole proposition expressed. Potts (2005, 406 2007a, 2007b) considered that an expressive like that in (26a) could give rise to a ‘non- 407 at-issue’ proposition like (26b), where it is evident that the speaker has a particular 408 attitude towards a certain event: 409 (26) a. I’ve lost my damn keys. 410 b. The speaker holds a derogatory attitude towards the fact that he has lost his 411 keys. 412 Expressive APs and expletive NPs cannot be detached from the head noun with which 413 they occur and moved to sentence-initial or final positions: 414 (27) a. *Fucking, the lady was so lucky that… 415 b. *The lady was so lucky that she did not get stuck between the car and the 416 wall, fucking! 417 Expressive APs and expletive NPs are part of a phrasal constituent: a noun phrase. 418 Therefore, they do not convey information about the speaker’s attitude to the whole 419 propositional content of an assertion or towards some fact or event, as Potts (2005, 420 2007a, 2007b) claimed, but only to a fragment of an assertion: the referent of the noun 421 that they accompany. Had the speaker intended to express a particular attitude 422 towards the fact that the lady in (27) was so lucky that something extremely negative 423 did not happen to her, the speaker would have had to use an interjection and/or the 424 appropriate intonation and facial expressions: 425 (28) (Oh) Fuck!/Shit!/Wow! That lady was so lucky that she did not get stuck 426 between the car and the wall! 427 Non-detachability of expressive APs and expletive NPs from a nominal head 428 suggests that the attitudinal description they enact must be about it. Addition of 429 attitudinal adverbials or interjections to an utterance containing an expressive AP or 430 an expletive NP further indicates that the respective procedural semantics of these 431 expressions differ: while that of the former facilitates the construction of higher-level 432 explicatures, that of the latter do not: 433 (29) a. Happily, that fucking lady did not get stuck between the car and the wall. 434 b. Wow, that fucking lady did not get stuck between the car and the wall! 435 The attitudinal adverbial and the interjection in these examples trigger a description of 436 the speaker’s happiness because the said lady was not stuck between the car and the 437 wall. However, the expressive AP would give rise to a description of an attitude of 438 disdain, scorn, criticism, etc., which the speaker projects towards the lady in question. 439 In this sense, expressive APs and expletive NPs behave like slurs such as ‘hori’ or 440 ‘chink’, which target a disparaging attitude at a social group that is identified by their 441 conceptual, descriptive content (Blakemore 2015) 442 In the examples above the attitudes expressed in relation to the proposition – 443 positive– and to the said lady –negative– are opposed, but the same sort of attitude 444 may simultaneously be expressed regarding both a proposition and one of its 445 constituents: 446 (30) Shit!/Oh, fuck! That fucking lady did not get stuck between the car and the wall! 447 Despite differences in the outcomes of the instructions encoded by interjections and 448 expressive APs and expletive NPs, their respective procedural semantics may interact 449 and result in more precisely built higher-level explicatures and attitudinal descriptions 450 about propositional constituents. In other words, occurrence of interjections and 451 expressive APs or expletive NPs in the same utterance could provide more evidence 452 for the construction of similar attitudinal descriptions about the whole proposition 453 expressed and a constituent therein, respectively. Thus, those tokens of showing 454 would mutually reinforce each other and lend support for the construction of similar 455 emotional-state descriptions, even if with distinct scopes. 456 A similar interaction may arise with the procedural meaning of intonation, mood or 457 word order, thus triggering attitudinal descriptions that might include information about 458 the speaker’s emotions or feelings, and/or the manner in which she performs an action. 459 If imperative mood and word order (31a) give rise to a speech-act description like that 460 in (31b), occurrence of an expressive AP (31c) could make the hearer develop the 461 higher-level explicature in (31b) as in (31d): 462 (31) a. Give me that pen! 463 b. The speaker tells the hearer to give her that pen x . 464 c. Give me that fucking pen! 465 d. The speaker (very) angrily tells the hearer to give her that pen x . 466 Likewise, if interrogative word order and intonation (32a) enact the higher-level 467 explicature in (32b), addition of an expressive AP (32c) would yield a more complex 468 higher-level explicature capturing nuances of the speaker’s attitude (32d): 469 (32) a. Do you have that pen? 470 b. The speaker asks/wonders whether the hearer has pen x . 471 c. Do you have that fucking pen? 472 d. The speaker angrily/anxiously asks/wonders whether the hearer has pen x . 473 Thus, expressive APs may emphasise imperative or interrogative force and add further 474 overtones of irritation, anger, indignation, curiosity, anxiety or wrath, to name but a few. 475 An anonymous reviewer of this paper and Raeber and Wharton (2017) point out that 476 the participles, adjectives and nouns that can be used as expressive APs or expletive 477 NPs may sometimes have the intensifying function characteristic of adjectives like 478 ‘sheer’, as in the following examples: 479 (33) This gives me sheer enjoyment. 480 (34) The cake is fucking delicious/damn good! 481 (35) a. John is a fucking idiot. 482 b. Curro es un puto crack. 483 ‘Curro is a fucking star’. 484 The expressive usage of those participles, adjectives and nouns must certainly be 485 differentiated from their usage as intensifiers. When intensifying, they would definitely 486 not trigger descriptions of the speaker’s attitude towards the noun they co-occur with, 487 so that procedure would be blocked or impeded. Rather, those participles, adjectives 488 and nouns could give rise to the construction of ad hoc concepts. The head nouns and 489 adjectives they intensify are scalar: there are degrees of enjoyment, 490 deliciousness/goodness, idiocy or stardom. In their intensifying usage, those 491 participles, adjectives and nouns would enact the construction of the occasion-specific 492 concepts ENJOYMENT *, GOOD */ DELICIOUS *, IDIOT * and CRACK */ STAR *, which would 493 idiosyncratically move something up alleged scales of enjoyment, 494 deliciousness/goodness, idiocy or stardom. 495 Arguing that those participles, adjectives and nouns could also trigger a procedure 496 that differs from the one they activate when used as proper expressives raises the 497 following question: if procedural meaning is monosemic (Carston 2016: 159-161), 498 which procedure would expressive APs and expletive NPs really encode? The answer 499 to this question may be given on the basis of the notion of meta-procedure, or an 500 instruction to activate a particular procedure from among a set of candidate ones 501 (Wharton 2009; Wilson 2011, 2012, 2016). Accordingly, the procedural semantics of 502 the participles, adjectives and nouns that may be used both as expressives and as 503 intensifiers would not be restricted to any of the two procedures mentioned above, but 504 would amount to a meta-procedure that forces, so to say, the selection of the 505 appropriate procedure from a set comprising construction of an attitudinal description 506 about a constituent or construction of an ad hoc concept. Activation of the latter 507 procedure may be favoured by the scalar nature of the noun or adjective they co-occur 508 with and meta-linguistic or encyclopaedic knowledge concerning the interpretation of 509 this lexical combination. Just as slurs give access to assumptions about the range of 510 attitudes that their users may hold towards the group of individuals that they denote 511 (Blakemore 2015: 31-34), individuals may access assumptions about the outputs of 512 the participles, adjectives and nouns under analysis when occurring with diverse 513 phrasal heads, which may bias the activation of one or the other procedure. 514 Accordingly, the participles, adjectives and nouns preceding nominal or adjectival 515 heads and lacking a clear descriptive content could put hearers in a state wherein they 516 have to activate a particular procedure out of a set of two candidate ones. The head 517 and encyclopaedic information about the specific element with which it combines may 518 determine which procedure must be enacted. 519 520 4.2. Possible implications 521 The more restricted scope of the attitudinal descriptions triggered by expressive APs, 522 expletive NPs and slurs might prevent those descriptions from being neatly subsumed 523 by the relevance-theoretic notion of higher-level explicature, provided this is to be 524 strictly applied to conceptual schemas about attitudes or emotions that are solely 525 projected towards whole propositions. As currently formulated, this notion does not 526 seem to capture the fact that hearers also forge descriptions reflecting speakers’ 527 attitude(s) towards specific conceptual constituents of the propositions that they 528 express, i.e. towards only a portion of what speakers say. Therefore, the distinction 529 between lower- and higher-level explicatures could need refining in order to 530 accommodate these attitudinal descriptions. Two solutions seem plausible, then. 531 The first, and simpler one, is modifying the definition of higher-level explicature so 532 that it is not solely restricted to attitudinal descriptions about whole propositions, but 533 also includes the shorter-ranging attitudinal descriptions that expressive APs, expletive 534 NPs and slurs trigger. This more parsimonious solution does not contravene Occam’s 535 razor principle and would involve loosening the term ‘higher-level explicature’ so as to 536 encompass any attitudinal description constructed regardless of whether it is about a 537 whole proposition or a constituent or portion of a proposition. 538 The second solution amounts to coining a new label as a way of differentiating an 539 additional type of attitudinal description that only takes within its scope a propositional 540 constituent: intermediate-level explicature. In relevance-theoretic terms, a lower-level 541 explicature is the truth-evaluable propositional form resulting from the pragmatic 542 enrichment of a decoded logical form, while a higher-level explicature is the conceptual 543 schema referring to the speaker’s emotions, feelings or attitudes, which is inferentially 544 constructed upon perception of some manifest fact, object, event or propositional 545 content and subsequent activation of some attitude- or emotion-related concept 546 (Sperber and Wilson 1986/1995; Wilson and Sperber 2002, 2004). Metaphorically, a 547 lower-level explicature could be portrayed as a Meccano block containing smaller 548 pieces –concepts and pragmatically enriched variables. That whole block may be 549 inserted inside another larger block corresponding to the higher-level explicature –the 550 block would be like some sort of subordinate structure, while the larger block within 551 which it is inserted would be like some sort of superordinate structure. However, the 552 pieces corresponding to concepts may in turn be inserted into bigger pieces or blocks 553 that would still fit in the lower-level explicature block. One such bigger piece or block 554 would be an intermediate-level explicature. Graphically, an intermediate-level 555 explicature could be represented as follows: 556 (Insert Figure 1 here) 557 The term ‘intermediate-level explicature’ would capture the idea that some attitudinal 558 descriptions need not subsume a whole proposition, but a smaller fragment or a portion 559 of it. To put it differently, this term would capture the intuitions that speakers express 560 emotional states –i.e. perform expressive acts– about propositions and their 561 constituents and that hearers may forge some mental structures to somehow portray 562 those acts. Were this second solution actually plausible, an intermediate-level 563 explicature could be defined as an emotional-state description constructed as a 564 consequence of a procedure activated by elements like expressive APs, expletive NPs 565 and slurs. That description would consist of attitude- or emotion-related concepts, and 566 would only take within its scope any phrasal constituent of a proposition, so it need not 567 be restricted to a particular syntactic constituent like the subject or the direct object, 568 but could affect the head of any propositional constituent. Accordingly, the 569 ‘explicaturial’ structure of an utterance like (36) could be glossed as in figure 2: 570 (36) Wow, that fucking lady was hit by the lamp! 571 (Insert Figure 2 here) 572 Higher-level explicatures are mental conceptual schemas that somehow reflect the 573 attitude expressed towards a proposition, or, in other words, an expressive act whose 574 sincerity is normally presupposed –it would make little sense to say that someone who 575 shrieks in pain is not feeling pain, for instance. Moreover, higher-level explicatures do 576 not affect the truth-conditions of the proposition that they take within their scope (see 577 the tests in Section 2). Inasmuch as an intermediate-level explicature also captures 578 another expressive act, even if this only affects a constituent of a proposition, it is also 579 non-truth-evaluable and has no impact on the truth-conditional content of that 580 proposition. Although such an act is connected with the proposition as long as it 581 consists of the expression of some emotional state caused by or felt about one of its 582 constituents, intermediate-level explicatures are independent of lower-level 583 explicatures and do not contribute to them. While lower-level explicatures are 584 constructed after a decoded logical form is pragmatically enriched through 585 disambiguation, conceptual adjustment, reference assignment and recovery of 586 unarticulated constituents, intermediate-level explicatures are constructed once 587 emotions are, so to say, ‘read’ on the basis of perceptible cues –i.e. paralanguage, 588 interjections, intonation, expressives, etc.– and a conceptual schema portraying them 589 is generated. 590 The appropriateness of the term ‘explicature’ to label this additional attitudinal 591 description has been questioned by another anonymous reviewer. ‘Explicature’ is used 592 here following the definition and characterisation of the two types of mental structures 593 –i.e. the pragmatically enriched logical form, or lower-level explicature, and the 594 conceptual schema portraying feelings, emotions or attitudes, or higher-level 595 explicature– making up the explicit content of utterances given by Sperber and Wilson 596 (1986/1995). Both structures are built on the basis of elements encoding conceptual 597 meaning, procedural meaning or a combination of both; in other words, explicatures 598 result from decoding and inference upon perception of certain intentionally produced 599 (para)linguistic elements, and this is precisely why they are part of the explicit content 600 of utterances. Explicatures are not the result of accessing and relating implicated 601 premises in order to derive implicated conclusions. 602 Lower- and higher-level explicatures may be strong or weak depending on the 603 reliance on decoding or the amount of inference needed for their construction. The 604 more reliance on encoded material, the stronger the explicatures; the more inferential 605 work needed, the weaker they are (Wilson and Sperber 2002, 2004). Accordingly, 606 something is the ‘holy host’ (41). Rather, a wide array of assumptions would be 757 manifest to speakers regarding, for instance, and respectively, suffering, penance, 758 torture, etc.; pleasure, delight, joy, etc.; superiority, supremacy, etc., or failure, ruin, 759 etc., which they would associate with the employed noun or verb. Such association 760 might further be possible because the encyclopaedic entries of their encoded concepts 761 could contain information about attitudes, feelings or emotions usually caused by the 762 denoted referents, which would be similar to the positive or negative ones that the 763 speakers intended to express at the moment of speaking. Reiteration of such 764 connections could broaden the conceptual content of those words and lead them to 765 progressively allude to something with which they share some properties or which is 766 evoked by them. A similar broadening extending the concepts encoded by certain 767 nouns, adjectives and verbs to more peripheral properties as a result of metaphorical 768 or hyperbolical usages might also in part underlie their expressive usage. 769 Children’s overextensions fall into two types (Wałaszewska 2011: 321-322): 770 a) Over-inclusion or categorical/classic overextension (Clark 1973, 1993; Rescorla 771 1980), whereby words are applied to members of other categories within the 772 same or an adjacent conceptual domain (Wałaszewska 2011: 321). For example, 773 ‘doggie’ may also refer to cats, cows or horses owing to perceptual similarities – 774 four legs– and conceptual contiguity –being animals. 775 b) Analogical (over)extension (Clark 1993; Rescorla 1980), whereby words 776 belonging to a particular conceptual domain are applied to other objects with 777 which they only share some resemblance. For instance, the roundness of cookies 778 and the moon leads some children to refer to the latter as a ‘cookie’. 779 The proceduralisation of content words as expressives cannot be accounted for as 780 a result of analogical extension because of a lack of perceptual resemblance between 781 their initial denotation and emotional states. Despite clearly perceivable reflexes, 782 emotions are ineffable, internal and personal. Over-inclusion seems a better candidate, 783 then: the encyclopaedic entry of their encoded concepts could include information 784 about properties of the referents, reactions or emotions they may cause, or common 785 attitudes towards them, which may resemble, be applicable to, evoke or be evoked by 786 certain emotional states. Repeated access to this information could create a link in the 787 logical entry of the concepts initially encoded with new, non-lexicalised concepts 788 referring to emotional states. Thus, the information of the logical entry of the initially 789 encoded concepts –e.g. that such words denote individuals, actions or conditions– 790 would be ignored and those words could activate concepts like POSITIVE / NEGATIVE 791 EMOTION or more specific ones like HAPPINESS , SADNESS , PAIN , NUISANCE , etc. 792 Accordingly, an encoded concept like FUCK could be broadened to FUCK * and allude to 793 some emotion, even if that broadened concept would still need adjusting as 794 POSITIVE / NEGATIVE EMOTION or even as HAPPINESS , SURPRISE , WRATH or HATRED . In turn, 795 DAMN could be broadened to the emotion-related concept DAMN * and be subsequently 796 adjusted as PAIN or NUISANCE . The output of such adjustment will depend on 797 paralanguage, assumptions about the speaker, setting, recent events, objects or 798 individuals alluded to, which explains why a particular expressive may result in various, 799 distinct attitudinal descriptions (Blakemore 2011: 3542-3544; Wharton 2009, 2016). 800 A broadened concept would somehow capture the emotional state that the speaker 801 needs to express, while inferential adjustment would yield a more specific concept 802 capturing actual nuances of the emotional state in question, such as the specific type 803 of emotion or the intensity with which it is experienced. Although the broadening of the 804 conceptual content of the words giving rise to expressive APs and expletive NPs may 805 be seen as a case of over-inclusion triggered by the need to allude to something 806 elusive like feelings or emotions, the existence of secondary interjections such as 807 ‘fuck!’ or ‘damn!’, which have already stabilised as expressives in their own right in a 808 language or language variety, could also enact or facilitate that broadening. Indeed, 809 the expressive functions and meaning of many secondary interjections –which, like 810 expressive APs and expletive NPs, are words coming from various lexical categories– 811 may be due to an extreme broadening that enables them to refer to emotional states 812 (Author 2017). Once the process whereby the words giving rise to those interjections 813 is spread and stabilises in a language or variety, it would be relatively easy to repeat 814 and apply it to other words in order to get new expressive devices. 815 The concepts initially encoded by the words giving rise to expressive APs and 816 expletive NPs, then, would be broadened and get their denotation over-extended in 817 such a way that those words would become associated with emotional states. The 818 generalisation of that broadening among language users would favour its stabilisation 819 in communities of practice or geographical groups, and specialise those words for 820 expressive functions. The interpretation of those expressives would in some cases 821 require the construction of occasion-specific attitudinal descriptions, for which 822 activated emotion-related concepts would have to be adjusted; in other cases, the 823 stabilisation of a broadening would restrict the range of possible attitudinal descriptions 824 amenable to construction, as the resulting expressives would be associated with more 825 specific emotion-related concepts in a more stable manner. 826 827 6. Conclusion 828 Expressive APs and expletive NPs partake in the procedural semantics of expressives 829 in general, but differ from some of them as regards their output. Like slurs, they enact 830 attitudinal descriptions that only affect a propositional constituent. Accommodation of 831 these shorter-ranging mental schemas in the relevance-theoretic notional apparatus 832 may be done in two ways. One of them would involve refining the distinction between 833 lower- and higher-level explicatures by introducing a new label for them. These 834 descriptions have also been shown not to be solely triggered by the expressive devices 835 discussed in this paper, but also by metalinguistic negation, irony, prosody, diminutives 836 or some evidentials. 837 Terms of endearment or affection often used as vocatives in order to refer to the 838 addressee –e.g. ‘honey’, ‘darling’– could analogously be thought to give rise to similar 839 emotional-state descriptions. However, the format and place where those descriptions 840 are constructed within the mental representation of utterances are issues that still 841 deserve more detailed consideration and onto which future research could probably 842 shed some light. Arguably, if vocatives refer to the addressee, and reference to him is 843 normally made in the higher-level explicature, the emotional-state description that 844 vocatives result in would not take within its scope a propositional constituent, but a 845 constituent of the higher-level explicature. Therefore, that sort of conceptual schema 846 could not be subsumed by the label proposed in this paper, provided it was accepted, 847 and would require a new one reflecting this peculiarity. 848 Finally, the development of expressive functions by items from distinct lexical 849 categories has been suggested to result from a broadening affecting their semantics. 850 More specifically, that broadening has been argued to resemble an over-extension of 851 their initially encoded concepts similar to those children often do. Thus, this paper has 852 additionally provided support for the applicability of the relevance-theoretic postulates 853 on lexical pragmatic processes (Clark 2016). 854 855 References 856 Amenós Pons, José. 2011. Cross-linguistic variation in procedural expressions: 857 Semantics and pragmatics. In M. V. Escandell Vidal et al. (eds.), Procedural 858 Meaning: Problems and Perspectives, 235-266. Bingley: Emerald Group Publishing. 859 Anglin, Jeremy. 1977. Word, Object and Conceptual Development. New York: Norton. 860 Bezuidenhout, Anne. 2004. Procedural meaning and the semantics/pragmatic 861 interface. In C. Bianchi (ed.), The Semantics/Pragmatics Distinction, 101-131. 862 Stanford: CSLI. 863 Blakemore, Diane. 1987. Semantic Constraints on Relevance. Oxford: Basil Blackwell. 864 Blakemore, Diane. 1992. Understanding Utterances. An Introduction to Pragmatics. 865 Oxford: Blackwell. 866 Blakemore, Diane. 2002. Relevance and Linguistic Meaning. The Semantics and 867 Pragmatics of Discourse Markers. Cambridge: Cambridge University Press. 868 Blakemore, Diane. 2011. On the descriptive ineffability of expressive meaning. Journal 869 of Pragmatics 43: 3537-3550. 870 Blakemore, Diane. 2015. Slurs and expletives: A case against a general account of 871 expressive meaning. Language Sciences 52: 22-35. 872 Blass, Regina. 1989. Pragmatic effects of co-ordination: The case of ‘and’ in Sissala. 873 UCL Working Papers in Linguistics 1: 32-51. 874 Blass, Regina. 1990. Relevance Relations in Discourse: A Study with Special 875 Reference to Sissala. Cambridge: Cambridge University Press. 876 Börjars, Kersti and Kate Burridge. 2001. Introducing English Grammar. London: 877 Arnold. 878 Carston, Robyn. 1994. Metalinguistic negation and echoic use. UCL Working Papers 879 in Linguistics 6: 321-339. 880 Carston, Robyn. 1997. Enrichment and loosening: Complementary processes in 881 deriving the proposition expressed. Linguistiche Berichte 8: 103-127. 882 Carston, Robyn. 2002. Thoughts and Utterances. The Pragmatics of Explicit 883 Communication. Oxford: Blackwell. 884 Carston, Robyn. 2013. Word meaning, what is said and explicatures. In C. Penco and 885 F. Domaneschi (eds.), What Is Said and What Is Not, 175-204. Stanford: CSLI. 886 Carston, Robyn. 2016. The heterogeneity of procedural meaning. Lingua 175-176: 887 154-166. 888 Carston, Robyn and Eun-Ju Noh. 1995. A truth-functional account of metalinguistic 889 negation, with evidence from Korean. UCL Working Papers in Linguistics 7: 1-25. 890 Carston, Robyn and George Powell. 2005. Relevance Theory: New directions and 891 developments. UCL Working Papers in Linguistics 17: 279-299. 892 Clark, Billy. 2016. Relevance theory and language change. Lingua 175-176: 139-153. 893 Clark, Eve V. 1973. What’s in a word? On the child’s acquisition of semantics in his 894 first language. In T. Moore (ed.), Cognitive Development and the Acquisition of 895 Meaning, 65-110. New York: Academic Press. 896 Clark, Eve V. 1993. The Lexicon in Acquisition. Cambridge: Cambridge University 897 Press. 898 Collins, Peter and Carmella Hollo. 2000. English Grammar. An Introduction. London: 899 Palgrave. 900 de Saussure, Louis. 2012. Temporal reference in discourse. In K. Allan and K. 901 Jaszczolt (eds.), The Cambridge Handbook of Pragmatics, 423-446. Cambridge: 902 Cambridge University Press. 903 Escandell Vidal, María V. and Manuel Leonetti. 2011. On the rigidity of procedural 904 meaning. In M. V. Escandell Vidal et al. (eds.), Procedural Meaning: Problems and 905 Perspectives, 81-102. Bingley: Emerald Group Publishing. 906 Escandell Vidal, María V., Manuel Leonetti and Aoife Ahern (eds.) 2011. Procedural 907 Meaning: Problems and Perspectives. Bingley: Emerald Group Publishing. 908 Fremgen, Amy and David Fay. 1980. Overextensions in production and 909 comprehension: A methodological clarification. Journal of Child Language 7: 205- 910 211. 911 Gehweiler, Elke. 2008. From proper name to primary interjection. The case of gee! 912 Journal of Historical Pragmatics 9 (1):71-93. 913 Goffman, Erving. 1981. Forms of Talk. Oxford: Blackwell. 914 Gottfield, Gail M. 1997. Using metaphors as modifiers: Children’s production of 915 metaphoric compounds. Journal of Child Language 24: 567-601. 916 Greenbaum, Sidney and Randolph Quirk. 1993. A Student’s Grammar of the English 917 Language. London: Longman. 918 Grice, Herbert P. 1957. Meaning. Philosophical Review 66: 377-388. 919 Grisot, Cristina and Jacques Moeschler. 2014. How do empirical methods interact with 920 theoretical pragmatics? The conceptual and procedural contents of the English 921 Simple Past and its translation into French. In J. Romero Trillo (ed.), Yearbook of 922 Corpus Linguistics and Pragmatics 2014. New Experimental and Theoretical 923 Paradigms, 7-33. Cham: Springer. 924 Grisot, Cristina, Bruno Cartoni and Jacques Moeschler. 2016. Conceptual and 925 procedural information for verb tense disambiguation: The English Simple Past. In 926 M. Padilla Cruz (ed.), Relevance Theory. Recent Development, Current Challenges 927 and Future Directions, 103-143. Amsterdam: John Benjamins. 928 Haegeman, Liliane and Jacqueline Guéron. 1999. English Grammar. A Generative 929 Perspective. Oxford: Blackwell. 930 Hall, Alison. 2007. Do discourse connectives encode concepts or procedures? Lingua 931 117 (1): 149-174. 932 Hall, Alison. 2011. Ad hoc concepts: Atomic or decompositional. UCL Working Papers 933 in Linguistics 23: 1-10. 934 Hopper, Paul J. 1991. On some principles of grammaticalisation. In E. C. Traugott and 935 B. Heine (eds.), Approaches to Grammaticalisation. Volume I. Theoretical and 936 Methodological Issues, 17-36. Amsterdam: John Benjamins. 937 Hopper, Paul J. and Elizabeth C. Traugott. 1993. Grammaticalisation. Cambridge: 938 Cambridge University Press. 939 Huddleston, Rodney. 1988. English Grammar. An Outline. Cambridge: Cambridge 940 University Press. 941 Huttenlocher, Janellen. 1974. The origins of language comprehension. In R. Solso 942 (ed.), Theories in Cognitive Psychology, 331-368. Potomac: Erlbaum. 943 Ifantidou, Elly. 1992. Sentential adverbs and relevance. UCL Working Papers in 944 Linguistics 4: 193-214. 945 Ifantidou, Elly. 1993a. Sentential adverbs and relevance. Lingua 90: 69-89. 946 Ifantidou, Elly. 1993b. Parentheticals and relevance. UCL Working Papers in 947 Linguistics 5: 193-210. 948 Ifantidou, Elly. 2001. Evidentials and Relevance. Amsterdam: John Benjamins. 949 Itani, Reiko. 1990. Explicature and explicit attitude. UCL Working Papers in Linguistics 950 2: 52-64. 951 Itani, Reiko. 1994. A relevance-based analysis of hearsay particles: Japanese 952 utterance-final tte. UCL Working Papers in Linguistics 6: 379-400. 953 Itani, Reiko. 1998. A relevance-based analysis of hearsay particles: With special 954 reference to Japanese sentence-final particle tte. In R. Carston and S. Uchida 955 (eds.), Relevance Theory. Applications and Implications, 47-68. Amsterdam: John 956 Benjamins. 957 Jucker, Andreas H. 1993. The discourse marker well: A relevance-theoretical account. 958 Journal of Pragmatics 19: 435-452. 959 Jucker, Andreas H., Sara W. Smith and Tanja Lüdge. 2003. Interactive aspects of 960 vagueness in conversation. Journal of Pragmatics 35: 1737-1769. 961 Kecskes, Istvan. 2004. Editorial: Lexical merging, conceptual blending, and cultural 962 crossing. Intercultural Pragmatics 1 (1):1-26. 963 Kecskes, Istvan. 2008. Dueling contexts: A dynamic model of meaning. Journal of 964 Pragmatics 40: 385-406. 965 Moeschler, Jacques. 2016. Where is procedural meaning located? Evidence from 966 discourse connectives and tenses. Lingua 175-176: 122-138. 967 Mustajoki, Arto. 2012. A speaker-oriented multidimensional approach to risks and 968 causes of miscommunication. Language and Dialogue 2 (2): 216-243. 969 Navarro Reyes, Jesús. 2009. Can we say what we mean? Expressibility and 970 background. Pragmatics & Cognition 17 (2): 283-308. 971 Oppermann-Marsaux, Evelyne. 2008. De l’injonction a l’interjection. La naissance des 972 marqueurs discursifs tien(s) et tenez en français medieval (XIIe-XVe siecles). 973 L’Information Grammaticale 118: 11-15. 974 Author. 2009a. 975 Author. 2009b. 976 Author. 2016. 977 Author. 2017. 978 Author. Forthcoming. 979 Potts, Christopher. 2005. The Logic of Conventional Implicatures. Oxford: Oxford 980 University Press. 981 Potts, Christopher. 2007a. The expressive dimension. Theoretical Linguistics 33 (2): 982 165-197. 983 Potts, Christopher. 2007b. The centrality of expressive indices. Theoretical Linguistics 984 33 (2): 255-268. 985 Raeber, Thierry and Tim Wharton. 2017. Intensifying and expressing: Two sides of the 986 same coin. Paper presented at the Beyond Meaning International Conference. 987 Athens, 13-15 September. 988 Rescorla, Leslie A. 1980. Overextension in early language development. Journal of 989 Child Language 7: 321-335. 990 Rouchota, Villy. 1995. Discourse connectives: What do they link? UCL Working Papers 991 in Linguistics 7: 199-212. 992 Sasamoto, Ryoko and Rebecca Jackson. 2016. Onomatopoeia – Showig-word or 993 Saying-word? Relevance Theory, lexis, and the communication of impressions. 994 Lingua 175-176: 36-53. 995 Sasamoto, Ryoko and Deirdre Wilson. 2016. Little words: Communication and 996 procedural meaning. Lingua 175-176: 1-4. 997 Schröder, Helga. 2016. Reference assignment in pronominal argument languages: A 998 relevance-theoretic perspective. In M. Padilla Cruz (ed.), Relevance Theory. Recent 999 Development, Current Challenges and Future Directions, 81-102. Amsterdam: John 1000 Benjamins. 1001 Scott, Kate. 2011. Beyond reference: Concepts, procedures and referring expressions. 1002 In M. V. Escandell Vidal et al. (eds.), Procedural Meaning: Problems and 1003 Perspectives, 183-203. Bingley: Emerald. 1004 Scott, Kate. 2013. This and that: A procedural analysis. Lingua 131: 49-65. 1005