BREAKING THE RULES: PRAGMATIC CONNOTATIONS OF (UN)MARKED USES OF FINAL NEOCLASSICAL COMBINING FORMS

This paper analyzes the way speakers perceive the transgression of the rules of word formation, particularly regarding neoclassical final combining forms, which are elements with a marked use and whose morphological combining standards are quite restrictive. This transgression consists in combining a native Catalan word and a final combining form, the result of which has been assigned an expressive value. In order to test these perceptions, we developed a questionnaire and administered it to 152 speakers of Catalan from different backgrounds. These speakers were asked to evaluate 20 neologisms documented in the Catalan press regarding their context of use and their connotations. The results show that speakers perceive that those forms that “break” the expected rules are more likely to appear in an informal or colloquial context, and they generally assign them an ironic or pejorative value. Finally, this study aims to present compounds as a promising topic of interest within morphopragmatics.


INTRODUCTION
The combination of morphology and pragmatics, two fields that seemed unlikely to combine, was made real with the arrival of morphopragmatics, the foundations of which took form with the research carried out by Dressler and Merlini Barbaresi ("Elements"; Morphopragmatics; "Morphopragmatics") and Dressler and Kiefer. According to these authors, pragmatics influences the production and reception of complex words, but, besides the study of diminutives and augmentatives, research that followed this path has not been in any way systematic (Meibauer, "Wordformation and Contextualism"). This paper aims to add to the existing research on morphopragmatics by analyzing the formation of neoclassical compounds in Catalan which combine an initial element that is not an initial combining form (henceforth, ICF) and a neoclassical final combining form (henceforth, FCF); therefore, they break the rules of "standard" neoclassical compounding, according to which we always find two neoclassical elements (an ICF and an FCF). We believe that this transgression in the form of the compound creates unexpected words to which speakers associate, in most cases, specific pragmatic effects. Through the analysis of the answers obtained in a questionnaire that was answered by 152 participants, we observe the context (formal, informal, or neutral) in which speakers think that these formations will appear, as well as the effects that speakers associate to these words (section 6). These evaluations guide a discussion (section 7) about the nature of the combining forms and help us to prove that the pragmatic effect is created by the very transgression of the formation process itself.

MORPHOPRAGMATICS AND COMPOUNDS
According to the authors that have worked in morphopragmatics, this subdiscipline is devoted to the general meanings of the morphological rules; more specifically, it is interested in the morphological rules that contain a pragmatic variable that cannot be omitted in the description of the meaning of a form (Dressler and Merlini Barbaresi, "Morphopragmatics" 1). As expressed by Kiefer (272), "morphology is relevant pragmatically in so far as word structure can be taken as an indication of the speech situation and/or of the speech event." Among the wide range of morphological phenomena, morphopramatics was conceived to deal, most of all, with derivation and flexion. The research by Dressler and Merlini Barbaresi focuses on the pragmatic value that can be attributed to the use of augmentatives and diminutives in speech acts or interaction strategies (Merlini Barbaresi), known as evaluative morphology. Initially, this line of research was applied to Italian and German, but it was later expanded to include other languages such as Spanish (Calvo Pérez; Cantero; Martín Zorraquino), Portuguese (Gonçalves), and French (Koehl and Lignon). According to this research, Merlini Barbaresi considers that among the pragmatic effects that result from evaluative morphology there is emotion and tenderness, irony, euphemisms, fury, sarcasm, and contempt.
Beyond augmentatives and diminutives, Dressler and Barbaresi ("Elements"; "Morphopragmatics") mention explicitly that compounding is rarely relevant in morphopragmatics. Kiefer, however, does include compounds as likely receivers of pragmatic activity, but he is only concerned with compounds created ad hoc in poetry, advertisements, and newspaper headlines. Therefore, this paper aims to show that the scope of morphopragmatics could be expanded to include this kind of forms, that is, neoclassical compounds created with a native word and a neoclassical FCF, although, as we see in section 5 their behavior differs from them. We explain how violating the rule of word formation for neoclassical compounds, which usually follow an ICF + FCF structure, is precisely what creates a pragmatic effect in receivers, who assign specific nuances to this kind of compounds when they encounter them. This connotation, therefore, does not emerge from the context in which this word appears (a feature that Merlini Barbaresi finds devoid of any pragmatic relevance), but to the formation process that this word undertook and that is included in the morphological rule.

CREATIVITY AND EXPRESSIVE MORPHOLOGY
Creativity plays a crucial role in the conception behind these compounds. In morphology, Bauer (63) states that creativity is "the native speaker's ability to extend the language system in a motivated but unpredictable (non-rule-governed) way." This absence of rules is what leads to the transgression in the use of specific combining forms in compounds, as we show in section 4, and to the fact that these compounds, in turn, are perceived as new by the speakers (Lüdeling and Evert).
Morphological creativity, understood as the "conscious coinage of a new word on an unproductive pattern" (Mühleisen 96), opposes morphological productivity, "the unintentional coinage of a number of formations." Zwicky and Pullum (335) propose a similar dichotomy, although they restrict it to derivative morphology: plain/core morphology and expressive morphology. Despite this limitation, we believe that two of the features that differentiate these two groups could also apply to compounds: • Pragmatic effect. This feature is particularly relevant for our study, and it refers to "expressive, playful, poetic, or simply ostentatious" effects that are not likely to appear in core morphology. It must be said, however, that the example offered by the authors comes from advertising, and they deem it "whimsical." • Promiscuity with regard to input basehood. The field of application of expressive morphology is much wider than that of core morphology, and it can apply to inflected forms, compounds and other elements. Although the authors are specifically referring to the category of the bases, in our case this promiscuity can be found in the nature of the initial element in the compound.
On the other hand, Bagasheva and Stamenov, and Winter-Froemel, talk about ludicity and they associate traces of intentionality to it and the pursuit of a special effect, of entertaining, surprising, etc. Finally, Baeskow, in her analysis of the lexical proprieties of non-English morphemes, addresses morphological creativity through compounds such as hamburgerology (example cited in Bauer), jassophile or queenomania, which imitate the formation of neoclassical compounds like vulcanology. This author explains that when the native base (non-classical) of the hybrid formation is simple, it can combine directly with a neoclassical root of the first order, following the structure: "freely occurring constituent + FCF" (in the previous examples it was -log, -phile and -mania). And this structure is precisely what gives them "their markedness and somewhat humorous connotation" (Baeskow 85).
Along the same lines, although in this case for phrasal compounds in German, Meibauer ("How Marginal") considers that the forms created ad hoc transmit a certain expressivity and account for almost all the features that Levinson (137) lists for marked forms: "morphologically more complex and less lexicalized, more prolix or periphrastic, less frequent or usual, and less neutral in register." Meibauer ("How Marginal") adds that these compounds are understandable and witty, even when speakers encounter them out of context. Therefore, this phenomenon has been approached from different perspectives and authors have qualified these compounds as expressive, stylistic, playful, etc. These adjectives refer to the result, but they do not address the process the compounds underwent in their creation; we believe that the approach presented here can provide some answers about a process that is not isolated, but replicable in all the words that result from the transgression of the rule of neoclassical compound formation.

USING (FINAL) COMBINING FORMS TO ADD PRAGMATIC EFFECTS
The status of combining forms 1 such as -itis, -mania or -oide, but also hidro-, aero-or pseudo-, has been discussed before in many languages (Scalise; Fuhrhop; ten Hacken) with an underlying assumption: word formation with neoclassical compounds differs from word formation with native compounds (see Lüdeling, Schmid and Kiokpasoglou for a summary), or, at least, it should be treated differently. Neoclassical compounding is based on the assumption that neoclassical combining forms can "only" combine among them, which justifies the differential treatment that they are usually given. 2 Although neoclassical compounding is especially frequent in the creation of specialized words, this is not an exclusive context, since Romance languages (and even more so Indo-European languages) incorporate neoclassical combining forms and their patterns of combination to their word formation processes. Generally, homogeneous neoclassical formations are given preference over hybrid formations, 3 and, within the latter, those that contain combining forms from the same language should also come first (Termcat), as seen in the following examples from Catalan: Greek Nevertheless, word formation that combines neoclassical and native elements goes far beyond specialized languages (terminology), and its combinatory possibilities reach much higher in languages like Catalan, and other Romance and Germanic languages: Greek + Catalan / English: enoturisme 'enotourism, wine tourism' / microsurgery Latin + Catalan / English: triatleta 'triathlete' / supermarket Catalan / English + Latin: lliberticida 'freedom + -cide' / meritocrat Catalan / English + Greek: aromateràpia 'aromatherapy' / Americanitis Cabré (100, our translation) explains that this kind of formations "are not neoclassical, but they have been formed on purpose by imitating the processes of creation of neoclassical compounding with an expressive purpose, and not a referential one." This explains why we can find so many neoclassical compounds in non-specialized contexts (Bernal et al.), which in turn places neoclassical compounding as a fairly productive word formation process in current Catalan (Observatori de Neologia 113). In fact, this situation may have contributed to a change in the view that speakers have on these combining forms, which they now conceive as prefixes and suffixes instead (for example, in the case of -itis and -oide) (Observatori de Neologia 127). Following this idea, some authors (Amiot and Dal; Bernal, "Nuevos prefijos") have wondered whether the morphological status of these combining forms is the same in these cases than in strict neoclassical compounds, especially in some cases like -itis (Feine; Lüdeling, Schmid and Kiokpasoglou; Bernal, "Valor neològic." These authors maintain that there are clear differences between words like tonsil·litis 'tonsillitis' vs. mundialitis 'World Cup-itis' or dipsomania 'dipsomania' vs. dinomania 'dinomania' (Bernal, "Valor neològic"). Other authors, like Díaz-Negrillo, have studied the type of native bases that can combine with neoclassical forms and the motivation behind the appearance of the linking vowel in English. However, beyond the morphological dimension, these authors also associate an undeniable expressive dimension to these new formations, which is determined by the deviation from the combination rules of their components, and that grants them pejorative or ironic hues.
Concerning Catalan language specifically, neoclassical combining forms are a frequent resource when it comes to creating new words: for the period 2007-2009, Cañete, Fernández, and Janer document 1,143 new different words (from a lexicographical perspective 4 ) created using combining forms. This amount corresponds to 11.63% of the total of neologisms documented in that period, which puts neoclassical compounding in second place when it comes to frequency, behind only prefixation and omitting loanwords from other languages. Out of this percentage, only 5% are homogeneous compounds, like biogènia 'biogeny' or megapolis 'megapolis', while the rest combine a neoclassical element and a native one. For these hybrid creations, Bernal et al. establish the following characteristics: • Hybrid compounds change the rules of formation of standard neological compounds, since they admit a word from the general language that can be a noun (or an adjective, though it is not as frequent) -common or proper -, a phrase, an acronym of a loanword from another language. • They are expressive neologisms, because they add a positive or negative value to the word (Observatori de Neologia 126), which can even be ironic. In the case of prefixes and suffixes with these values, Provencio Garrigós (247-248) points out that hearers or readers must deduce the evaluations hidden behind the neologisms. • In order to achieve this expressivity, the neoclassical compound must appear in final position, which usually crates the need for a linking vowel between the two components: an -o-for Greek forms (saltòdrom 'jumpodrome') and an -i-for Latin forms (austericidi 'austericide'). • Speakers play constantly with these neoclassical combining forms, with the goal of catching the attention of the receiver (Guilbert). Speakers can play with them because they know what they mean and how to use them, and this knowledge is precisely what allows them to break the corresponding word-formation rule. As for the readers or hearers, they are able to interpret them and understand the motivation behind them thanks to that very same knowledge that allows them to create them (Bernal and Sinner; Winter-Froemel). • These formations are usually short-lived, and they will seldom leave the context and register in which they were created. However, even if most cases presented here are nonce-words, some of them have attained certain frequency (sentimentaloide lit. 'given to easy and shallow emotion', meritocràcia 'meritocracy') or that have been documented in other languages (in Spanish, aplausómetro / aplaudímetro, in English clapometer¸ clapometer and applause meter, etc.).
The capacity that speakers possess to create hybrid neoclassical compounds that break the prototypical formation has led some authors to talk about a trivialization of the formation process (Adelstein; Guerrero Ramos and Pérez Lagos; Ginebra, Llagostera and Rull). However, beyond the changes in meaning, there is a change in register and a humorous tone that cannot be found in traditional neoclassical compounding, as we see below in section 6.

METHODOLOGY
Within the framework of emotive communication, Caffi and Janney (358) present two possible analytical approaches: as a process, where it is interactive and the emotive signals are negotiated between participants, and as a product, in which the emotive profile appears as a quality and can be associated to stylistics. In this paper we focus on the latter, since the emotive signals are specific words that result from the process of compounding. Since they are small units (compared to discourse, sentences, or texts), a static perspective is more pertinent.
Concerning the specific morphopragmatic phenomenon under discussion, we identified it following the steps proposed by Dressler and Merlini Barbaresi ("Morphopragmatics"): • Make sure that it concerns a morphological rule (productive and regular), instead of it being an extragrammatical operation or a phenomenon like suppletion. • Carry out a semantic analysis of the morphological rule in order to establish what can be inferred from the meaning of its forms without involving any pragmatic factors. In our case, each compounding form retains its original meaning (although some cases like -itis deserve special attention), and so do the native initial forms. Therefore, the attribution of a specific (and diverse) pragmatic value finds its origin in the union of these two elements to create a new word. • Analyze the different elements from the point of view of pragmatics in order to prove that generalizations do not originate individually or syntactically, and that they go beyond the semantic properties of the units under study.
This last step is the focus of our analysis, which is carried out from a micro perspective: we analyze specific ICFs and FCFs in order to exemplify how this morphopragmatic rule works, and we assess them through the evaluation and the values that speakers attribute to them. The chosen words are the following: 5 In general, the neoclassical combining forms presented above retain their original etymological meaning (for example, -meter means 'measure [instrument]', -leg means 'specialist', etc.). However, in some cases this original meaning has been expanded metaphorically, as in -drom, which went from 'racecourse' to simply 'place or premises', or -man, which went from 'maniac' to 'lover, fan' (Bernal et al.). These twenty neologisms were chosen according to the following criteria: • They are all words that have not been included in the dictionary, but which appear with relative frequency in the media. 7 • Most of them combine a native initial word with a neoclassical FCF. However, we also included five "proper" neoclassical compounds that follow the word formation according to which both forms must be neoclassical: cleptocràcia 'kleptocracy', equinoteràpia 'hippotherapy', eurofòbia 'Europhobia', fructícola 'fruit + -cole', geopatia 'geopathy' and rinoconjuntivitis 'rhinoconjunctivitis'. These six words work as distractors, because, considering they follow the expected rules of formation, they should be considered neutral or unmarked by the speakers. • There are five cases in which the final form is repeated, a decision which also answers to the need to include distractors and obtain more complex answers: -cràcia (cleptocràcia 'kleptocracy', comentocràcia 'commentocracy'), -itis (dissenyitis 'design + -itis', rinoconjuntivitis 'rhinoconjunctivitis'), -leg/-logia (rumorologia 'rumorology', totòleg 'everything + -logue', zarzuelòleg 'Zarzuela + -logue'), -man/-mania (balletòman 'balletomane', obamamania 'Obamamania') and -metre (aplaudímetre 'clapometer', criticòmetre 'criticism + -meter').
These twenty words were included in a questionnaire in which collaborators had to evaluate the neologisms, which were presented in a real context extracted from a newspaper or a magazine in Catalan (and presented as such): La llegenda excèntrica … no va parar mai de créixer, gràcies a la rumorologia, al valor immens de la seva literatura i a l'ús pervers que feren alguns pertorbats -l'assassí de Lennon, l'assassí fallit de Reagan -dels seus llibres. (Ara, February 1, 2014) També estan afectats per la crisi econòmica global i pateixen greus problemes de corrupció i cleptocràcia governamental. (Avui, February 18, 2011) The informants had to answer two questions. First, they had to choose the context in which they thought this neologism could appear: formal or neutral or informal and colloquial (mutually exclusive). Then, they were asked to assign a pragmatic effect to the word (non-exclusive choices in order to give them freedom): humorous and ironic, neutral, affective, derogatory, or other (which allowed informants to add any additional effects they thought of, or, as it happened in some cases, to justify their answers if they felt the need to do so). In this case, the decision to have an open question followed the results of previous studies (Grandi; Provencio Garrigós; Bernal, "Valor neològic"), which found that frequently more than one value can be attributed to one form. As established in Llopart-Saumell, perception studies, which value linguistic intuition, are relevant in order to obtain data about vocabulary processing and classification tasks. They are complemented by more experimental approaches that have been gaining attention, such as the use of eyetracking (Llopart-Saumell et al.) or Natural Language Processing tools (Lejeune and Cartier; Torres Rivera and Torres-Moreno).
In order to reach a wide variety of informants, the questionnaire was shared online. A total of 152 people completed forms, 76.32% women and 23.68% men.
Concerning age, 25% of the informants were 16-24 years old, 23.7% were 25-34 years old, 26.3% were 35-44 years old, 20.4% were 45-59 years old, and, finally, 4.6% were over 60 years old. Out of all of them, 150 were native Catalan speakers and the other 2 were German speakers with an advanced knowledge of the Catalan language, analogous to that of a native speaker, so their answers were also included. The results of the questionnaire is presented in section 6.

DATA AND RESULTS
As explained in section 5, the questionnaire was divided into two parts. In the first one, informants needed to choose the context in which the words were more likely to appear, and in the second one they had to attribute to them an effect (which could be neutral). The results of the first question are shown in Table 1.
In seven cases, there is almost a unanimous answer in the classification of the contexts. On the one hand, informants consider that the words equinoteràpia 'hippotherapy' (96.1%), fructícola 'fruit + -cole' (97.4%), geopatia 'geopathy' (97.4%) and rinoconjuntivitis 'rhinoconjunctivitis' (98%), which are well-formed neoclassical compounds, would typically appear in formal contexts. On the other, informants determined that dissenyitis 'design + -itis' (94.7%), follòdrom 'to fuck + -drome' (90.8%) i xupòpter 'to suck + -ptherus' (94.1%) clearly belonged to informal contexts. In these three cases, the transgression in quite obvious: dissenyitis does not refer to an illness or disease of any kind, follòdrom has a very vulgar initial form (the verb follar 'to fuck'), and xupòpter has nothing to do with flying insects and takes a loanword from Spanish as the initial form (chupar 'to suck').
There is a second group of words for which the answers were quite uniform. The forms cleptocràcia 'kleptocracy' (79.6%), eurofòbia 'Europhobia' (82.9%) and telegènia 'TV + -geny' (73%) are attributed a high likelihood of appearing in formal contexts, while criticòmetre 'criticism + -meter' (84.9%), sabatòfon 'shoe phone' (85.5%) and totòleg 'everything + -logue' (77.6%) are clearly assigned to informal contexts. This distinction between formal and informal contexts corresponds, once again, to them being well-formed words with two neoclassical forms, in the first case, and words that result from breaking the formation rule in order to denominate entities that do not actually exist in real life: an instrument to measure criticism, a shoe phone, 8 or a specialist in everything.
Moving from the clear-cut distinction, we find five words that were classified with less certainty: a 60-70%. On the one hand, aplaudímetre 'clapometer' (61.2%) and rumorologia 'rumorology' (68.4%) were considered typical of formal contexts. This is probably due to the fact that their use is stabilized (Metcalf) in Catalan, and they are actually included in some secondary dictionaries and resources (aplaudímetre is listed in ésAdir, the style guide of the Catalan media we found to be more likely to appear in informal contexts, probably because the transgression presented by the initial form is more obvious: in two cases, it is a proper noun (Obama and Zarzuela), and in the other the expected linking vowel -o-does not appear (serièfil). Finally, balletòman 'balletomane' and comentocràcia 'commentocracy' are the words that presented more doubts for the informants, who considered them more likely to appear in informal context with only a 51.3%. Both words can be formally assimilated to neoclassical compounds due to the appearance of the linking vowel, but the initial combining form has a clear common origin: ballet and comentari 'commentary'.
Concerning the second question of the questionnaire, in which speakers were asked to attribute pragmatic effects to the compounds, the answers were very diverse, since, besides the available options and the possibility to combine more than one of them, there was an additional space to add more options or to give an explanation. In order to simplify the presentation of the results, Table 2 only includes the results for the most frequent categories (those which were selected by more than 10% of the informants), 9 and the more salient results are highlighted in bold.
In Table 2 we find that there is more than an 85% of agreement on four words that are perceived as neutral: fructícola 'fruit + -cole' (93.4%), equinoteràpia 'hippotherapy' (88%), geopatia 'geopathy' (90.8%) and rinoconjuntivitis 'rhinoconjunctivitis' (92.7%). These are precisely four of the five proper neoclassical compounds that were introduced as distractors, which the informants considered likely to appear in formal context in the first part of the questionnaire. The fifth distractor, cleptocràcia, was, however, only considered neutral by 21.2% of the informants, even though it follows the morphological rule of formation. The reason behind this circumstance can be attributed to the concept it refers to, which is considered to be derogatory or clearly negative by most informants (42.1%). There is no government that could be based on theft, and this word is actually only used by the government's political opposition party, and it never refers to the speaker that pronounces it.
On the other hand, the results for rule-breaking compounds are more varied, with the exception of criticòmetre 'criticism + -cracy' and follòdrom 'to fuck + -drome', which are clearly perceived as humorous and ironic. In the rest of the cases, the informants seem to struggle when it comes to deciding which effect they attribute to them (although it is never neutral), and that is why they usually choose more than one option. In comentocràcia 'commentocracy', totòleg 'everything + -logue', xupòpter 'to suck + -ptherus' and zarzuelòleg 'Zarzuela + -logue', speakers opt for a derogatory effect over the others (39.5%, 40.1%, 38.8%, and 30.3%, respectively), while in dissenyitis 'design + -itis' the dominant effect is the humorous and ironic (44.7%).
Finally, it must be noted that the affective effect, which usually attributed to diminutives and augmentatives, does not seem to be especially relevant for hybrid neoclassical compounds: it only receives a notable percentage in four words (and  it never surpasses the 35%). For two of these words, it becomes the first effect, balletòman 'balletomane' (34.6%) and serièfil 'TV show + -phile' (31.58%); for obamamania 'Obamamania' (23%), it is the second effect chosen by the informants after humor and ironic, and for telegènia 'TV + -geny' (11.84%) it is the third most frequent effect. All these words denominate positive concepts, such as hobbies or virtues (in the case of telegènia), and we think these are precisely what makes informants choose the affective effect, more so than the ICFs themselves.

DISCUSSION
After presenting the results separately, we now match the answers to the two questions posed in the questionnaire in order to explore the informants' evaluations more closely. In most cases, the classification was expectable, and the choices they made concerning the context and the effect are coherent, but there were also some surprises that must be addressed.
First of all, we will talk about five proper neoclassical compounds (out of a total of six) that were hidden as distractors: equinoteràpia 'hyppotherapy', eurofòbia 'Europhobia', fructícola 'fruit + -cole', geopatia 'geopathy', and rinoconjuntivitis 'rhinoconjunctivitis'. For all of them informants agree in considering that they belong in formal contexts (> 88%), and they attribute an undisputable neutral effect to four of them (> 88%). These results show that informants have no doubts when it comes to the orientation of proper neoclassical forms. The fifth neoclassical compound, eurofòbia 'Europhobia', is mostly considered a neutral word (54.61%), but 26.32% of informants see it as a derogatory word. This is probably due to the Greek form -fòbia 'phobia', which means 'fear, aversion' and is usually associated to negative things or features; this negativity might have conferred a pragmatic value on the Greek combining form.
Secondly, there is a second group of neologisms that informants classified as pertaining to a formal context, but with a less clear distinction (between 61% and 79%): cleptocràcia 'kleptocracy' (the sixth distractor), telegènia 'TV + -geny', rumorologia 'rumorology' and aplaudímetre 'clapometer'. In the last three words, the neutral effect is the dominant one, although the percentages they obtained were inferior to those of the proper neoclassical compounds: between 44% and 52%. This unmarked value might be a result of the connotation that speakers assign to -gènia '-geny', -logia '-logy', and -metre '-meter', and educated language: identifying these FCFs as pertaining to a higher register of language somehow softens the surprise of having an initial native form. However, the case of cleptocràcia is different. As seen above in section 6, it refers to a concept that is clearly negative: therefore, even if speakers identify clepto-'klepto-' and -cràcia '-cracy' as neoclassical elements, the meaning prevails, and that is why they consider it has a derogatory effect (42.1%) instead of neutral (21.1%). All these cases, however, share the fact that they are relatively frequent words with a tendency to find stabilization in use, which seems to resonate with the speakers, who place them mostly in formal contexts.
In the middle of the list there are two words for which the informants could not decide on a context of use: balletòman 'balletomane' and comentocràcia 'commentocracy'. Nevertheless, the effects attributed to them are quite different. On the one hand, balletòman is one of the cases where affective effect is relevant (34.6%), followed by neutrality (26.1%) and irony (15.7%). The affective value might be related to the fact that it seems hard to value someone who loves something as artistic as ballet, as ironic or derogatory; however, the informants wanted to reflect that its meaning went beyond what is purely referential. On the other hand, neutrality does not appear in comentocràcia, which informants identify as derogatory (39.5%), humorous and ironic (30.9%), or both (18.4%) connotations. Just as happened for cleptocràcia, the concept behind comentocràcia is negative and critical, but at the same time it is clearly a playful creation that refers to a strange (but truthful) idea: the power that comments (made by people on social media) have over other people and the decisions they make. The difference between considering these two words as formal or informal, therefore, depends on the nature of the first element: while clepto-is a neoclassical form, coment-is clearly a native Catalan word, a fact that, in the eyes of speakers, lowers its formality, even if they both retain the pragmatic effects.
Within the group of words that have been considered informal, there is a first group where we find obamamania 'Obamamania', serièfil 'TV show + -phile' i zarzuelòleg 'Zarzuela + -logue' (with between 64 and 68%). In these three concepts the neutral effect is minor (between 13% and 25%), behind the other pragmatic effects. Two of them contain proper nouns (obamamania 'Obamamania', which is considered humorous and ironic, and zarzuelòleg 'Zarzuela + -logue', which is seen as derogatory), and two of them gather a lot of affectivity (obamamania 'Obamamania' and serièfil 'TV show + -phile') thanks to the fact that the concepts they name are clearly positive. In general, the forms -mania, -mane and -philia, -phile (excluding any sexual of pathological issues) are linked to positive hobbies, even if in some cases they can get a bit out of hand.
Finally, there are six neologisms that have been classified as clearly pertaining to an informal context (>75%), to which informants have associated clear pragmatic effects: totòleg 'everything + -logue', criticòmetre 'criticism + -meter', sabatòfon 'shoe phone', follòdrom 'to fuck + -drome', xupòpter 'to suck + -ptherus' and dissenyitis 'design + -itis'. Three of them were seen as humorous and ironic, possibly due to how impossible it would be for these things to actually exist and for the humorous intent that triggered their creation: criticòmetre 'criticism + -meter', sabatòfon 'shoe phone' and follòdrom 'to fuck + -drome'. As for totòleg 'everything + -logue', xupòpter 'to suck + -ptherus' and dissenyitis 'design + -itis', they combine a humorous and ironic and a derogative effect, as was seen before: they are also playful, but they designate negative or socially unacceptable concepts, such as a know-it-all, someone who takes advantage of others, and an unhealthy obsession for anything made by a renowned designer. All these cases share a native initial word that speakers recognize immediately.
The results presented in sections 6 and 7 reflect that speakers are clearly able to distinguish proper neoclassical compounds from those compounds that break the rules of word formation: homogeneous compounds are considered suitable for formal contexts, while compounds with a native initial form are perceived as informal. Nevertheless, there are some cases, such as rumorologia 'rumorology' or cleptocràcia 'kleptocracy', where the stability and the frequency of use of the words has reduced the transgression of their form in the eyes of the informants.
Additionally, the degree of metaphorization of the meaning of the compounds also seems to play a role in their evaluation: in the examples with -metre, which always retains the meaning of '[instrument used to] measure', the perception of the informants changes according to the plausibility of those instruments that actually exist. If they are real or possible instruments (aplaudímetre 'clapometer'), which measure real things that can be measured in real units (meters, grams, etc.) -just like in proper neoclassical compounds -they are unmarked, while those that designate hypothetical instruments that measure abstract notions are considered more expressive (criticòmetre 'criticism + -meter').
To sum up, the humorous and ironic effect is present and dominates in almost all informal (i.e., transgressor) compounds: in some cases as the main pragmatic effect, in others as the second one, right behind the derogatory effect that arises in negative concepts. Therefore, informants have univocally assigned a pragmatic value to those forms that express meanings that go beyond the sum of their components, as a consequence of their breaking the rules of neoclassical compound formation.

CONCLUSIONS
This study shows that breaking the rules of word formation by creating marked forms that combine a native word in first position and a neoclassical FCs results in pragmatics effects that are associated with the compound. These effects have an impact on the register, because these forms are much more likely to appear in informal contexts, and they carry additional nuances that lead speakers to interpret them as ironic, derogative, or both. According to our analysis, the strength of these effects is blurred if the word becomes stabilized in use: the more a neologism is used, the less transgressive it becomes. In this sense, neologisms that initially presented a surprising composition, such as aplaudímetre 'clapometer' and rumorologia 'rumorology', have lost their strangeness in the eyes of speakers due to their continued use and are now considered neutral; on the contrary, follòdrom 'to fuck + -drome' and sabatòfon 'shoe phone' have not had such extensive use and are assigned humorous and ironic values. Despite our results, we must admit that the design of the questionnaire had some limitations, especially when it comes to the somewhat scattered answers to the second question. However, we believe that the data we collected is significant and makes for a consistent analysis, given the high coincidence we found in the answers of the informants, and that our study can open a new path for a new approach on compounds from a morphopragmatic approach.
In his study, Plag makes the following remark about compounds: [A]lthough compounding is the most productive type of word-formation process in English, it is perhaps also the most controversial one in terms of its linguistic analysis and I must forewarn readers seeking clear answers to their questions that compounding is a field of study where intricate problems abound, numerous issues remain unsolved, and convincing solutions are generally not so easy to find. (132) 4. A word is considered a neologism if it does not appear in the exclusion corpus, which in this case are the normative or more representative dictionaries of a language. 5. For all the Catalan neologisms, an equivalent word in English or a literal equivalent of its parts is offered, and then each part of the word is explained separately. 6. The case of equinoteràpia deserves special attention, since it is a hybrid neoclassical compound which combines a Latin and a Greek form: the Greek combining form equino-(in English, echino-) actually means 'hedgehog', as in equinoderm 'echinoderm'; therefore, the appropriate compound to talk about the therapy method where horses are used to improve the condition of people with disabilities, conduct disorders or social rehabilitation would be hipo-(in English, hippo-) 'horse'. Instead, the form equino-is formed from the Latin adjective equí -ina 'equine', to which a linking vowel is added in order to combine it with the Greek form -teràpia 'therapy'. 7. All the neologisms were extracted from the database of the  Get Smart, broadcast between 1965 and1970]. If shoe phones actually existed, I would have considered it a neutral word" [our translation]. 9. The percentages were calculated individually for each word based on the total of answers obtained, since this number could vary due to the possibility of choosing more than one option.