Zs. Schnell (1) -Varga, E. (2) -Járai, R. (3

Zs. Schnell (1) – Varga, E. (2) – Járai, R. (3) (1) Institute for Cultural and Social Sciences, (2) Dept. of Psychiatry and Psychotherapy, Medical School (3) Institute of Psychology, University of Pécs, Hungary Contextual awareness in pragmatic meaning construction – An experimental study Abstract Research focus Our experimental pragmatic study aims to explore the role of contextual cues such as prosody in children’s understanding of irony and explain why prosodic patterns, functioning as surface cues in comprehension, facilitate non-compositional meaning construction at an early age. Research methods We investigate preschoolers’ linguistic performance with tasks testing the understanding of literal (semantic-) and pragmatic (context-based, intended) meaning. Our trials are based on simile, metaphor, three types of verbal humor-, irony, irony with surface cue and control tasks, and a nonverbal humor task). The linguistic tasks are viewed in relation to children’s mentalization skills, that is, their ability to understand intentions and desires attributed to the self and others. This social-cognitive skill we see as crucial in the unfolding of pragmatic competence is tested with verbal and nonverbal False Belief Tests (Baron-Cohen – Leslie – Frith 1985) to see if mentalization ToM skills predict success in pragmatic skills. Results Our results indicate that contextual cues in fact facilitate understanding, but the lack of context does not automatically ensure difficulty in interpretation, especially in nonverbal conditions. Irony, surprisingly, is processed with more success by preschoolers than metaphor or humor, due to its salient prosody and intonation patterns that serve as contextual surface cues in interpretation. In the case of metaphor, contextual constrains have a moderate facilitating effect, while in the case of irony contextual cues that target the mental state of the speaker specifically, have a significant facilitating effect in interpreting implicit meaning. When contextual cues make the implicit mental terms explicit, semantic processing suffices and no mentalization is required in utterances with implicit meaning. Main conclusion Contextual effects thus may vary according to what they target, and the successful deciphering of implicit meaning is even more strikingly influenced by contextual cues than in literal language. Our findings confirm results in developmental research that surface cues help in the recognition of communicative intent and contribute to the successful resolution of the intended meaning at hand. Keywords: mentalization, irony, nonverbal modality, surface cue 1 1. Introduction and purpose of the study The experimental investigation in the present dissertation aimed to give a bird’s eye view of the unfolding of pragmatic competence and identify the cognitive background mechanisms of this process, among these the cognitive ability that is essential in non-compositional, holistic meaning construction. Overall, the crucial role of mentalization was supported in trials measuring metaphor, humor, irony processing and conversational abilities. In the case of humor and the conversational maxims the mindreading ability is further completed by complex social and cultural knowledge, together with memory skills improving in school years. 1.1. The psycholinguistics of non-compositional meaning construction The first chapter reviews some of the theoretical issues surrounding metaphor, delineating the major models of metaphor processing, their strong claims and their weaknesses. Pragmatic competence is defined as the ability to understand non-compositional constructions, where the central component of the interactive process of interpretation is based on theory of mind, stemming from intentionality, which enables the hearer to decipher speaker’s intention, and thus to decipher intended meaning. The cognitive developmental background of the unfolding of this competence is sketched, and an empirical test of the model is outlined, targeting the role of mentalization in metaphor processing. In the mentalistic model proposed in the present study (Fig. 1.), mentalization plays a central role, as this is a basic skill crucial in both the semantic and the pragmatic aspects of language acquisition that may account for the ease and the equal reaction times in current psycholinguistic findings (Gibbs—Nayak—Cutting 1989, Gibbs 1994). Context Holistic Interpretation Mentalization Figurative meaning Conventionality Semantic analysis Decompositionality Figure 1. The mentalistic model, building on mentalization 2 2. Empirical study on idiomaticity in view of social cognition The study tests pragmatic competence with metaphor tasks (see Appendix), while the simile condition served as the control condition, given that in the case of similes, the word ‘like’ makes the comparison explicit. In metaphor, however, it stays implicit, where in deciphering intended meaning one needs to rely on intention reading, thus ToM abilities. The empirical study tested preschoolers’ mentalization skills with a puppet-play adaptation of the classic false belief task (Wimmer—Perner 1983) based on unseen displacement in the framework of Baron-Cohen’s Sally-Anne test version (1985) (Fig 2a,b.). Fig. 2 a,b The FBT testing mentalization skills 2.1. Methodology Different ToM tasks were administered (1.: Verbal first order, verbal second order, 2: Nonverbal Eyes test) to preschoolers individually, in a private room on their kindergarten. They were shown the puppet play of unseen displacement that provided the ToM test and yielded a group of passers (ToM group), and failers (NoToM group) in the verbal tasks, which was contrasted with the Eyes test to reveal the latter’s validity and reliability. 2.1.1. ToM tests Verbal – First-order ToM test: ToM question: Where will the Brown Bear look for the honey? Memory question: where did the Brown Bear put the honey in the beginning? Reality question: Where is the honey now? Only if the participant answered all 3 questions correctly, was the point given for first-order ToM. 3 Verbal second-order ToM question: Where does the White Bear think that the Brown Bear thinks the honey is? 2.1.2. Eyes test story The participants were read out the scenarios individually in person, and were shown the eyes to choose from for the reading the mind in the eyes task. Scenario example: Peter and Dorothy got a puppy for Christmas. They were so happy, they even slept with it! One day when they woke up, they did not find their favorite puppy. “It must have got lost”! – they thought. Target sentence (question) at the end: How did Peter feel the morning he did not find his puppy? (Which picture shows his eyes?) (Child has to point at the picture). Eyes Test (Schnell 2011) 6 basic emotions The performance of the two ToM groups was compared in the linguistic tasks targeting different aspects of pragmatic competence (Simile-, Metaphor-, Humor-, Irony- processing). 4 The sample consisted of 45 (final: 41) preschoolers, between the ages 3-6, 19 boys and 26 girls, of a mean age of 5;1 years (ranging from 3;7 to 7;3). Pragmatic tests contained 5 tasks in each condition, 1 point was earned for all correct answers in each condition. The findings confirmed that theory of mind is a core component of metaphor processing (Fig. 3a,b.). Fig. 3a. Simile and Metaphor comprehension in view of ToM skills The ToM group was significantly more successful in the metaphor condition where due to its implicit nature, the deciphering of intended meaning required inferential strategies, i.e. ToM (F(1,43)=134, pmetaphor<0,01). There was no interaction between the simile and the metaphor conditions. In the simile (control) condition (F(1,43)=0,5, psimile>0,05), where semantic interpretation suffices, there was no significant difference in the performance of the two groups, suggesting that pragmatic competence is largely dependent on the ability to mentalize, which enables the efficient decoding of intended, figurative meanings. 5 Estimated Marginal Means of MEASURE_1 csoport 5,00 Estimated Marginal Means noTOM TOM 4,00 3,00 2,00 1,00 1 2 Simile / Metaphor Fig.3b. Idiomatic and pragmatic skills in line with ToM competence 1.: Simile condition 2.: Metaphor condition 3. Mentalization and the development of humor comprehension The next chapter targeted cognitive interpretative processes of non-literal meaning in general, and of humorous utterances in particular. The continuum of mentalization and humor is outlined in identifying the cognitive, linguistic and psycholinguistic background of humor research. The same empirical test used in the idiomaticity chapter was extended with 3 verbal humor trials (decontextualized: one-liners and riddles, and a contextual trial involving classical jokes ending in punchline) and 1 nonverbal, visual based humor task, together with a second-order verbal ToM test and an Eyes test of semi-verbal nature (see App.), in order to see if verbal performance limitation is significant in the classical false belief tests of mentalization, and to see the interaction of verbal- and nonverbal results in the pragmatic tasks. 3.1. Validity and reliability of the Eyes test The Eyes test was not designed to break down the sample into further groups of ToM competence, rather, it was included to see if the nonverbal variety can be predictive and passed earlier, ensuring the avoidance of performance limitation. The results seem to back up the expectations, supporting a moderate performance limitation of verbalization. While in verbal ToM tests children were successful at the age of 5;4, in the nonverbal test measuring complex 6 mentalization mean age of passers was 5. This suggests that by eliminating verbal performance limitation, the nonverbal Eyes test can be completed 4 months earlier (Sullivan 1994). Our eyes test thus proved to be valid and predictive concerning mentalization abilities, even though it was aimed to be less demanding due to its semi-verbal nature (Fig. 4.) Fig. 4. Area under RoC curve showing the diagnostic accuracy of the Eyes test Somers’d measure in the statistical analysis has shown that there is a significant difference in the performance of the two groups in terms of Eyes test results Somers’s d T=4,503; p< . 001 suggesting that the ToM group is significantly better in the Eyes test, which confirms its predictive value and efficiency. 3.2. Humor processing in view of mentalization The empirical investigation aimed to identify if mentalization is an important factor in humor processing, since ToM subjects performed better in humor tasks, however, ToM in this particular case does not seem to be sufficient, since the ToM group’s performance was weaker in the humor trials than in the metaphor trial F(2,82)= 222,29, p< 0,01, Interaction: F(2,82)=22,97, p<0,01. (Fig. 5a,b.). 7 This suggests that beyond ToM, social and cultural factors also contribute to the acquisition of productive discourse skills in terms of humor (Pexman et al. 2005). Nonverbal results (Fig. 7.) support the Cognitive Congruence principle in that those in the NoToM group preferred nonverbal humor trials, and performed weaker in both verbal humor conditions, while the ToM group performed better in both verbal humor tasks, and was least successful in the nonverbal trial F(2,82)=2,744; p<0,1. csoport 5,00 Estimated Marginal Means noTOM TOM 4,00 3,00 2,00 1,00 1 2 3 Simile hasonlatmetaforahumor / Metaphor / Humor Fig 5a 1. Simile 2. Metaphor 3. Humor processing in relation to ToM skills 8 1 2 3 4 5 6 7 8 9 Fig 5 b. 1. Simile 2. Metaphor 3. Humor processing in relation to ToM skills 1 – Eyes Test; 2 – H - One-liners, 3 – H-Riddles, 4 – H-Jokes, 5 - H-visual, 6 - Similes (Dctx), 7 – Simile (Ctx), 8 – Metaphor (Dctx), 9 – Metaphor (Ctx). 00: NoToM group (blue, left side) 01: ToM group (green, right side) 4. Contextual effects in the metaphor and humor condition The decontextualized condition measured production, while the story condition focused on perception skills. We found no significant difference between the two conditions (Ctx/Dectx), meaning, that production and perception actually correlate, and thus measuring one or the other shall suffice and is predictive in terms of judging competence. 4.1. Contextual effects in the idiomaticity trial Figure 6. shows contextual effects in idiomatic processing. Contextual effects support the basic psycholinguistic axiom that contextual cues facilitate comprehension, since in contextual tasks context enhanced subjects’ performance F(1,41)= 0,228 n.s. (Fig. 6.). 9 csoport 4,50 Estimated Marginal Means noTOM TOM 4,00 3,50 3,00 2,50 1 2 Decontextualized idiomaticity trials / Contextualized id. trials Fig. 6. Contextual effects in idiomatic processing in view of mentalization (1) Decontextualized (multiple choice) (2) Contextual (stories ending in metaphor / simile) There is no significant difference between decontextualized and contextualized stimulus processing, only a slight advantage on the side of longer, contextual input. In other words, the contextual conditions were slightly more successful in both groups, which in the case of the metaphor tasks confirm the facilitating effect of the context in interpretation. The contextual cues apparently help in the resolution of implicatures and in the selection of the intended meaning out of competing alternatives F(1,41)= 0,228 n.s. All in all, contextual cues of this general, linguistic descriptive type, together with situational constraints apparently do facilitate interpretation, but they do not make a significant difference in results. 4.2. Contextual effects in the humor condition In humor tasks context’s facilitating effect (Fig. 7.) is apparently present in the NoToM group’s performance, where probably, in short of a fully-fledged ToM contextual information facilitates comprehension. Fig. 7. also reflects contextual effects in humor processing: the ToM group was more successful in short, decontextualized humor (Jean-jokes or One-liners) than in the contextualized humor task (i.e. jokes ending with a punchline). It may be because longer context may burden processing in humorous utterances where cognitive effort is greater than in semantic processing, because the resolution of incongruity requires an interactive dynamic 10 meaning construction; while short verbalized input, being more congruent in cognitive terms, triggers a more optimal effect. Independent of tasks we did find a difference between the ToM and the NoToM group: F(1,41)=7,612, p<0,01; meaning, the ToM group and the NoToM group rely on different cognitive strategies. This proves that the ToM group makes use of a different mental strategy and relies on mentalization. csoport Estimated Marginal Means noTOM TOM 2,00 1,50 1,00 1 2 3 Decontextualized / Nonverbal / Contextualized trials Fig. 7. (1) Decontextualized (2) Nonverbal and (3) Contextualized humor comprehension skills in view of mentalization The Eyes test maintains the possibility of a wide range of further investigations. Future methodologies building on the Eyes test can give even more insight into the interaction of nonverbal mentalizing skills and nonverbal inferential processes, and possibly clarify issues surrounding the levels of pragmatic competence and its associations to performance limitation. 5. The development of irony comprehension in view of mentalization The chapter on social cognitive abilities and irony briefly summarized the psycholinguistics background of irony, touching on some definitional problems in theory and research, discussing its distinctive features which make it salient in interpretation, and outlined its differences compared to humor (Wilson – Sperber 1992, Wilson 2009). The empirical test aims to contribute to the resolution of long-standing debates on irony processing (Wilson 2013). Its 11 methodology is based on the previous chapters’ procedure, extended with three irony tasks: genuine contextual ironies based on hyperbole, an irony with linguistic task, and a control task. Irony and humor are treated as dichotomies in linguistics, whereas in humor research the two form rather fuzzy, blurry, overlapping categories. The present findings may clarify this issue, confirming that irony triggers different cognitive processes than humor processing. Children were successful in irony, but a lot less competent in the humor tasks (see Fig. 8). This suggests the detachment of irony and humor, although as for social function, the two may serve similar or converging goals. All figures of speech, after all, tend to have similar functions: they help formulate abstract phenomena we cannot otherwise understand (Lakoff-Johnson 1980), conveying playfulness, poetic style, and wittiness in communication. As Gibbs—Gregory—Colston (2014) point out, finding the humor in irony is not the same as in simple jokes, and demands the examination of a complex host of contextual factors not always considered in linguistic theories of humor. Irony is not necessarily difficult (ibid), confirming the findings of the present study, that situational cues set the stage for a directly available, mentalization based inferential interpretation. The findings indicate that irony in fact is processed fairly early in preschool years (Fig. 8a). It was the easiest of all the non-compositional constructions tested (Fig. 8b.). 12 1 2 3 4 5 6 7 8 Fig.8a. Relationship of [V-ToM skills] and [NV-ToM, Humor and Irony competence] 1 – Eyes test, 2 – H-One-liners, 3 – H-Riddles, 4 – H-Jokes, 5 – H-Visual, 6 – Irony, 7 – Irony with lg. help, 8 – Control (Irony). 00: NoToM group (blue, left side); 01: ToM group (green, right side) Maxim of Quantity Maxim of Humor Manner Maxim of Quality Maxim of Relevance Metaphor Irony Simile Irony + lg. help Literal level (Control task) Fig. 8b. The relative place of irony in the continuum of non-compositionality, in view of ToM 13 A Mann Whitney analysis of the mean ranks of two independent samples indicated that there was a significant difference in the performance of the NoToM and the ToM group in Irony tasks, i.e. the difference in results is due to their mentalization ability. U=115,5, p= .019, r= .36. In the irony with linguistic cue task U=150, p= .104, r= .25; there was no significant difference in the two group’s performance (mean rank for the ToM group being 24, for the NoToM group: 19), but this difference is visibly bigger than in the case of the control task. This suggests that the surface cue providing explicit information about the speaker’s mental state indeed facilitated non-compositional meaning construction, and the NoToM group was more successful in interpreting this, explicitly put irony than the genuine irony. This confirms the role of mentalization in irony processing and in pragmatic competence, along with the facilitating effect of surface cues, and context in interpretation. In the control condition requiring only semantic interpretation the analysis indicates that there is no significant difference in the results of the two groups, 20,44 being the mean rank in the NoToM group, whereas 21,88 in the ToM group. This shows that the difference in mean performance is very small, meaning, that the NoToM group, in short of a fully fledged mentalizing ability is equally successful in semantic tasks, as the ToM group, since in this trial no mentalization and thus no pragmatic meaning construction was needed, and semantic interpretation sufficed: U= 186, p= .664, r= .067. In general humor is apparently more difficult for preschoolers than irony (Fig. 8a), and metaphor (Fig. 5a,b), see also Fig. 8b. for relative difficulty. The relative placement of the different types of non-compositional constructions yields a developmental trajectory of pragmatic development: among the three targeted forms of non-compositionality, the easiest to handle is apparently irony, then comes metaphor, and eventually humor, for which a number of social and contextual situational, and cultural factors need to be considered in order to explain full productivity (Colston—Gibbs 2002, Pexman et al. 2005). Since in the case of irony the comparison to be drawn is based on conflict of meanings, a shortcut strategy seems to be employed by subjects, which also explains early success in comprehension. The Linguistic cue condition significantly facilitated interpretation, since the significant difference in the two groups’ performance disappeared. The control conditions required only semantic processing, and as expected, there was no significant different in the performance of the two groups. Irony is apparently easier for preschoolers than metaphor, due to its distinctive features (ironical tone of voice, mocking attitude, normative bias) which function as ostensive cues 14 (Csibra 2010) and trigger the shortcut strategy (Győri et al. 2002) based on a heuristics of taking the opposite meaning (Fig. 9.). Fig. 9. Metaphor vs. irony processing in view of mentalization Lower line (00): NoToM, Upper line (1,00) ToM group 1- Decontx. metaphor; 2- Contextualized metaphor; 3- Irony, 4- Irony with surface cue; 5Control task (irony condition) Significant differences were found in the performance of the two groups in the irony task, confirming the key role of ToM in irony processing. ToM group was significantly more successful in the irony task than the NoToM group. Mann Whitney U=115,5; p<0,01. In the irony with surface cue condition no significant difference was found between the performance of the two groups MWU=150; p= .104; Contextual cues made the mental state of speaker explicit, thus, semantic interpretation suffices (Csibra 2010); and also in the control condition, the No-ToM group was equally successful as the ToM group. MWU=186; p= .664. since in these scenarios no human agents participated and thus children did not have to mentalize about intentions and intended meanings. A pure semantic interpretation suffices. Apparently metaphor is more complex, probably due to the emergent properties involved in metaphors, where comparison is based on an optimal relevance based traits of similarity, making it a more graded conceptual phenomenon (Colston—Gibbs 2002, Wilson – Carston 2006). The findings reveal that metaphor and irony performance of the ToM and NoToM group differ significantly. Metaphor is more difficult than irony. p<0,01; mean: 2,503; (F) 4,156=44,305 Mixed Way ANOVA. Context facilitates comprehension – in ToM group 15 contextual stimuli was decoded with more success; which supports Cognitive Congruence (Zigler et al. 1966). Irony also proved to be easier to process than humor, which is again indicated to be a more complex competence where incongruity is not purely linguistic but often broader, rather contextual, based on social-constructions and cultural patterns that come with socialization and with age, thus delegating productivity in humor comprehension to early school years (Gibbs 1994, Pexman et al. 2005). The Irony with linguistic help trial also confirmed findings of the facilitating effect of contextual cues in interpretation. Eventually a continuum of pragmatic phenomena is drawn up reflecting the levels of cognitive difficulty and mentalization based inferential activity required in each pragmatic phenomenon. The pyramid thus formed reveals the order of the emergence of the examined non-compositional constructions in development, from similes, through metaphor, irony, to humor. 5.1. Some thoughts on the irony results The early success in irony tasks may in part be explained by ostensive and behavioral cues proposed by Csibra (2010) who claims that infants demonstrate pragmatic abilities in the resolution of the referential ambiguity of nonverbal gestures and of verbal utterances prior to verbal abilities, as they capture the holistic, intention based message of utterances, before they handle meaning. Irony involves an ironic situation, where ambiguity is present in several modalities: in the characteristic tone of voice, in situational and contextual cues and in the social setting itself. Infants seem to have an inclination, an instinct to detect such cues that modify meaning. Csibra (2010) claims that infants recognize the presence of communicative intentions in others’ behavior before the actual content of these intentions is accessed or inferred. Thus, they seem to rely on specific cues, the distinctive features of irony (Wilson 2009, 2013), among these ostensive signals, the decoding of which enables them to detect the communicative intentions of others. This attitude, rooted in the so-called pedagogical stance (Csibra 2010) ensures that infants and children rely on dedicated mechanisms that enable efficient social communication. Such strategies may account for the ease with which children handled irony, compared to other non-compositional constructions, in our study. Metaphor and irony are known to involve different processes of comparison: similarity vs. contrast, respectively. This suggests that irony, since it is the exact and sheer opposite of what is said, may be interpreted with different heuristics, on the grounds of the “infer the opposite meaning” heuristics, where slight ironic intonation is enough to trigger such algorithm. 16 6. Contextual effects in the irony condition The Irony with linguistic help condition provided surface cues, which can also be seen as being contextual cues. This factor significantly improved performance of subjects in the irony with linguistic help condition, compared to the implicit irony condition (see also Varga et al. 2013). The facilitating effect of the linguistic surface cues confirms that context facilitates pragmatic processing and comprehension in general, since all children’s performance was better in the linguistic help condition. Therefore, the facilitating effect of context has been born out in all three (metaphor, humor and irony) conditions as well, confirming its central role in holistic interpretation. As expected, no significant difference was found between the performance of the two groups in two cases: in the ‘Irony with linguistic help’ and in the ‘Control’ condition. In the Irony with linguistic help condition the implicit meaning was made explicit with the overt formulation of the speaker’s mental state (‘angrily’ said) – therefore, a semantic interpretation sufficed for the correct decoding of the ironic remark. Due to the linguistic help, the extra contextual cue consisting of the explicit formulation of the speaker’s mental state, the meaning was not implicit anymore, which significantly facilitated the inference of the intended meaning. The NoToM group and the ToM group were equally successful in this trial, i.e. no significant difference was found in the performance of the two groups. The NoToM group was obviously successful in this because the contextual aid, namely, the linguistic surface cue facilitated pragmatic interpretation. In the control condition (Fig. 5ab, 8a): As expected, this trial proved to be equally easy for both groups, since in these cases no human agents participated and thus children did not have to mentalize about intentions and intended meanings. A pure semantic interpretation sufficed in this case. Significant differences were found in the performance of the two groups in the irony task, confirming the key role of mentalization in irony processing, given that the ToM group, i.e. the children who passed the False Belief test requiring mentalization skills, was significantly more successful in the irony task than the NoToM group, who failed the test of mentalization based on the deciphering of intentions and false beliefs. This social cognitive skill we call mentalization or ToM skills enabling people to decipher intentions and therefore, the intended meaning o the utterance, correlates with success in irony comprehension. This suggests that mentalizing abilities do facilitate the interpretation of ironic utterances where the intended meaning is not the literal meaning but, in the case of irony, it is the opposite (see Fig. 5ab, 8ab). The NoToM group was apparently successful in the Irony task, compared to the rest of the pragmatic tasks of metaphor and humor. Surprisingly, the Irony condition was the easiest task measuring pragmatic skills, despite data in literature that it requires a second order theory of mind, 17 namely, the ability to understand not only where Sally would look for the ball (first order ToM), but also where Anne thinks Sally would look for the ball, in other words, Anne’s thoughts about Sally’s beliefs and intentions. In our study irony was fairly predictable which might be because of the salient features (Wilson 2009, 2013) that distinguish irony, and the heuristics (Csibra 2010) applicable for irony like “take the opposite meaning” triggered by some surface cues as ironic tone and intonation, mocking attitude, among others, functioning as distinguishing features of irony (Wilson 2009, 2013). This suggest that children in fact handle irony at an earlier age than previously demonstrated (Gibbs 1994), hypothesizing an irony processing skill at age 8-13 (Fig. 8,9). All in all, the findings suggest a continuum between literal and non literal language, in line with pragmatic literature (Sperber—Wilson 1986, 2002, Abu-Akel – Shamay-Tsoory 2011). Literal meaning may sometimes be ambiguous, and therefore, the intended meaning is not always strictly the literal or the idiomatic one (Giora 2001). Our findings thus support Giora’s views (2001) on the graded salience hypothesis. Literal and non-literal domains of speech, therefore, are not treated as a dichotomy in our research. As for cognitive strategies, pragmatic interpretation definitely shows a continuum of cognitive effort needed in the different cases of non-compositionality, which also boils down a continuum of contextual effects in the comprehension and productive incorporation of utterances. ToM abilities seem to make no significant difference in some domain (Similes, Irony with linguistic help, Control condition) exactly because in these, semantic interpretation is sufficient for the deciphering of intended meaning. The simile and the irony with linguistic help trial are two conditions where the implicature is made explicit, and therefore, literal interpretation suffices. These constitute the initial stages of the wide division line forming a continuum between literal and not literal domains. ToM is needed and is sufficient in other cases: (e.g. Irony, Metaphor). Still, in some other cases ToM cannot, in itself, account for non-compositional interpretation: in cases of Humor. Humor is apparently more difficult for children to process than metaphor (Fig. 5ab, 8a), but surprisingly, irony was easier for children than metaphor (Fig. 8, 9). This contradicts previous findings in irony research (Gibbs 1994), which claim that irony is mastered around ages 8-13. 7. Interpretation of contextual effects in the study Our results demonstrate that contextual effects differ in different contexts. This may seem like a pun, but int different modalities, in different constructions whose interpretation is influenced by mentalization to a different extend, the context of agents vs. inanimate objects all provide different factors of interpretation in a simultaneous process of human communication integrating several linguistic, psychological and social cognitive factors. 18 Although in the Simile-Metaphor (Figure 6) and in Humor tasks (Figure 7) the facilitating effect of context is not as significant as in the Irony condition (Figure 9), it is easy to see in Figure X showing metaphor results (Figure 6) that there is a very mild facilitating effect in the contextual metaphor condition, where the interpretation of metaphorical meaning was supported by the context, versus in the decontextualized condition, where the target sentence was more isolated, thus figurative meaning was not anchored and confirmed by contextual and surface cues. This slight facilitating effect is not significant statistically F(1,41)=0,228 n.s, but is visible in the chart, and we can definitely see and feel the effect in real-life discourse. In the case of humor results the effect of context appears as a trend-like interaction (Figure 7; F(2,82)=2,744, p<0,1). The question whether contextual effects in the nonverbal modality influence interpretation in different ways, the effect of modality on the task and on cognitive processes, also the effect of context in the verbal tasks themselves require further research to see the clear picture integrating subtle effects of linguistics, social-cognitive and other factors that serve and function as cues and therefore, influence contextual effects. 7.1. Context as a provider of optimal relevance In the Humor tasks (Figure 7) the longer context was easier for the preschoolers, and in the two decontextualized conditions their performance was lower in both groups. This supports the facilitating effect of context, however, also the humor tasks show a very interesting result, namely, that the visual and in the verbal modality performance on the tasks was not predictable on the basis of mentalization skills, as it was most preferred in the No-Tom group, compared to the verbal humor conditions, and it was the least preferred in the ToM group, i.e. for those preschoolers who passed the false belief test (FBT) and thus are seen as having a fully-fledged theory of Mind and mentalization skills. It is not easy to interprets such intertwined results and clearly identify cognitive processes and contextual factors of interpretation. We have many variables in the study, all the more, because the psycholinguistic principle of the facilitating effect of context claims that contextual information functions as cues and constraints to narrow down meaning, and thus facilitate interpretation in general; but, we all know that too much information that is not fully relevant, can actually hinder efficient interpretation and thus be distracting in the process of meaning construction – this is what the Maxim of Quantity stipulates in the Gricean framework of cooperative stances discourse. This itself is ambivalent, and thus the close examination of context always has to keep the principle of optimal relevance in the foreground, and see the effect of context as possibly having an effect of hindering or in an optimal scope and range, if 19 it targets the intended meaning then it can facilitate the deciphering of intended meaning and even pragmatic interpretation. This ambivalence, however, is not surprising in the field of cognitive science and psycholinguistics, and thus this duality of contextual effects manifests itself in our results as well (see Fig. 7). 7.2. Context as modality - Verbal vs. Nonverbal (visual) All in all, the results are weaker in the no-Tom group, in the nonverbal condition, but the fact that those kids who have not passed the FBT preferred visual humor to verbal, whereas the kids who passed FBT liked it the least, calls for a hypothesis of interpretation concerning cognitive congruence of stimuli and mental structures. This interesting finding, namely, that older children who have a fully-fledged mentalization skill prefer verbal humor to nonverbal one, is supported and may be explained by the hypothesis of Cognitive Congruence (CC) (Zigler et al. 1966) stating that the most enjoyable cognitive challenge is always the one that is most congruent with the given cognitive schemes of the person involved. In other words, those who already developed mentalizing competencies, are at a higher level in social cognition and also have linguistic skills find it easier to handle jokes and utterances at a more symbolic, i.e. verbal level, whereas visual humor might provide these more linguistically advanced participants too little context or not enough cues, compared to the verbal condition. All in all, the principle of Cognitive Congruence states that if the complexity of cognitive schemes are in harmony with the cognitive complexity of the given task, then the task is optimally challenging for the person, this is most enjoyable and may be preferred. This is what happened in the visual humor condition, where the nonverbal was highly preferred by the No-Tom group, but dispreferred by the ToM group of kids with stronger linguistic and social cognitive skills. This may be explained by the fact that Martin (2007) and Séra (1983) have pointed out, that children start to enjoy verbal jokes later in development, as they slowly dispose of linguistic skills, they can handle complex verbal tasks more easily and more readily. They thus become more competent in verbally and mentally flexible situations, tasks and are thus ahead in their natural discourse skills as well, where we have long known that more is hinted at than explicitly stated (Grice 1957, 1975). This explains why the No-Tom group, of less fully-fledged social cognitive skills that were not sufficient to pass the FBT preferred nonverbal and some decontextualized, more isolated utterances and jokes. These were more congruent with their actual cognitive schemes. Here the principle of CC actually brings in a factor of optimality, which might clear the picture for us, in that optimal relevance explains why in some instances 20 less context can be more relevant and thus facilitate interpretation, whereas in other tasks or conditions more context might be facilitating the anchoring of meaning. Optimality is effective in guiding interpretation and serves as an efficient cue in narrowing down meaning, especially in the social use of language as in pragmatic utterances, that heavily build on social-cognitive abilities (Sperber—Wilson 1986, Schnell 2007, 2015, 2012, 2019, Schnell et al. 2019). Social and situational context largely determine interpretational outcome (aggressive humor is often seen as insult humor or irony – (for definitional problems on this topic see chapter 7). Implicit meaning evokes interpreting strategies and inference. In fact, Sperber and Wilson argue for an optimal relevance in interpretation, serving as a guideline in the myriads of factors to consider in non-compositional meaning construction. Therefore, pragmatic decoding is not argued to be a plain synthesis of holistic, chaotic mass of impressions and inputs from linguistic, social, cognitive domains, but an optimal relevance determines the rule of thumb in inferencing. Similarly, Giora (2001) emphasizes that jokes and humorous instances often capitalize on the activation of the “least relevant” meaning. Jokes are stories ending with a punchline, so it’s easy to see how discourse coherence matters in interpretation. Going further, inference helps in the dissolution of incongruity, creating imagination based cognitive strategies that also contribute to the decoding of meanings (Schnell 2010, 2012). Figure 7. shows not only contextual effects but also the effect of the modality on interpretation: verbal vs. nonverbal. The verbal tasks include a decontextualized condition (Riddles and Jean-jokes) and a contextualized one as well (classic jokes with longer context ending in a punchline). The results here, thus show not only contextual effects but also the effect of modality, which can also be seen as a type of context, of course. If we closely examine the graph, and we interpret it in the visual vs. verbal modality framework, we can claim that the verbal tasks are more readily taken as contextual, (given that verbal context is a classic type of context). In both verbal variables the results of the ToM group are higher, so this confirms that in humor comprehension, for those with stabilized social cognitive skills the verbal context facilitated interpretation to a greater extent, compared to the context given by visual modality. The effect of verbality as a contextual effect and cue is confirmed here, at a level confronting modalities. In between modalities verbal context facilitates interpretation to a greater extent than nonverbal context, for those who have fully-fledges mentalizing and thus developed discourse skills. This result highlights the effect of context that goes along with verbal modality, stemming from the verbal nature of the task. Verbality yields a context that has a different effect on interpretation than visual context. The effect of context on this sample, therefore, is confirmed at the level of trend- interactions (Fig. 7). It manifests 21 itself strongly rather in a broader sense, in a comparison of the modalities of verbality (as linguistic context) vs. nonverbality (visual stimuli). For more exact results we need further research on particular patterns of interpretation in the different forms of pragmatic constructions with special attention to contextual cues and their impact on processing. 7.3. Contextual effects in irony processing – optimally relevant and targeting intentions As for our irony trials the effect of context is much clearer: the surface cue added significantly facilitated understanding and successful decoding, and the differences disappeared when the surface cue making the mental state of the speaker explicit was added. This optimal context, in a cognitively and interpersonally motivated pragmatic meaning construction is crucial, and although there is no lengthy explanation, its optimal nature makes it so strikingly efficient in deciphering intended, implicit message as the ironic meaning. This surface cue making the mental state of the speaker explicit directly effects the intentional nature of the utterance, thus facilitates the identification of speaker’s intent and speaker’s meaning. It is this targeted, efficient nature that made the role of context so clear-cut in the case of the irony trials. The linguistic cue or surface cue provided optimally relevant information, namely, the mental state of the speaker, which is a key component in the construction of pragmatic meaning; making it explicit also made it directly accessible in interpretation. Hence the name, surface cue. This way it diminished the need for lengthy inferential activities and thus, mentalization enabling these. Surface cue therefore, provides information that is more relevant than average linguistic information of the same length, as contextual elements and cues can be more or less relevant (depending on the topic, the speakers, their intentions, their interests and the social purposes served by their utterances). In harmony with the principle of Relevance (Sperber—Wilson 1986) and optimal relevance (Sperber—Wilson 1986, Giora 2001) our results confirm that Relevance is a general cognitive guideline for our inferential activities in identifying intended meaning. In the tasks on metaphor and humor the context we provided was not directly targeting mental states thus making implicit inferentially uncovered details explicit, therefore, context there facilitated interpretation to a smaller degree, only to the level of tendency-like interaction. But in the irony trials, the context provided specifically intention-based information, which is the cornerstone of identifying intended meaning and taking perspectives. This latter case made contextual effects strikingly significant, as they directly targeted mental states and socialcognitive aspects of meaning. The purely linguistic context provided in the metaphor, humor condition, which basically ensure only a longer story as context, was not directly relevant to 22 identification of speaker’s meaning, hence, the numerical data on contextual effects is also less significant and effects are less intensive. Our results confirm the facilitating effect of contextual cues in the interpretation of pragmatic constructions to a smaller or a greater extent. 8. Conclusion Our results indicate that (1.) within trials (Metaphor, Humor) the facilitating effect of ‘average’ context is not significant (Figures 6 and 7: decontextual vs. contextual), (2.) on this sample in Humor tasks the effect of context stays at the level of interactional trends and is more pronounced in between modalities (verbal- vs. nonverbal-, as context) (Figure 7: nonverbal vs verbal). Here the principle of Cognitive Congruence is key in understanding the interaction of contextual levels and cues. In the case of contextual (surface-) cues relevant for mental state of the speaker, hence optimally relevant for the deciphering of the intention of the speaker, produces significant differences in interpretation, and thus provides significance of contextual effects. Our findings confirm results in developmental research that surface cues help in the recognition of communicative intent and contribute to the successful resolution of the intended meaning at hand. Contextual effects thus may vary according to what they target, and the successful deciphering of implicit meaning is even more strikingly influenced by contextual cues than in literal language. 8.1. Contributions to the cognitive pragmatic investigation of context The present investigation aimed to contribute to the resolution of a number of long-standing debates in cognitive linguistics and developmental psycholinguistics. With its novel approaches in research methodology, it aims to clarify boundaries of metaphor, irony and humor, testing humor in spontaneous settings, in both verbal and nonverbal modalities, not only through prefabricated jokes commonly administered. It also examines all the four maxims concurrently, shedding light on the developmental implications of their acquisition and understanding. It relies on neuroimaging based experiments in trying to explain the relationship of mentalizing skills and pragmatic competence in general, and of difficulties of irony processing in schizophrenia stemming from a mentalization deficit in the disorder in particular (Varga et al. 2013, 2014, Schnell et al. 2016, Varga et al. 2016). It also entails a potential for creating a therapeutic training material in neurocognitive disorders with the aim of reintegrating the 23 patient, given that it shows how surface cues can enhance success of comprehension in instances of pragmatic constructions based on mentalization and idiomaticity. The testing material provides a comprehensive pragmatic test (to be standardized) which can be used as a clinical measure of general pragmatic competence in several neurocognitive impairments and neuropsychiatric disorders (Down Syndrome (DS), Williams Syndrome (WS), Specific Language Impairment (SLI), Autism Spectrum Disorder (ASD) currently not available, functioning as a differential diagnostic measure in research. 24 References Abu-Akel, A., – Shamay-Tsoory, S. G. 2011, Neuroanatomical and neurochemical bases of theory of mind. Neuropsychologia, 49(11), 2971–2984. Baron-Cohen S. – Leslie A. M. – Frith, U. 1985, Does the autistic child have a theory of mind? Cognition, 21(1), 37-46. Baron-Cohen, S. 1995, Mindblindness. An essay on autism and theory of mind. Cambridge: Mass., MIT Press. Colston, H. – Gibbs, R. 2002, Are irony and metaphor understood differently? Metaphor and symbol, 17, 57-80. Csibra, G. 2010, Recognizing communicative intentions in infancy. Mind & Language, 25, 141-168. de Villiers, J. 2007, The interface of language and Theory of Mind. Lingua. 117(11) 1858– 1878. Gibbs, R. – Nayak, N. – Cutting, C. 1989, How to kick the bucket and not decompose: Analyzability and idiom processing. Journal of memory and language, 28, 576-593. Gibbs, R. 1994, The Poetics of Mind. Cambridge: Cambridge University Press. Gibbs, R. – Gregory, A. B. – Colston, H. 2014, Where is the humor in verbal irony? Humor, 27(4), 575-595. Giora, R. 2001, Irony and its discontent. In Psychology of language: In honour of Elrud Ibsch. Eds. G. Steen and D. Schram. Amsterdam: John Benjamins, 165-184. Győri M. et al. 2002, Naiv tudatelmélet és nyelvi pragmatika magasan funkcionáló autizmusban: reprezentációs zavar, performanciakorlát, vagy kompenzáció? [Naive theory of mind and linguistic pragmatics in high-functioning autism: representational disorder, 25 performance- limitation or compensation?] In Racsmány, M. – Kéri, Sz. (eds.): Architektura és patológia a megismerésben [Architecture and pathology in cognition] Budapest: BIP, 11-39. Katsos, N. – Bishop, D.V.M. 2011, Pragmatic tolerance: Implications for the acquisition of informativeness and implicature. Cognition 120. 67–81. Lakoff, G. – Johnson, M. 1980, Metaphors we Live by. Chicago. The University of Chicago Press. Martin, R. A. 2007, The psychology of humor: An integrative approach. Burlington, MA: Elsevier Academic Press. Musolino, J. – Lidz, J. 2006, Why children aren’t universally successful with quantification. Linguistics 44–4 (2006), 817–852. Pexman, P. – Glenwright, M. – Krol, A. – James, T. 2005, An acquired taste: Children’s perceptions of humor and teasing in verbal irony. Discourse processes 40.(3). 259-288. Schnell, Zs. 2007, Metaphor Processing and the Acquisition of Idioms – A mentalistic model. Acta Linguistica 54 (2007) 1, 73-104. Schnell, Zs. 2010, When the Penny Drops – in the Cognitive Dimension. In: Litovkina, A. – Barta, P. – Daczi, M. (eds.) Linguistic Shots at Humour. Krakow: Tertium. 189-209. Schnell, Zs. 2012, The Development of Humour Competence of Hungarian children - A Cognitive Approach. In: T. Litovkina A. – Szőllősy J. – Medgyes P. – Chłopicki W. (eds): Hungarian Humour. Humor and Culture 3. Krakow: Tertium Society for the Promotion of Language Studies, 2012. 235-253. Schnell, Zs. 2015, Social cognitive and pragmatic aspects of language acquisition from a developmental perspective. http://pea.lib.pte.hu/handle/pea/14467 Schnell, Zs. 2019, Fuzzy Boundaries in Interpretation: Empirical Study on the Common Traits and Differences of Irony and Metaphor as the Hypothesized Infringements of the Maxim of 26 Quality. In: Furkó et al. (2019). Fuzzy Boundaries in Discourse Studies. Theoretical, Methodological, and Lexico-Grammatical Fuzziness. Palgrave-Macmillan. Cham, Switzerland. Part of the Postdisciplinary Studies in Discourse book series (PSDS). Print ISBN 978-3-03027572-3. Online ISBN 978-3-030-27573-0. E-book Chapter 8: https://link.springer.com/chapter/10.1007%2F978-3-030-27573-0_8 Schnell, Zs., - Dr. Herold, Róber – Prof Dr. DSc. Tényi, Tamás – Dr. Varga Eszter 2019, Fuzzy Boundaries and Fuzzy Minds: Interpretation Strategies and Discourse Processing in Schizophrenia. In: Furkó et al. 2019, Fuzzy Boundaries in Discourse Studies. Theoretical, Methodological, and Lexico-Grammatical Fuzziness. Palgrave-Macmillan. Cham, Switzerland. Part of the Postdisciplinary Studies in Discourse book series (PSDS). Print ISBN 978-3-03027572-3. Online ISBN 978-3-030-27573-0 E-book chapter 9: https://link.springer.com/chapter/10.1007%2F978-3-030-27573-0_9 Schnell, Zs. – E. Varga 2012, Humour, Irony and Social Cognition. In: T. Litovkina A –, Szőllősy J. –Medgyes P. – Chłopicki W. (eds): Hungarian Humour. Humor and Culture 3. Krakow: Tertium Society for the Promotion of Language Studies, 2012. 253-271. Schnell, Zs. – Varga, E. – Tényi, T. – Simon, M. –Hajnal, A. – Járai, R. - Herold, R. 2016, Neuropragmatics and irony processing in schizophrenia – the role of the pragmatic module. Journal of Pragmatics 92. 74-99. Séra, L. 1983, A nevetés és a humor pszichológiája. [The psychology of laughter and humour.] Budapest: Hungarian Academy of Sciences. Sperber, D. – Wilson, D. 1986, Relevance: Communication and Cognition. Cambridge USA: Harvard University Press. Sperber, D. – Wilson, D. 2002, Pragmatics, modularity and mindreading. Mind & Language 17: 3-23. Special issue on pragmatics and cognitive science. Sullivan, K. et al. 1994, Preschoolers can attribute second-order beliefs. Developmental Psychology, Vol. 30(3), 395-402. 27 Varga, E – Simon, M. – Tényi, T. – Schnell, Zs. – Hajnal, A. – Orsi, G. – Dóczi, T. – Komoly, S. – Janszky, J. – Füredi, R. – Hamvas, E. – Fekete, S. – Herold, R. 2013, Irony comprehension and context processing in schizophrenia during remission – A functional MRI study. Brain and Language 126. 231-242. Varga, E. – Zs. Schnell – T. Tényi – N. Németh – M. Simon – A. Hajnal – R. Horváth – E. Hamvas – S. Fekete – R. Herold – R. Járai 2014, Compensatory effect of general cognitive skills on non-literal language processing in schizophrenia. Journal of Neurolinguistics. Vol. 29. 1–16. Varga, E. – Herold, R. – Schnell, Zs. – Horváth, R. – Simon, M. – Hajnal, A. – Tényi, T. 2016, The processing of humour by individuals suffering from schizophrenia. European Journal of Humour Research. Vol 4. No. 1. http://www.europeanjournalofhumour.org/index.php/ejhr/article/view/138/pdf Wilson, D – Sperber, D. 1992, On verbal irony. Lingua 87. 53-76. Wilson, D. 2009, Irony and metarepresentation. UCL Working Papers in Linguistics 21: 183-226. Wilson, D. – Carston, R. 2006, Metaphor, relevance and the “emergent property” issue. Mind and Language 21, 404–433. Wilson, D. 2013, Irony comprehension: a developmental perspective. Journal of Pragmatics 59. 40-56. Wimmer, H. – Perner, J. 1983, Beliefs about beliefs: Representation and constraining function of wrong beliefs in young children's understanding of deception. Cognition, 13, 103-128. Zigler, E. – Levine, J. – Gould, L. 1966, Cognitive processes in the development of children’s appreciation of humour. Child Development 37(3) 507-518. 28 29