Inherent linguistic impoliteness: The case of insultive YOU þ NP in Dutch, English and Polish

This article conducts a corpus analysis of insults in the form YOU þ NP (e.g. you (stupid) idiot ), an impoliteness formula, in Dutch, English and Polish. It argues that impoliteness can be inherently associated with linguistic structures, a claim which contradicts the widely held view in current (im)politeness research that impoliteness, and indeed politeness, is primarily determined by context. However, whilst we show that our insultive form is strongly conventionalised in similar ways across languages, it is never completely conventional. We suggest that the generally high level of conventionalisation found for this form is a result of the addressee evaluation inherent in the structure, as well as the pragmatic explicitness, and thus directness, of referring to the target with a second person pronoun. The form was found to be most conventionalised for impoliteness in Polish, something which is probably attributable to the decline of the vocative case in that language. The article also considers the nature of exceptions, i.e. cases which ﬁ t the form but were not impolite.


Introduction
Politeness studies, dating from the early days of classics such as Lakoff (1973), Leech (1983), Brown and Levinson (1987), and later impoliteness studies, dating from works such as Culpeper (1996), have experienced a significant shift.What has generally emerged is the view that (im)politeness is not inherent in linguistic form or structure but rather a contextual judgement.This shift has been galvanised by work such as Eelen (2001) and particularly the advent of the "discursive" approach (e.g.Locher, 2006) and the post-structuralist approach (e.g.Mills, 2003).Although support for this view is not uniform, detractors including, for example, Terkourafi (e.g. 2001) for politeness and Culpeper (e.g. 2011) for impoliteness, it has developed into something approaching a new orthodoxy.While it would be absurd to argue that (im)politeness is only a matter of meanings inherent in linguistic forms (e.g.impoliteness can be implied), our position is that linguistic form plays a substantial role and one that has been somewhat neglected in the literature.A key aim of this article, then, is to bring form back to the fore, by showing not just that there exist structures in language that are conventionally associated with impoliteness but also that impoliteness can be strongly conventionalised in structures across languages.
Our particular focus is the form YOUþNP, as in you bastard.As we will elaborate, there are multiple reasons for this choice.Firstly, items such as you bastard constitute a type of impoliteness formula (Culpeper, 2011), namely, an insult.Importantly, insults involve negatively evaluative linguistic content (e.g.bastard) and an orientation, usually explicit, to a target (e.g.you).With respect to other impoliteness formulae, such as threats, dismissals, curses and so forth, they are the most frequent type (Culpeper et al., 2017: 15).Secondly, YOUþNP is not restricted to English.This form also occurs with impoliteness functions and effects in other languages, including Dutch and Polish, as explored in the present article.Thirdly, YOUþNP is computersearchable in a corpus.Of course, whether a linguistic item is (im)polite is not a quality that a computer can confirm.We therefore extracted randomised samples from our corpora, and then manually examined those samples.
Our key research questions are: (i) is YOUþNP a construction conventionalised for impoliteness?; (ii) does YOUþNP vary, not least with respect to conventionalisation, across the languages under consideration, and, if so, how?; (iii) what are the components of this linguistic form that affect its function?;(iv) why is YOUþNP conventionalised for impoliteness across languages?A particular point of interest is the fact that YOUþNP need not always have negatively evaluative linguistic content, as complimentary you star makes clear.
In Section 2, we review relevant literature, covering the debates on whether (im)politeness is inherent in linguistic form, the notion of impoliteness formulae, insults and the form YOUþNP. Section 3 describes our data and more particularly our procedures for extracting instances of YOUþNP for the three languages under consideration and identifying relevant impolite cases.Section 4 provides the results of our analyses for each language and, in Section 5, finally, we discuss these findings.Note that this paper, dealing with insults in naturally occurring data, inevitably contains examples that some may find offensive.

Literature review
2.1.Debates about the (non-)inherence of (im)politeness One of the most enduring debates in studies of (im)politeness concerns the question whether it is inherent in language or not.These were aired in Culpeper (2011: 117e126).We more briefly rehearse them here and, moreover, update them to include what has happened over the last dozen or so years.A necessary preliminary, although one neglected by much scholarship, must be to define what is meant by (im)politeness being 'inherent in language'.We take a traditional view of inherent meaning viewing it as formal semantic meaning (encoded, literal or explicit) which is (i) more a matter of truth conditions than felicity conditions, (ii) more conventional than non-conventional and (iii) more non-contextual (and thus non-relative) than contextual (cf.Grice, 1989: 25, 87).
Perhaps the earliest statement articulating the view that (im)politeness is not inherent in form is Fraser and Nolan's (1981: 96): "[N]o sentence is inherently polite or impolite."The not-inherent movement was given a significant boost by the development of the "discursive turn" in (im)politeness studies, involving, amongst other things, a focus on participants' situated and dynamic evaluations of politeness.Locher (2006: 250e251), for instance, writes: "What is perceived to be (im) polite will thus ultimately rely on interactants' assessments of social norms of appropriateness that have been previously acquired in the speech events in question ... As a result, we claim e with many others e that no utterance is inherently polite."(See also e.g.Eelen, 2001;Locher and Watts, 2008.)In addition, the not-inherent movement was championed by those pursuing a post-structuralist or post-modern approach (e.g.Mills, 2003), which favours cultural and individual relativism and disfavours universalising generalisations, as might be captured by (semi-)conventional (im)politeness forms of language.More recently, this approach seems to have merged with the discursive approach generally (e.g.Van der Bom and Mills, 2015).However, most articulations of the discursive approach do allow a half-open door on the idea that some words or expressions have some inherent (im)politeness meaning.Watts (e.g. 2003) and Locher (e.g. 2004), for example, embrace the notion of a (cognitive) 'frame' and use it to account for how people make judgements about things they have never before experienced: they draw on frame-based knowledge about such things.This mixed approach has led to the charge that discursive approaches are incoherent (Haugh, 2007).
Importantly, as Culpeper (2011: 120e121) points out, no mainstream (im)politeness theorist argues that (im)politeness is simply a matter of what is inherent in linguistic expressions.If (im)politeness meanings were entirely fixed in particular forms, one could not account for how: (i) (im)politeness can be achieved in the absence of those forms (e.g. in the right (British) context, I'm thirsty could be perceived as a polite request for a cup of tea); (ii) linguistic forms can have different degrees of (im)politeness (or none at all) when perceived by different people, in different situations or cultures and so forth; and (iii) polite expressions might be used for sarcasm.However, a strict position that (im)politeness is not at all inherent in particular words and expressions is not plausible.People can communicate (im)polite meanings with relative ease, which suggests they are quite stable.Clark (1996) argues that conventions, including those flowing from words and structures, enable participants to coordinate their thoughts and actions.Surely, this also applies to (im)politeness meanings.Perhaps the most compelling evidence for the strict position's implausibility is that people are able to judge the differing degrees of (im) politeness of particular words and expressions out of context (e.g.Jain, 2022: 389).In reality, then, (im)politeness has to be inherent to a degree.How might such issues be accounted for theoretically?
A starting point might be to look back at the classic works on politeness.Here, far from the fully inherent position being dominant, we actually find dualism.Leech (1983: 83e84, 102) distinguishes between "absolute politeness" and "relative politeness".These terms were updated in Leech (2014: 88) to "pragmalinguistic politeness", i.e. "degrees of politeness in terms of the lexigrammatical form and semantic interpretation of the utterance", and "sociopragmatic politeness, i.e. "politeness relative to norms in a given society, group, or situation".Early dualist positions were not confined to Leech (1983) (e.g.Craig et al., 1986).However, whilst being sufficiently broad to accommodate the varying politeness phenomena, they did not offer a full theoretical account, especially not of the pragmalinguistic side of things.For example, Leech's (2014: 74e76) comments on how pragmalinguistic features evolve, alluding to pragmaticalisation, a process by which pragmatic meanings become conventionalised to some degree in specific linguistic forms, but does so only briefly.A fuller account emerged in the work of Terkourafi (e.g. 2001Terkourafi (e.g. , 2002;;2005a;2005b), which is the focus of the next section.

(Im)politeness formulae
Terkourafi (e.g.2001) proposes a frame-based approach to politeness, arguing that we should analyse concrete linguistic realisations and particular contexts of use, which together co-constitute frames.While Terkourafi was not the first to think in terms of the notion of frames in pragmatics (e.g.Aijmer, 1996) or in terms of sociopragmatic routines (e.g.Coulmas, 1981), she was the first to connect the two in the pursuit of politeness and anchor everything in pragmatic theory.In her view, it is "the regular co-occurrence of particular types of context and particular linguistic expressions as the unchallenged realisations of particular acts that create the perception of politeness" (Terkourafi, 2005a: 248).That such 'politeness formulae' are both regular and unchallenged accounts for the observation that politeness often passes unnoticed (e.g.Kasper, 1990: 193).In Britain, for instance, the please's and thank you's that accompany requests will not be the talking point of the day.
Politeness formulae are conventionalised to some degree.Terkourafi (2005b: 213) defines conventionalisation thus: a relationship holding between utterances and context, which is a correlate of the (statistical) frequency with which an expression is used in one's experience of a particular context.Conventionalisation is thus a matter of degree, and may well vary in different speakers, as well as for the same speaker over time.This does not preclude the possibility that a particular expression may be conventionalised in a particular context for virtually all speakers of a particular language, thereby appearing to be a convention of that language.
In other words, candidates for politeness formulae become conventionalised to some degree for a particular context of use.Note here that there is no claim that such items have completely stable polite meanings and that they are polite in all contexts of use.These conventionalised meanings are positioned midway between semantics and pragmatics, between fully conventionalised and non-conventionalised meanings (Levinson, 2000: 25).They do not amount to what Terkourafi (2005b: 211) calls "sentence meaning", encoded in the form and indefeasible, but they are also not mere "utterance-token meaning", derived in nonce contexts through particularised conversational inferences.Rather, they constitute "utterance-type meaning", which is presumed/preferred for the form but is still defeasible.Terkourafi (2005a: 251, original emphasis) supplies an elegant Neo-Gricean account of the pragmatic inferencing of such conventionalised polite expressions: Politeness is achieved on the basis of a generalised implicature when an expression x is uttered in a context with which based on the addressee's previous experience of similar contextsexpression x regularly co-occurs.In this case, rather than engaging in full-blown inferencing about the speaker's intention, the addressee draws on that previous experience (represented holistically as a frame) to derive the proposition that "in offering an expression x the speaker is being polite" as a generalised implicature of the speaker's utterance.On the basis of this generalised implicature, the addressee may then come to hold the further belief that the speaker is polite.
Terkourafi's work became a starting point for Culpeper's (2011) impoliteness formulae.Unlike with politeness, evidence of impoliteness is in the fact that it is typically challenged by counter impoliteness and that it attracts meta-pragmatic comments (e.g.'that was so rude') and/or displays of humiliation, hurt, or anger.Generally, formulae can be studied through empirical methods, such as corpus linguistics.Culpeper (2011: 113e154) collected candidates for English impoliteness formulae from discourses in which impoliteness is central (e.g.exploitative TV shows, graffiti) and impoliteness event report forms completed by 100 British undergraduates.All candidates were then examined in the two-billion-word Oxford English Corpus, ensuring that more than 50% of the instances of each type occurred in contexts which could be interpreted as involving impoliteness (e.g.angry reactions, counter impoliteness).Table 1 displays some of the resulting impoliteness formulae.Importantly, the same types of impoliteness formulae have been confirmed to exist in other languages and cultures (see Kleinke and B€ os, 2015 for German data, Lai, 2019 for Chinese data, Andersson, 2022 for Swedish data and Tsoumou, 2023 for multilingual data from Congo-Brazzaville).

Insults
Of all impoliteness formulae, insults e which often take the form central to the present article, i.e.YOUþNP e are by far the most frequent (e.g.Autor et al. Year).In an attempt to define genuine insults for British English, Jucker and Taavitsainen (2000: 73) suggest that they have three features at their core: First, a predication about the target (or about some part of his/her social identity, e.g.his/her profession).That is to say that the speaker utters something about the target or uses words to characterize him/her, or uses an epithet to address him/her.Second, this predication is perceived as inappropriate and demeaning by the target.And third, the target experiences this predication as a face-threatening speaker intention, that is to say he or she believes that the speaker made the predication with the intention to hurt or demean him or her.
The second and third features are widely shared by many impoliteness definitions (cf.Culpeper, 2011: 19e20); these features specify what it is to be insulting, which overlaps with what it is to be impolite.It is the first feature where we find something specific to insults and which is most relevant to the linguistic concerns of this article.In fact, as we see it, there are two components to this feature: (i) "a predication about the target (or about some part of his/her social identity, e.g.his/her profession)" and (ii) an "address" to a target.
As for the predication, the idea that insults must say something negative about the target, usually an aspect of their identity, is reflected in the fact that the literature is well-stocked with work exploring such content (e.g.Bright, 1979;Nkara, 1996).Caution is needed in our understanding here.First, you pig, for instance, could be taken as a bantering, sympathising expression when said by a friend following a large meal, but it would most likely be understood entirely differently if said by a stranger.In other words, the effects of insults can be cancelled by context (cf.O'Driscoll, 2020: 73 on taboo predication).Second, there is actually nothing literally negative about the word pig, denoting a type of animal.At issue in you pig are the connotations that the word has acquired in our culture.As Allan (2007Allan ( : 1049) ) argues, they "rely on … encyclopaedic knowledge, experiences, beliefs, and prejudices about the" denotatum, and, crucially, such pragmatic effects "will vary from community to community".Thus, what counts as negative differs between languages/cultures.Third, not all insults are as conventional as you pig.You porcine personage has the same basic meaning but breaks the conventional pattern in selecting more formal, obscure alliterating lexis.More standardised insults are likely associated with a high degree of anger on the part of the insulter, who relieves their agitation by turning to the most readily available word at their disposal (Martínez and Yus, 2000).Innovative insults, as Labov's (1972) early work in this area made clear, have long been associated with friendly interactions and banter.This is not an absolute distinction, of course: insults can also be creatively modified or created anew to intensify their offensiveness (see Culpeper, 2011: 139e151, 239e244).As for the component of addressing a target, finally, the first three insult structures in Table 1, including the YOUþNP example you fucking moron, achieve it through an explicit form of the second person pronoun.The fourth relies on a strong implication generated in context that a target is being addressed.

YOUþNP
Our focus on YOUþNP is motivated by a number of considerations, which we discuss in the four following sections.Section 2.4.5, finally, introduces the question central to the present article: is YOUþNP a construction conventionalised for impoliteness?

Languages
YOUþNP is not limited to English.Similar patterns exist in other Germanic languages, such as Swedish in (1a).Like in English, someone can be addressed not just as 'idiot' in Swedish but also as 'you idiot'.Unlike in English, the second person singular in (1a) occurs in its possessive form din 'your' (Julien, 2016: 97).Dutch, the first author's native language, has YOUþNP too.The second person singular takes its full subject form jij 'you' here, as in (1b) (cf.unstressed je).1 (1) a. din idiot 'you idiot' (Julien, 2016: 102) b. jij idioot 'you idiot' The pattern is not restricted to the Germanic group, though.According to Corver (2008), for instance, it is also found in Brazilian Portuguese, where the second person is possessive again, as masculine singular seu 'your' in (2a) shows.Polish, the second author's native language, has YOUþNP as well.A man can be called an idiot just by employing the vocative of masculine idiota, which is idioto (cf.feminine nominative idiotka, vocative idiotko).However, one may also add second person singular vocative/nominative ty 'you', like in (2b).
(2) a. seu idiota 'you idiot' (Corver, 2008: 52) b. ty idioto 'you idiot' In other words, YOUþNP yields an opportunity to examine whether a pattern that has been put forward as a possible impoliteness formula in one language functions in the same way in other languages.Given the requirements for analysing insults proposed in Section 2.3, we will investigate Dutch, English and Polish here.

Searchability
For a corpus-based study, it is of course desirable that the pattern at issue can easily be located in the data.YOUþNP satisfies this requirement (see Section 3.2) e unlike, for instance, less linguistically specified impoliteness formulae, like condescensions and threats.

Frequency
Of the formulae in Table 1, YOUþNP e referred to above as 'personalised negative vocatives' e is one of the most common ones across the languages under investigation.Exploratory evidence for this claim from the English enTenTen18 and Polish plTenTen19 corpora is given in Table 2 (see Section 3.1 for more information about these corpora).It compares the frequencies of the YOUþNP strings you bastard/bitch and the corresponding 'personalised negative assertion' strings you are/'re (such) a bastard/bitch, as well as their respective Polish equivalents ty draniu/suko and jeste s draniem/suka ˛.2 YOUþNP is more frequent than personalised negative assertions in Dutch too.For this language, we can also compare the YOUþNP string in (3a) with the string in (3b).This example translates literally as 'slut that/who you are' and instantiates a distinct construction described by Corver (2016: 385) as an exclamative relative vocative (ERV) and as expressing "a (negative) evaluation of the addressee by the speaker".The frequencies of all these strings in the nlTenTen14 corpus, for a selection of nouns, are presented in Table 3.
(3) a. jij slet 'you slut' b. slet dat/die je bent 'you're such a slut' The numbers are low and the differences minimal, but the pattern proves fairly consistent across nouns.

Constructional status
YOUþNP has been argued to be an actual construction.We take this term to refer to a "conventionalized" pairing "of form and function" (Goldberg, 2006: 3), combining a distinctive set of formal features and constraints with a specific meaning.As Corver (2008: 52e55) explains for Dutch, YOUþNP is distinct from appositive patterns of pronoun and noun like wij/jullie fietsers 'we/you cyclists' in (4a), despite the superficial similarities.Unlike (4b), which is an example of an address, appositives are 2 To be clear, we did not check whether all the hits in Tables 2 and 3, for that matter e actually function as insults in context (see Section 3.3).Assuming, with caution, that YOUþNP, personalised negative assertations and the construction in (3b) do not vary drastically in their ratios of impolite uses, we still believe that these numbers e especially with nouns meaning 'bastard', 'bitch' and the like e can serve to support our claim that YOUþNP is the more frequent insult formula in the languages under investigation.part of the syntax of the clause and can only be plural, as (4c) makes clear.The same difference exists in English, as evidenced by the translations.
'You, idiot, are spoilt.' In Polish too, the nominative plural wy rowerzy sci 'you cyclists' would be acceptable in (4a) and the nominative singular ty rowerzysta 'you cyclist' unacceptable in (4c).3Unlike its counterparts in Dutch and English, the latter cannot serve to address someone either; for this function, the vocative instead of nominative case is required in Polish (e.g.rowerzysto 'cyclist').The distinction only exists in the singular, however, because of syncretism between the vocative and the nominative in the plural.Note also that, in all three languages, singular subject 'you' could e in principle, but only in a fairly marked way e be followed by a direct address, but that pattern differs from both YOUþNP and appositives in requiring some kind of prosodic break, as indicated by the commas in (5).Corver (2008: 47e50) and Potts and Roeper (2006: 187e188) also mention semantic restrictions specific to YOUþNP.For Dutch, it is said that only degree nouns can appear in the construction on their own but other nouns can be made to fit by adding a gradable adjective.We can illustrate these claims with Polish, which exhibits similar behaviour.'Idiot' in (6a) is perfectly acceptable but 'linguist' in (6b) is somewhat awkward unless, as in (6c), an evaluative adjective like 'stupid' is inserted.In the same vein, it is argued for English that nouns can only be modified by expressive items.We can again use Polish to exemplify the point: (6c) with głupi 'stupid' is unproblematic, (6d) with wysoki 'tall' is slightly peculiar.It remains to be seen, however, how categorical these constraints truly are.
'You tall idiot.' The above arguments motivate Corver (2008) and Potts and Roeper (2006) to propose special (generative) syntactic structures for YOUþNP.The details are beyond the scope of the present article, though.What is crucial for us here is that the formula identified by Culpeper (2011) clearly has a distinct form.
Moreover, this form is said to pair, in a conventionalised way, with a particular function (see Section 2.2).For English, Potts and Roeper (2006: 184) characterise it, generally, as expressive or conveying "emotive force" and, more precisely, as "selfdisapprobation".The description of, say, you idiot as some kind of attack suggests a close link with impoliteness.Yet, it fails to capture our intuition that speakers normally direct YOUþNP not at themselves but at their addressee(s).Corver (2008: 47) appears to agree, for Dutch, and calls the construction an "evaluative vocative" (cf.Julien, 2016: 90).He indicates, however, that the value judgment need not be negative, like in the cases discussed so far, and provides the affectionate example in (7) (see Jain, 2022: 371 for a similar point on English YOUþNP).
(7) jij duifje van me you little dove of mine' (Corver, 2008: 47) Evidence that YOUþNP is evaluative as a construction and this meaning is not simply due to the NPs occurring in it comes in two related forms.As shown in (6b), a typically non-evaluative noun such as 'linguist', if not modified by an adjective like 'stupid', seems rather incompatible with YOUþNP's overall function, at first glance.At the same time, if the construction nonetheless featured such a noun, it would coerce an evaluative reading. 4As Davies (1986) points out, boy and man can be used on their own to address someone in a respectively authoritative and informal but non-evaluative way.YOUþNP, by contrast, "is sufficient to impose a pejorative or complimentary tone; so that you boy, if used at all, would probably be understood to express the speaker's disgust at the addressee's behaving [sic] childishly, while you man might serve to express admiration of the addressee's manly qualities" (Davies, 1986: 99).She does add that this evaluative coercion is only unavoidable in the singular.Plural you boys may just be used to address and name a group of people.It remains to be seen, however, whether this identifying use is indeed restricted to plural 'you'.

Impoliteness
YOUþNP can thus be regarded as a construction, in the sense introduced in Section 2.4.4,dedicated to addressee evaluation.This fact does not, however, equate to the claim that it is a conventionalised impoliteness formula.As pointed out by Jain (2022: 366), "expletive NPs … are natural expressive labels" and thus, obviously, highly compatible with YOUþNP.So are positively evaluative NPs, though, as evidenced by (7). 5 Culpeper (2011) is nevertheless not the only one who seems to attribute a sense of impoliteness to the construction.Finkbeiner et al. (2016: 4), for instance, refer to (German) cases like 'you human being' as "pejoration based on syntactic constructions" and both Julien (2016: 91) and Jain (2022: 371) suggest (for Scandinavian languages and English respectively) that negative readings of YOUþnon-evaluative NP are preferred.Given examples like (7), it is so far unclear why the construction tends to be associated with impoliteness.
We would contend that the answer lies in usage.One of the tenets of the usage-based framework is that "instances of use impact the cognitive representation of language" (Bybee, 2010: 14).Moreover, when repeatedly experiencing instances that have certain features in common, language users may generalise over them (e.g.Langacker, 2000: 93).Such generalisations or schemas can vary in their level of abstraction and would be part of a network of more/less abstract ones, "the relationships between [which] may change" (Traugott and Trousdale, 2013: 16) over time.The implications for YOUþNP e and us e are as follows: if we can establish that, in actual usage, the construction (most) frequently serves impolite purposes, it is not unreasonable to assume that language users generalise over such instances and there exists a schema e alongside a more abstract evaluative one e where the form YOUþNP is associated with the 'function' of impoliteness.Its relative prominence might then explain, for example, the hypothesised negative interpretation bias for non-evaluative NPs.This line of argumentation is compatible with earlier work on (im)politeness formulae (see Section 2.2).YOUþNP would constitute an impoliteness formula displaying conventionalisation of the kind proposed by Terkourafi (2005b: 231) and Culpeper (2011), i.e. as utterance-type meaning that is presumed/preferred but still defeasible.Crucially, conventionalisation is a matter of degree in this approach and YOUþNP would not need to possess an invariably impolite meaning or be impolite in every context.If YOUþNP is indeed seen to be employed mostly for impoliteness, the final question that still needs to be answered, in Section 5, is why the usage of this evaluative construction is skewed this way.

Corpora
For our study, we used the multilingual TenTen corpus family.It contains large bodies of texts, with billions of words (see Table 4), that "can be regarded as comparable corpora" as the same "technology specialized in collecting only linguistically valuable web content"6 is employed to construct a corpus for every language considered in our study.To our knowledge, there exist no other readily accessible corpora with comparable data for Dutch, English, as well as Polish.Moreover, TenTen's sheer size and variety of texts e including, for instance, not only newspaper articles and Wikipedia pages but also discussion forums and online fiction, which approximate speech in at least some respects e would ensure a sufficient amount of relevant information.TenTen also has the advantage of being tagged with parts of speech, which allowed for the relatively straightforward retrieval of 'you' followed by (an adjective and) a noun.Relying on web-crawled data comes with certain drawbacks too, however (e.g.Jakubícek et al., 2013).It is, for example, hard to control for language variety (e.g.Northern/Southern Dutch, (non-)native English) or time.For many online sources, it is simply impossible to know when they were first produced.Corpora like the TenTen ones are thus not unlikely to contain much less recent material as well.We will return to these issues in the interpretation of our results.

Data selection
The Dutch nlTenTen14, English enTenTen18, and Polish plTenTen19 components were explored with Sketch Engine's concordance tool and corpus query language. 7The string in (8), for English, is indicative of our searches.We essentially looked for all capitalized and non-capitalized second person forms able to occur in the construction (e.g.not your) that are immediately followed by a noun or a pairing of first adjective and then noun.A few comments are in order, however.First, for the second person forms, we restricted ourselves to the standardized language, to ensure comparability: Dutch second person singular jij, plural jullie and polite u; English you; Polish second person singular ty and plural wy (and not, for instance, Brabantic Dutch gij or English second person plural youse).Second, the inclusion of a single adjective as an optional element allowed us to extract cases like 'you idiot' and 'you little idiot' simultaneously.Multiple adjectives are possible but, to keep the data manageable, we decided not to search for cases like 'you stupid little idiot'.Culpeper's (2011: 239) corpus results for you bastard with zero to four words between the second person and the insult suggest in any case that a single optional adjective would catch approximately 80% of relevant instances of YOUþNP.Third, we also limited the number of non-target hits (e.g.you financial advice in if someone is licensed to give you financial advice) by requiring the presence of punctuation marks at the beginning and the end of the string.This constraint is motivated by the parenthetical/autonomous nature of the construction.
The third column of Table 4 gives the number of hits for each language.The Dutch and Polish data were downloaded entirely with the maximum amount of context permitted (i.e. 100 characters to the string's left and right) and randomized in Excel.As the English data exceeded Sketch Engine's download limit, a random sample of 10,000 hits was extracted.
We then manually selected the first 200 target cases for every language, forming the dataset to be analysed in detail (see Section 3.3).This process meant removing obvious non-target hits like (9).We also disregarded any hits where the string fulfils a syntactic role in the clause.Jullie Joden 'you Jews' in (10), for instance, is not of interest here, as it functions as the subject of the sentence.(9) ...to recline and relax in one of our massage basins, pedi-spa or massage rooms allowing you the chance to spend some real "you time" while you get pampered.(enTenTen18-10867806) (10) ...bonkige en witte Nederlanders die ongegeneerd beginnen te melden ... dat "jullie Joden" dat moeten aanvaarden.(nlTenTen14-5337355) '… chunky and white Dutch people who begin to mention unabashedly … that you Jews have to accept that.' We did, however, include what may be described as metalinguistic instances.In (11), for example, the string itself is talked about and serves to make a point about language.A first reason for retaining such uses is that they are far from uncommon e making up 6.5%, 2% and 11% of our Dutch, English, and Polish data respectively.This fact perhaps reflects the idea that impoliteness "is more noticed and discussed than politeness" (Culpeper, 2011: 131).A second, related reason is that such instances can offer insight into how YOUþNP is actually perceived by speakers (cf.Culpeper, 2009: 66e67).
(11) Wiem, _ ze tym komentarzem troszeczke ˛odbiegne ˛od tematu, ale je_ zeli zwr ocicie sie ˛do mieszka nca e powiedzmy e Afryki, "Ty murzynie", to be ˛dzie to dla niego najwie ˛ksza obelga.(plTenTen19-1206406) 'I know that with this comment I will deviate a bit from the topic, but if you address an inhabitant of e say e Africa, with "you negro", it will be the biggest insult to him.' For the sake of completeness, the fourth column in Table 4 presents, for each language, the number of hits that we had to go through to reach the goal of 200 target cases.

Data analysis
Each hit in the resulting data sets was coded for the following features.
(i) the singular (e.g.Dutch jij) versus plural (e.g.Polish wy) number of YOUþNP (for number-neutral English you and Dutch u, the subsequent noun's singularity or plurality was the deciding factor; for Dutch u, we also made a note of its polite nature); (ii) the absence versus presence of an adjective in the NP; (iii) the specific noun and, potentially, adjective occurring in the NP; (iv) the metalinguistic versus 'normal' use of YOUþNP; (v) its impolite or non-impolite function in context, which will be explained in more detail in the next paragraphs.
The Dutch data was primarily analysed by the first author and the Polish data by the second author.They collaborated on the English data.The coding, for feature (v) especially, underwent multiple rounds of reanalysis, after regular meetings in which the first two authors discussed their work with each other and with the third author. 8he analysis of corpus data as impolite or not is notoriously difficult.It would be erroneous, for instance, to label everything that looks like an insult, conventionally meant to cause offense, as impolite (see also Section 2.3 on banter).For this reason, amongst others that we discuss below, we considered the co-text of every hit carefully, in line with  Culpeper (2011: 11e12), and coded as impolite those instances where there are reasonable indications that YOUþNP is meant and/or taken to have negative emotional consequences.The evidence was varied in nature.We found, for example, explicit evaluations of the construction as impolite, particularly in metalinguistic uses like (11).A non-metalinguistic case is (12).
(12) usmiechnol sie do mnie szyderczo gdy widzial ze policja mnie powstrzymuje, krzyknołem "ty pedale!" a on do mnie "ty heteryku!":/ od kiedy heteryk to cos zlego?:| ale skoro on mnie tka obraza to uwaza ze to cos zlego (plTenTen19-390136) 'he smirked at me when he saw that the police were stopping me, I shouted "you faggot!" and he said to me "you heterosexual!":/ since when is being straight something wrong?:| but since he is insulting me in this way, he must think it's wrong' The addressee's verbal or described non-verbal reactions were often revealing, too.Unlike politeness, which typically goes unnoticed, as noted above, impoliteness tends to be countered in some way (e.g.Culpeper and Tantucci, 2021).In (13a), for example, the second speaker's retort suggests that they take offense at the first one's words, disputing them defensively (cf.Culpeper et al., 's 2003Culpeper et al., 's : 1563 typology of impoliteness response options).In (13b), the narrator clearly describes the addressee's non-verbal reaction as one of fury.Their use of unprovoked, which regularly modifies nouns like attack, is another indication that you stupid bitch is impolite here.Narratives also frequently afforded insight into the speaker's aggressive intent and/or psychological state.In ( 14), for instance, the female driver is described as incensed by the boy's conduct.( 14) De bestuurster draaide zich naar de jongen en keek hem aan met ogen vol vuur."Jij snotjong, dit is de allerlaatste keer dat ik je uit die teringkroeg vandaan moet halen.Uit mijn ogen voor ik..." (nlTenTen14-3841903) 'The female driver turned toward the boy and looked at him with eyes full of fire."You young snot, this is the very last time that I have to get you out of that damn pub.Get out of my sight before I…"' Another co-textual sign of impoliteness was the acts co-occurring with YOUþNP.The aggressive dismissal (another impoliteness formula; see Section 2.2) at the end of ( 14) is a case in point.
Co-text was especially relevant for (adjectives and) nouns not readily interpreted as insults.Consider ( 15).It is only evident from the presentation of the soldiers' words as taunting and the reference to their typical behaviour that you heroes is sarcastic here.
We also encountered cases of YOUþNP that exhibit no signs of impoliteness and, unlike banter, cannot be regarded as prima facie insults that are "neutralised or even made positive" in "highly restricted … contexts" (Culpeper 2016: 435).An example is ( 16).The speaker here is actually encouraging the addressee and complimenting this Podlachian on their valour in the face of defeat.Other types of YOUþNP cases that were not analysed as impolite are discussed in the following section.

Results
In order to give the reader a good idea of our data for each language, we list in tables the particular nouns and adjectives appearing in YOUþNP (only the recurrent ones and the first ten unique ones, in alphabetical order, will be given here; see the Supplementary Material for the others), as they proved revealing even without context.The frequency of each noun or adjective will be included too, as will the number of times that it occurs in an impolite instance of the construction in the corpus.Other tables will offer summary information about the overall proportions of cases with(out) an adjective, of impolite ones and of singular/plural ones.

Dutch
Table 5 presents an overview of all recurrent nouns and adjectives and the first ten unique ones (alphabetically) found in our random sample of 200 instances of YOUþNP (see also the Supplementary Material).As indicated in summary Table 6, most hits are singular (85.00%).This figure includes the two (impolite) attestations of formal second person u, that is, u huichelaar 'you hypocrite' and u bastaard 'you bastard'.Table 6 also points out that Dutch YOUþNP appears with an adjective in roughly half of the cases (51.00%).
Given the many nouns like viezerik 'dirty person' and leugenaar 'liar' and the many adjectives like vuil 'dirty' and lelijk 'ugly', it should come as no surprise that YOUþNP is primarily used for impolite purposes in Dutch (68.50%).Consider (17) for some more co-textualised examples and in particular the evaluative noun in (17a), the evaluative adjective in (17b), the combination of the two in (17c) and the non-evaluative NP in (17d). 9The speaker's accusations of selfishness and hypocrite in the latter example signal that 'modern-day Christians' is meant as an insult.YOUþNP can be said to facilitate this intention, encouraging through coercion an evaluative reading of the NP that is likely to be negative because of the primacy of impolite uses.Cases like (17d) account for only a small number of the hits, though.
(17) a. Mijn oude ik had meteen de mond open gedaan en was meteen in de aanval gegaan.Jij mafkees, wie denk je wel dat je bent.(nlTenTen14-284060) 'The old me would immediately have opened their mouth immediately and would immediately have gone on the attack.You nutcase, who do you think you are.' b.
'"Who has eaten one of the cleaners??" Hesitantly, one of the other four raises his hand… "You stupid dickhead!Idiot!!" screams the leader at him.' d.O! jullie hedendaagse christenen, jullie spijzen ons met broodkruimels en hosties, en laten ons die niet samen met jullie dopen in de schotels … Dat is een goede manier om met de christelijke gemeenschap de draak te steken en haar arme christenbroeders de schijn van het liefdesmaal, in plaats van der werkelijkheid daarvan te geven.(nlTenTen14-642548) Note also in Fig. 1 that singular and plural cases are similar in their proportions of impoliteness (67.65% and 73.33% respectively) but that there is a significant difference between NPs without an adjective (76.53%) and NPs with an adjective (60.78%; c2 ¼ 5.74, p < 0.05).
A closer look at the data suggests that most nouns in the impolite cases may e as in (17c) e but need not e as in (17a) e be modified to convey an evaluation of the addressee.By contrast, many of the nouns in the non-impolite instances require an adjective to express a non-negative evaluation (e.g.jij lieve lezer 'you sweet reader', jij bijzondere vrouw 'you remarkable woman').Without the adjective, an impolite interpretation of such cases would seem more likely.
The figure of 68.50% is arguably not the entire story for Dutch YOUþNP and impoliteness, however.There are two relatively common settings in the corpus where ostensible insults do not appear to be meant or taken as an offense.The first one is the well-known phenomenon of banter between interlocutors who are close to each other.The speaker and their cat in (18a) can serve as an example.The second one involves 'dirty talk' (erotic talk) and is linked to nouns like slet 'slut' and voyeurtje 'little voyeur' and adjectives like ondeugend 'naughty' and geil 'horny'.This type of setting is illustrated in (18b).
(18) a. [Ik] ging naar binnen terwijl ze zich tussen mijn benen schoot richting zetel."Jij luie trien" lachte ik haar toe.Ik aaide haar tussen de oren en spontaan toonde ze mij haar buik.(nlTenTen14-5364121) '[I] went inside while she darted between my legs toward the sofa."You lazy stupid woman," I said to her laughingly.I stroked her between the ears and she spontaneously showed me her belly.'b.Ik werd wild and riep: "Neuk me, asjeblieft!" "Zwijgen!"riep hij.Hij haalde zijn stijve lul uit zijn broek, "Zuigen, jij geile slet!" Gretig nam ik zijn lid in mijn mond en begon hem te pijpen.Hij begon zachtjes in mijn mond te stoten en ik werd [er] wild van.(nlTenTen14-6406021) 'I got excited and yelled: "Fuck me, please!" "Shut up!" he yelled.He got his hard dick out of his trousers, "Suck it, you horny slut!"I eagerly took his member in my mouth and started to suck him off.He started thrusting gently into my mouth and I got excited by [it].' Our sample contains nine cases like (18a) and eleven cases like (18b), making up an additional 10.00%.The majority of the remaining (21.50% of) examples resemble (19a) in that they explicitly convey a positive assessment of the addressee.The other instances include a few expressions of sympathy and a few attributions of luck, like (19b) and (19c) respectively.For the final seven attestations of YOUþNP, of which (19d) is one, the co-text only allows an identifying reading (even if a more evaluative interpretation is tempting; see Section 2.4.4 on you boys).'"Now I do have to say that my parents came round years ago..."You lucky person!My foster parents never came round.' d.Alles echter, wat afkeurenswaardig is, wordt openbaar gemaakt door het licht; want het licht maakt alles openbaar.Daarom wordt er gezegd: "Ontwaak, jij slaper; sta op uit de doden en Christus zal over je lichten."'(nlTenTen14-1282968) 'However, everything that is reprehensible will be exposed by the light; because the light exposes everything.This is why it is said: "Wake up, you sleeper; rise from the dead and Christ will shine on you."'Interestingly, four of the cases like (19d) come from religious texts, which may suggest that Dutch YOUþNP need not be evaluative in this type of very formal and perhaps archaic discourse, even in the singular (pace Davies, 1986: 99, who claims that identifying uses are restricted to the plural).It is also worth noting that such religious cases often involve the exercise of power, something that is an important function of impoliteness (e.g.Culpeper, 2008).

English
All recurrent nouns and adjectives in English YOUþNP and the first ten unique ones (alphabetically) are listed in Table 7 (see also the Supplementary Material) and the summary information is given in Table 8.The proportion of cases modified by an adjective in this language (52.50%) is virtually the same as that in Dutch (51.00%).There is, however, a dissimilarity in number: there are significantly more plural instances of YOUþNP in English (27.00%) than in Dutch (15.00%; c2 ¼ 8.04, p < 0.05).More research is needed to find out the reason for this difference.
As the numerous nouns like idiot and bastard and the numerous adjectives like stupid and filthy suggest, YOUþNP is predominantly employed for impoliteness in English.Its proportion of impolite cases (75.00%) is slightly, but not substantially, higher than the Dutch one (68.50%).For co-textualised examples with an evaluative noun, an evaluative adjective, the combination of the two and a non-evaluative NP that YOUþNP coerces into an evaluative reading, see (20a) to (20d) respectively.
Fig. 2 shows that, like in Dutch, there is no difference in the ratio of impoliteness between singular (74.83%) and plural instances (75.47%) but there does exist one between NPs without an adjective (84.21%) and NPs with an adjective (66.67%; c2 ¼ 8.19, p < 0.05).
The reason why the NPs with adjectives exhibit a lower ratio of impolite instances appears to be same as in Dutch too: the nouns in the non-impolite cases need modification more often to secure a non-negative evaluation e as in you gorgeous girl e or to 'alter' it e as in you romantic bastard.
This last, oxymoronic NP belongs to a group of twenty-four attestations (12.00%) that could still be argued to hinge on the impoliteness associated with English YOUþNP.They contain surface insults but are not intended or perceived as impolite in the settings in which they are used, like in (21) (cf.Culpeper 2011: 207e215 on mock impoliteness).The two types of setting in question, each of which accounts for twelve cases, are known from Section 4.1: sexually charged exchanges like (21a) and playful interactions between close participants like (21b).
The rest of the non-impolite corpus examples (13.00%) include a small number of cases involving compassion or luck, as in (22a) and (22b).English again resembles Dutch in this regard.Unmistakably positive evaluations of the addressee such as (22c), however, seem relatively less frequent here (eight instances) and simple identifications like (22d) relatively more frequent (twelve instances).
(22) a. [Lefko] shook in spasms of choking, wheezing laughter.Rabinowitz got up and put a comforting hand on Lefko's right shoulder."You poor man.
You poor, poor man."(enTenTen18-34261709) b.Happy Travels… Fun, think I've said this before, you lucky ducks!So neat hearing how your travel opportunities arise, and watching them unfold is like finding treasure.(enTenTen18-22338769) c.Another set of guys hanging around at the edge of the high school gym, waiting for you to ask them to dance.You sexy stud, you!Whether you need a quarterback to handle some bye-week blues or need some flex help because you're starting Shonn Greene and you hate your life, we've got your covered.(enTenTen18-8952886) d.Whoever wanted to drink had to be served.The other brother did the same with a jug of water calling out: "You Lords, water from God."When all had quenched their thirst, both came with a copper full of warm water.(enTenTen18-36477827) Table 9 Nouns and adjectives in Polish YOUþNP.

Polish
The overview of all recurrent nouns and adjectives in Polish YOUþNP and of the first ten unique ones (alphabetically) is provided in Table 9 (see the Supplementary Material too), and the summary information in Table 10.Of the three languages studied, Polish turns out to be most heavily skewed towards impoliteness.As Table 10 indicates, 92.00% of the Polish data has an impolite function, compared to 68.50% in Dutch and 75.00% in English (c 2 ¼ 34.9, p < 0.000).
As many as 96.00% (192/200) of the tokens are singular.Notably, as Fig. 3 makes clear, all plural instances in the data are impolite (compared to Dutch and English, for which we also found entirely positive evaluations, attributions of luck and so on).Consider (23).The reason for this difference between Polish and both English and Dutch merits further investigation.It might relate to the norms of Polish etiquette, according to which the informal forms of appeal/address to both wy and ty are per se perceived as less polite than the traditional ones Pan/Pani/Pa nstwo 'Sir/Madam/Ladies and Gentlemen' and should be used only in familiar relationships.It is therefore not unlikely that the speaker in ( 23) is offending the interlocutors both by using the negatively evaluative NP and the informal form of address.As it is perhaps more natural to be familiar with a single addressee rather than a group which can be addressed by the informal pronoun wy, impolite expressions in the plural like (23) may be less frequent in Polish due to their heavy pragmatic marking (a basic search for ty and wyþNP yields 5842 and 831 hits, respectively, pointing in the same direction).Interestingly, no examples of identifying YOUþNP in the plural e recall (22d) in English e were found in our sample.They do exist, however e as revealed by an additional search for WYþNP and one of its results in (24).
(25) a. Nie _ zartuj sobie, Hewo! Nigdy nie nazywaj mnie tak wie ˛cej, ty parszywa pluskwo!To, _ ze z toba ˛wsp ołpracuje wcale nie znaczy, _ ze cie ˛lubie ˛. (plTenTen19-3091330) 'Stop kidding me, Hewo!Never call me that anymore, you tacky bug!Just because I'm working with you doesn't mean I like you.' b.Zrozumiałe s? Tyle masz szarych kom orek _ ze chyba tak.p.s Z dupo kombinatorek a nawet tego nie potrafiłe s, ty miernoto!(plTenTen19-5305303) 'Got it?You have so many gray cells that I think so.p.s You're such a wide boy, but you were not even able to do that, you mediocrity!' As we argue in Section 5, this fact may be related to the construction's very strong functional bias towards impoliteness in Polish.
Like in English and Dutch, the bias towards impoliteness is also evident from what can happen to evaluatively positive or neutral NPs in YOUþNP.'Eagle eye', for instance, would traditionally be a positive assessment, but the speaker in ( 26) is clearly being sarcastic and expressing a negative evaluation of the interlocutor.The NP can be said to be coerced into an insult here.(26) Nie ma argument ow to zaczynaja ˛sie ˛wyzwiska?I taka kultura ma przemawia c za zostawieniem targowiska.No gratuluje, ty sokole oko.Lepiej patrz na łapy urze ˛dasom.(plTenTen19-2730506) 'When there are no arguments, insults begin?And such a culture is to speak in favor of preserving the marketplace.Well, congratulations, you eagle eye.Better look at the dirty hands of the officials.' Impolite uses of this type are perfectly acceptable in Polish spoken discourse, but scarce in our sample (four cases), where the vast majority of the data conveys impoliteness in a straightforward way.
Of the non-impolite cases in Polish, many are usages where the NP involved may actually be negatively evaluative on the surface but the co-text shows that YOUþNP is used for: banter (eight of the eleven instances); to express admiration, as in (27a); or to convey a sense of familiarity, like in (27b).Polish resembles both Dutch and English in this regard, except that its sample has fewer such 'superficially negative' uses in sexual situations.The only example is (27c).This comparative paucity may be an effect of the corpus composition and/or the traditionally vulgar/offensive nature of words relating to sexuality in Polish. 10One could therefore hypothesize that, in the context of a sexual encounter, 'you' followed by an overtly rude NP (e.g.'perv', 'slut') would be interpreted as offensive rather than arousing.Overall, the non-impolite instances in Polish are too scarce to generalise about potential underlying reasons for their occurrence.Note, however, that many of the positively evaluative and identifying cases come from high-register literary language such as poetry and religious texts (cf.24), not unlike in English and Dutch.Note also that, in half of the non-impolite hits, what prompts the non-negative evaluation is the accompanying adjective, like in ( 16) with ty dzielny Podlasiaku 'you brave inhabitant of Podlachia!'.

Discussion
As indicated above, YOUþNP clearly has a strong functional bias towards impoliteness in Dutch, English and Polish.The study in Section 4 of actual instantiations of the construction in co-text shows that they are generally meant and/or taken as offensive.In most cases, YOUþNP actually features negatively evaluative nouns and/or adjectives.As argued above for examples of banter and dirty talk, such NPs do not always need to be impolite, of course.Yet, with Leech (1983: 142e145) among others, we would argue that these uses essentially depend on the potential for offense and that they therefore provide further evidence for the construction's predisposition for impoliteness.The bias towards impoliteness is also supported by two other findings from Section 4. When YOUþNP contains an NP that is, at face value, evaluatively neutral, it is apparent from the co-text that it typically functions as an insult after all.The construction could be said to coerce an impolite reading in such cases.Moreover, there is a relationship (statistically significant for Dutch and English in our data) between non-impolite interpretations and the presence of adjectives.YOUþNP often seems to require additional modification of nouns if one is to make sure that they are not perceived as conveying a negative evaluation.
Notwithstanding its obvious bias, the construction under examination evidently does not always serve impolite purposes.Other recurrent, though significantly less frequent, usages of YOUþNP in our Dutch, English and Polish data include attributions of luck, expressions of sympathy and genuinely positive assessments.Contrary to claims in the literature about the intrinsically evaluative nature of the construction, we even have found singular instances of YOUþNP that just identify the addressee.This use has been argued to be possible only for plural cases.Intuitively, we are inclined to agree that pure identification is more acceptable with multiple addressees.However, the low numbers of plural and identifying examples in our samples do not allow us to establish any significant correlation between number and identification.It is still interesting to note that the few cases in the data of singular YOUþNP with an identifying function occur only in literary and religious contexts.This phenomenon could thus well be restricted to such highly formal and somewhat archaic discourse.
In our view, the results summarised in the two previous paragraphs justify an analysis of YOUþNP as an impoliteness formula.The construction may not be impolite at all times but, to paraphrase Terkourafi's (2005a: 251) work on politeness (see Section 2.2 too) and her 'utterance-type meaning' type in particular, its very frequent co-occurrence with negatively evaluative NPs and with contexts in which offense is intended and/or taken creates a frame that makes the addressee likely "to derive the proposition that 'in offering an expression x [i.e.YOUþNP] the speaker is being [im]polite' as a generalised implicature", even if the specific NP itself conveys no negative assessment and without "full-blown inferencing about the speaker's intention".We can formulate this idea in terms more typical of the usage-based model too (see Section 2.4).If repeated use indeed affects cognitive representation, it is reasonable to assume that the high frequency with which YOUþNP is associated with impoliteness in usage in Dutch, English and Polish is reflected in the manner in which the construction is stored mentally.One possibility would be the existence of a schema that directly links the form in question to the function of negative addressee evaluation and occurs beside a schema for YOUþNP and addressee evaluation in general.We will remain largely agnostic about the details of this proposal (e.g. the nature of the network relations).However, we do contend, in line with usage-based insights (e.g.Schmid, 2007: 119e120), that any schema of the former kind would be more cognitively salient and more easily accessible than any schema of the latter kind, because of its sheer frequency and corresponding extent of entrenchment.
Importantly, as previously emphasised, a frame's 'strength' or schema's salience is a matter of conventionalisation and therefore of the degree or, put differently, "a correlate of the (statistical) frequency with which an expression is used in one's experience of a particular context" (Terkourafi, 2005b: 231).Our data in Section 4 indicates that it is very high for YOUþNP as an impoliteness formula in each of the languages under investigation.This fact can explain the following range of related phenomena that, though not necessarily attested in our samples, exist in all of them in our view: the construction tends to be interpreted as an insult (without any context too!) even if the NP is (i) evaluatively neutral (like English you theoretician; see Section 4.2), (ii) contains pseudowords (such as Dutch jij blug; see also Jain, 2022: 389), or (iii) is not actually spelt out, as in the Polish example in (28).Put differently, YOUþNP tends to coerce not just an interpretation as addressee evaluation but one as negative addressee evaluation in particular.'Charles was sitting in one of the armchairs not far from the window.He seemed to be working on some assignment.e How could you do this to her! You…! e And then he hit him…' The extent to which YOUþNP is conventionalised as an impoliteness formula may also fluctuate between languages.Our corpus data suggests that this form-function pairing exhibits a higher level of conventionalisation in Polish (92.00% of cases are impolite) than in English (75.00%) and Dutch (68.50%) and therefore that Polish YOUþNP probably merits some further attention.The degree to which it is conventionalised for impoliteness may be correlated with its significantly lower number of adjectives (and its lack of a difference in impoliteness between modified and non-modified instances; see Section 4.3).In Dutch and English, YOUþNP, adjectives frequently appear to serve to prevent a negatively evaluative interpretation.The Polish construction, however, is so heavily biased towards impoliteness that adjectives rarely get to fulfil this function.One may also wonder about the reason(s) for YOUþNP's high level of conventionalisation for impoliteness in Polish.In our view, the status of the vocative case in the language likely plays a role.The vocative is losing ground in speech, being perceived as too elevated a form of appeal (particularly with names) for 'normal' conversation.YOUþNP, which requires a vocative NP, would thus be unusual for most types of addresses and have specialised even further into impoliteness.In general, given the special position of Polish, we expect the tendencies described at the beginning of this paragraph (e.g.YOUþNP with pseudowords) to be even stronger in this language than in Dutch and English.This hypothesis has to be left for future research, though. 11 The final question that needs to be answered now is why YOUþNP is so well-suited to express impoliteness.On the surface, the second person pronoun seems to be somewhat superfluous in an address.For instance, even without the presence of 'you', parenthetical 'idiot' on its own would normally be understood as applying to the addressee.From a relevance-theoretical point of view (e.g.Sperber and Wilson, 1986), though, the second person pronoun must be assumed to make some contribution of its own in YOUþNP.What distinguishes 'you idiot' from just 'idiot' in our view is that the former overtly attributes the meaning expressed by the noun to the addressee.It could be said, in the words of Culpeper and Haugh (2014: 170), to increase the pragmatic explicitness of the address, spelling out its second person target openly.People might, depending on their culture, eschew such directness for reasons of politeness, particularly if they believe that the other person may find the content of their utterance disagreeable.In fact, it is one of Brown and Levinson's (1987: 131) politeness strategies to "impersonalise S[peaker] and H[earer]" and "avoid the pronouns 'I' and 'you'".Concerns of this kind do not come into play in impoliteness.On the contrary, someone hoping to offend another person may want to be as clear as possible about the focus of their attack or, as Culpeper (2005: 41) argues, to "explicitly associate the other with a negative aspect".Because of the presence of 'you', YOUþNP lends itself perfectly to this task.We would even hypothesise that it does the job of hurting the addressee's feelings better than an offensive address that does not contain a second person pronoun. 12This conjecture obviously needs to be tested, for example, by checking experimentally how offensive the impolite 'idiot' and 'you idiot' are perceived to be.However, such a study is beyond the scope of the present article.
The argument above is probably still only part of the story of YOUþNP's well-suitedness to impoliteness.It is entirely imaginable that someone wishing to evaluate another person in a positive way may also want to be pragmatically explicit and to overtly ascribe their assessment to their addressee.We have actually come across such cases in our data in Section 4 (e.g.Dutch jij prachtig medium 'you magnificent medium').It is nevertheless our contention that YOUþNP goes especially well with negative addressee evaluation.The rationale is that the second person pronoun may not only increase the directness of the

(
13) a.I'm surprised at your arrogant post hasn't gotten you flamed yet; you certainly deserve to be, you dolt.e I don't see how I would be considered a dolt and the post was not arrogant.(enTenTen18-35133812) b. "Great!Thanks a lot, you stupid bitch!" screamed Vultureman.Chilla was enraged by the unprovoked profanity.(enTenTen18-34608742)

( 15 )
Bassam explained that the Border Police soldiers were driving by the school in Anata, taunting the children by saying, "Come out, you heroes."… They routinely use the loudspeakers to yell profanity at homes while on patrol.(enTenTen18-13452138)

(
21) a.A babysitting AB/DL dome is awaiting and willing to gratify you, you mucky perv!Allow this nappy lovers stepmother to administer a bit of specialised adult baby regimes to you.(enTenTen18-27051412) b. "Doug, you're a legend for pulling this off."Here are some select other quotes: Vinnie: "Doug, you magnificent bastard.It definitely takes balls to remix this source, and I pretty much think you pulled it off as best as one could."(enTenTen18-4969597)

Table 2
Frequencies of YOUþNP versus personalised negative assertions in English and Polish.

Table 3
Frequencies of YOUþNP versus ERV in Dutch.

Table 6
Summary of Dutch YOUþNP.

Table 5
Nouns and adjectives in Dutch YOUþNP.

Table 8
Summary of English YOUþNP.

Table 7
Nouns and adjectives in English YOUþNP.

Table 10
Summary of Polish YOUþNP.