eJournals Arbeiten aus Anglistik und Amerikanistik 41/2

Arbeiten aus Anglistik und Amerikanistik
0171-5410
2941-0762
Narr Verlag Tübingen
Es handelt sich um einen Open-Access-Artikel der unter den Bedingungen der Lizenz CC by 4.0 veröffentlicht wurde.http://creativecommons.org/licenses/by/4.0/
The paper addresses the treatment of negation-related phraseological units in two specialised monolingual dictionaries (the Cambridge International Dictionary of Idioms and the Oxford Idioms Dictionary for Learners of English). The research is based on a database consisting of 595 phraseological units that are listed in the two dictionaries as negative or co-occurring with negation. Following the well-established distinction between inherently negative lexical items (e.g. nobody, nowhere) and non-negative lexical items licensed by negation and other polarity licensers (e.g., anybody, yet), the paper examines the dictionary entries to determine to what extent this dichotomy is presented in the dictionaries‟ introductory sections as well as on the level of concrete dictionary entries. The negative or the polarity-sensitive status of a phraseological unit is established by analysing the corpus data obtained from three British English corpora (the BNC, the enTenTen [2012], the UkWaC), and one American English corpus (the COCA). The data analysis shows that the divide between negative and polarity-sensitive phraseological units is not clearly and consistently presented in the two analysed dictionaries, and this may result in a dictionary user‟s wrong comprehension and usage. The paper puts forth a practical and user-friendly solution in the form of a tripartite model of classification that builds on theoretical considerations as well as real language use and that may be readily adopted by lexicographers.
2016
412 Kettemann

The Lexicographic Treatment of English NegationRelated Phraseological Units

2016
Gašper Ilc
The Lexicographic Treatment of English Negation- Related Phraseological Units Gašper Ilc The paper addresses the treatment of negation-related phraseological units in two specialised monolingual dictionaries (the Cambridge International Dictionary of Idioms and the Oxford Idioms Dictionary for Learners of English). The research is based on a database consisting of 595 phraseological units that are listed in the two dictionaries as negative or co-occurring with negation. Following the well-established distinction between inherently negative lexical items (e.g. nobody, nowhere) and non-negative lexical items licensed by negation and other polarity licensers (e.g., anybody, yet), the paper examines the dictionary entries to determine to what extent this dichotomy is presented in the dictionaries‟ introductory sections as well as on the level of concrete dictionary entries. The negative or the polarity-sensitive status of a phraseological unit is established by analysing the corpus data obtained from three British English corpora (the BNC, the enTenTen [2012], the UkWaC), and one American English corpus (the COCA). The data analysis shows that the divide between negative and polarity-sensitive phraseological units is not clearly and consistently presented in the two analysed dictionaries, and this may result in a dictionary user‟s wrong comprehension and usage. The paper puts forth a practical and user-friendly solution in the form of a tripartite model of classification that builds on theoretical considerations as well as real language use and that may be readily adopted by lexicographers. 1. Introduction In linguistic theory, negation is treated as a functional/ semantic operator that changes, i.e., negates, the truth value of a syntactic unit within its scope, be it a word, a phrase or a clause. In his seminal paper on negation in English, Klima (1964) proposes that the negator not, including its enclitic form n‟t, should be analysed as a spelt-out form of the abstract operator, which he refers to as neg. Even though the negative operator neg is AAA - Arbeiten aus Anglistik und Amerikanistik Band 41 (2016) · Heft Gunter Narr Verlag Tübingen 2 Gašper Ilc 74 most frequently spelt-out in English as the negator not (cf. Biber et al. 1999: 159ff.), it can alternatively be incorporated into lexical items, rendering them semantically negative. These items have traditionally been labelled as negative items (=NI). For example, the NI nobody can be decomposed into the operator neg and the indefinite pronoun anybody: [neg+anybody]. Similarly, never and impossible can be analysed as [neg+ever] and [neg+possible], respectively. Although the incorporation of the operator neg into nouns, adverbs, and quantifiers has been left undisputed in the literature, its realisation as the negative affixes unand inremains questionable. While it is the case that a lexical item with the negative affixes unand incan trigger negative syntactic environments in the same way as other NIs (cf. Klima 1964: 291-2), it is often the case that (at least) semantically, these items do not necessarily constitute negatives, but form opposites, in terms of Lyons‟ (1977: 279) gradable or complementary antonyms (frequent/ infrequent and born/ unborn, respectively). In addition, Horn (2001: 280) observes that lexical items with the negative prefixes unor inare often understood pejoratively, whereas the same items with the prefix non- are understood as neutrally negated (e.g., irrational and un-Christian vs. nonrational and non-Christian). Hence, it can be claimed that the so-called affixal negation by unor in-, despite its syntactic similarities to proper negation, should be analysed as an antonymic word formation process (cf. Verhagen 2005, Paradis and Willners 2013 among others). For this reason, the affixal negation is excluded from the present investigation. Any account of negation must distinguish between the NIs and the lexical items that are incapable of expressing negation, yet typically occur within the scope of negation. These items have traditionally been referred to as negative polarity items (= NPIs). In English, the difference between NIs and negative polarity items can be observed in pairs such as nobody~anybody and never~ever. In (1a,b), the negator not and the NI nobody syntactically negate the sentence, their negative status being confirmed by the presence of a positive question tag. As (1c,d) show, the negative polarity item anybody 1 can appear in negative sentences but not in their positive counterparts. (1) a. Peter has not seen anybody, has he? b. Nobody has seen him, have they? c. *Peter has seen anybody. d. Peter has not seen anybody. It has to be pointed out that the term negative polarity item may be a misnomer that gives rise to possible misinterpretations. Detailed analyses of 1 A clear distinction must be kept between the negative polarity item anybody, as in (1a), and the free-choice item anybody, as in You can meet anybody you like. For details, see: Chierchia (2006) and Ginnakidou and Quer (2013). English Negation-Related Phraseological Units 75 negative polarity items - items like anybody in (1a,d) - by Fauconnier (1975) and Ladusaw (1979, 1983) have, in fact, shown that they are also found in syntactic contexts that are non-negative. What these contexts have in common - with each other and also with negation - is that they are established by the presence of special licensers that are downward entailing and/ or non-veridical (cf. Zwarts 1995, van den Wouden 1997, Hoeksema 1997, Giannakidou 1998). 2 These studies have also shown that there is cross-linguistic variation as to what qualifies as a polarity licenser. In English, for example, the list of the most typical licensers includes lexical items such as without, lest, before, until, since, few, little, only as well as various syntactic structures including negatives, questions, conditionals, imperative-like conditionals, and comparatives (cf. van den Wouden 1997, Hoeksema et al. 2000, Condoravdi 2010). The described dependencies are exemplified in (2). (2) a. I stumbled but was able to catch myself before anything tragic occurred. 3 b. Few saw anything morally wrong with their crimes. c. By the way, has anybody seen Redmond O'Neill recently? d. If you have seen anything, please tell me. e. The Daedalus ship was far more complex than anything mankind had yet created. f. There are many people who think that it is too difficult to do anything on their own. 2 What detailed investigations into the phenomenon of negative polarity have shown is that it is extremely difficult to identify a sole criterion that can be applied universally for determining negative polarity contexts. Downward-entailment (cf. Fauconnier 1975, Ladusaw 1979, 1983) and non-veridicality (cf. Zwarts 1995, Giannakidou 1998) are the two criteria that have been typically used for detecting negative polarity contexts. Since the focus of the paper is not negative polarity per se, the paper adopts these two criteria for identifying English negative polarity contexts (for alternative approaches, cf. van den Wouden 1997). Downward-entailment: an expression is downward entailing if the meaning of A entails the meaning of the subset of A, e.g. it didn't rain (A) downward entails it didn't rain heavily (subset of A). By contrast, an expression is upward entailing if the meaning of A entails the meaning of the superset of A, e.g. it rained hard (A) upward entails it rained (superset of A). (Non-)veridicality: a propositional operator f is veridical if f entails p, i.e., iff for all propositions p, whenever f(p) is true, p is true, too. A propositional operator f is non-veridical if f does not entail p, i.e., iff for all propositions p, whenever f(p) is true, p may or may not be true (cf. Ginnakidou 1998: 160ff.). For example, the sentence George saw Mary is veridical since it asserts the truthfulness of the claim, whereas George might have seen Mary is non-veridical, because it does not assert the truthfulness of the claim. 3 As pointed out by the anonymous reviewer, this typically occurs in the counterfactual usage of before. Compare, for example, with the factual usage in Before she fell in love with someone really intelligent, she dated all kinds of men. Gašper Ilc 76 Before addressing the question of the lexicographic treatment of negation-related phraseological units, a word on phraseological units themselves is necessary. In phraseology, no agreement has yet been reached about how to properly name lexical units that have been traditionally referred to as idioms (cf. Cowie 1998). In lay terms, idiom has been used as a cover term for prefabricated lexical units or word combinations that consist of two or more lexemes functioning syntactically as well as semantically as inseparable units. Different terminological solutions have been proposed, including phraseological units (Gläser 1988, Vrbinc and Vrbinc 2014), set combinations (Zgusta 1971), phrasemes (Mel‟čuk 1995), and word-combinations (Cowie 1988, Howarth 1996). Since the present investigation is not theoretical in nature, but focusses on the treatment of negation-related phraseological units in specialised dictionaries, the paper simply adopts the terminological limitation as proposed in Vrbinc and Vrbinc (2014: 135ff.): the term phraseological unit (henceforth: PU) is applied to all word combinations that consist of two or more lexemes and display fixed syntactic and semantic properties. It is imperative for the analysis presented hereafter, however, that a distinction be made between PUs with a sentence-like structure and the PUs with a word-like structure. Following Gläser‟s (1988) classification, the terms propositions and nominations are used to label the two categories, respectively. Examples of nominations from my database include the PUs not harm/ hurt a fly; not be half bad; in the middle of nowhere, and instances of propositions include if you can‟t beat them, join ‟em; beggars can‟t be choosers; don‟t count your chickens before they are hatched. The nominations are typically referred to as proverbs and integrated only textually but not structurally. The paper also follows the assumption that PUs are retrieved from the mental lexicon as units (cf. Jackendoff 1995), and that “the selection and processing of an idiom representation is highly similar to the selection and processing of a single word” (Sprenger, Levelt and Kempen 2006: 176). To differentiate between single-word lexical items and multi-word items, the authors propose the term superlemma for the entire multi-word PU, and the term simple lemma for single-word lexical items constituting a superlemma. When a superlemma is retrieved from the mental lexicon, only its syntactic features and not the syntactic features of its constituent simple lemmas are accessible to syntactic operations (Sprenger, Levelt and Kempen 2006: 176-8). In other words, the internal structure of a PU, however similar it may be to a freely composed syntactic structure, is by and large inaccessible to syntactic processes. For example, the transitive verb move, which constitutes the superlemma move a muscle in (3a), is inaccessible to the syntactic process of passivisation (3b). 4 4 It is noteworthy that the notion of fixedness applies to non-compositional and nontransparent phraseological units (cf. Glucksberg 2001). The fixedness of these English Negation-Related Phraseological Units 77 (3) a. Jane didn‟t move a muscle. b. *A muscle wasn‟t moved (by Jane). 2. Aim and objectives Re-examining the PU move a muscle in (3a), a question may arise as to whether negation, spelt out as the negator n‟t, is part of the PU or not. Theoretically, there are at least three possibilities, each resulting in different syntactic patternings of the PU. Thus, we can talk about three distinct syntactic classes: ‣ CLASS A. Negation is an inalienable part of the PU, rendering the PU inherently negative, i.e., an NI (cf. nobody in (1b)). ‣ CLASS B. Negation is not a constituent part of the PU, but is a licenser under whose scope the PU appears, rendering the PU an NPI (cf. anybody in (1a)). ‣ CLASS C. Negation is the operator that freely (i.e., only grammatically) modifies the PU - in the same manner as the tense operator [+ PAST ], for example, modifies the PU in (3a). A quick glance at the treatment of the PU move a muscle in the two specialised idioms dictionaries included in the present investigation - the Cambridge International Dictionary of Idioms (henceforth: CIDI) and the Oxford Idioms Dictionary for Learners of English (henceforth: OIDLE) - reveals that the entries for the PU are not move a muscle in both dictionaries, which favours CLASS A analysis. A search in the three English corpora, viz. the BNC, the UkWaC, and the enTenTen[2012] displays 24, 103 and 1,093 hits for move__a__muscle respectively. The corpus of contemporary American English COCA displays 97 hits for the same query. On the basis of the collected examples, it can be concluded that most of the examples, indeed, appear in negative contexts. However, as shown in (4) below, a few examples of the PU move a muscle have been identified, which appear in non-negative contexts, but in all of these cases there is a polarity licenser present. Consequently, these examples lead us to the conclusion that the PU move a muscle is not an NI but is an NPI - a CLASS B item. No example of move a muscle was found in non-negative and non-polarity sensitive contexts that would support CLASS C analysis. items is relative, since there is a possible variability in terms of lexical and/ or grammatical modification (cf. Moon 1998, Vrbinc and Vrbinc 2011). Gašper Ilc 78 BNC a. It seemed all over in the 63rd minute when Clough, a few yards outside the penalty area, volleyed a headed clearance instantly into the roof of the net before Hardwick could move a muscle. → licenser: before ukWaC b. Almost before any of the party had time to move a muscle, two sharp cracks were heard, and both swans fell stone dead, with a heavy splash, at the margin of the lake. → licenser: before enTenTen[2012] c. He had me sit on the sitting room‟s floral sofa and told me he‟d forcefeed me jalapeno-chili enchiladas for lunch if I moved a muscle. → licenser: conditional d. Like anybody would do, I hid under my covers too scared to move a muscle. → licenser: adjectival intensification with too COCA e. But before Hunter could move a muscle, Devlin started talking again. → licenser: before f. Move a muscle and you won't have a date for the prom. → licenser: imperative-like conditional (i.e., If you move a muscle, then you…) g. If the Bostonian moved a muscle in retaliation, she would be dead. → licenser: conditional A question may arise at this point as to why there appears to be a discrepancy between the dictionary entries and the real language use. The answer can be sought in Schmidlin (2007: 554), who points out that “[t]he relationship between a dictionary entry and its verifiability in texts is one of the main problems in lexicography,” so to overcome this problem, lexicographers often rely on the frequency distribution criterion (cf. Moon 1998). Treating negation-related phraseological units solely within this approach may have two undesirable side effects. First, since negation is by far the commonest licenser of the NPIs (cf. Hoeksema 1997, Richter, Fritzingher and Weller 2010: 87), the frequency distribution criterion cannot clearly differentiate between NIs and NPIs: an NI always occurs with negation, because negation is its constituent part, an NPI mostly (4) English Negation-Related Phraseological Units 79 occurs with negation, because negation is its commonest licenser. In both cases, the frequency is high. Second, the application of the frequency distribution criterion affects the type of the examples selected to illustrate the usage of NPIs. Once again the negative sentences display the highest frequency, and are thus, by default, selected to illustrate the usage. As a consequence, more PUs may be classified by lexicographers as NIs than there really should be according to their syntactic nature. This is not an insignificant problem and has already been addressed by Stantcheva (2006), who in her analysis of German negation-related PUs and their lexicographic treatment concludes that many German PUs classified as NIs are not inherently negative. Therefore, in subsequent sections, the paper analyses the negationrelated phraseological units as they appear in the CIDI and the OIDLE and addresses the following research questions. ‣ RQ1: To what extent do the dictionary entries reflect the syntactic patterning of PUs as found in corpora? ‣ RQ2: Is there a clear divide between options PUs that are NIs, NPIs or freely occurring with negation (see CLASSES A, B and C above, respectively)? ‣ RQ3: More specifically, how are the NIs and NPIs differentiated? ‣ RQ4: Which information, if any, is provided for the syntactic context of the NPIs? 3. Methodology The methodology used for the present investigation is both quantitative and qualitative in nature. For the purposes of my investigation two specialized monolingual dictionaries of idioms with the highest number of entries - the Cambridge International Dictionary of Idioms and the Oxford Idioms Dictionary for Learners of English - were consulted. The PUs that are listed in the CIDI and/ or the OIDLE together with negation or labelled as either negative or occurring in negative sentences were manually extracted from the dictionaries. The final data list contains 362 entries from the CIDI and 529 entries from the OIDLE, 296 of which appear in both dictionaries. This list of dictionary entries served as the dataset for the investigation. The PUs from the dataset were then analysed and checked in three British English corpora: the BNC, the enTenTen [2012], and the UkWaC, as well as in one American English corpus, the COCA, to determine whether they are NIs or NPIs or whether they frequently co-occur or collocate with negation. Gašper Ilc 80 4. Results 4.1. The treatment of negation-related PUs in introductory sections and lexical entries The front matter of the CIDI (1998: xiii) explains that the negator not preceding the boldfaced idiom in the dictionary entry indicates that “[t]his idiom is always used in negative sentences.” 5 The dictionary provides no other information for the PUs listed in the macrostructure with regard to their (non)-negative or polarity status. The OIDLE, on the other hand, gives a more detailed treatment of the negation-related PUs on the level of the concrete entries, and shows that these PUs do not form a uniform category. Analysing the entries in detail, we can observe four different treatments of the PUs regarding negation: i. the PU is listed together with the negation (e.g., not care/ give a hoot); ii. the negation is listed together with the PU but appears in parentheses (e.g., (not) breathe a word); iii. the PU is listed either with negation or without negation (e.g., (not) at all, anything like), and special usage information “used with a negative, in a question or in an if-clause” is given in parentheses; iv. the PU is listed as non-negative (e.g., if and/ or but), and special usage information “often used in negative sentences” is given in parentheses. The front matter of the dictionary, however, does not provide a detailed explanation of the four categories, so we can only speculate about their differences. PUs belonging to (i.) are perhaps the most straightforward, since it seems that they are treated as NIs. In category (ii.), the parentheses may either indicate that the PU is licensed by the negation or simply frequently occurs with negation. Since negative, interrogative and conditional clauses constitute by far the most frequent polarity licensing contexts, it can be deduced that PUs belonging to category (iii.) are NPIs. It is unclear how we can classify category (iv.). This category may again involve NPIs or PUs often co-occurring with negation. What is more, there seems to be an overlap between categories (ii.), (iii.) and (iv.): PUs belonging to categories (ii.), (iii.), and (iv.) can include any PU belonging to CLASSES B and C of the theoretically based classification presented in section 2.2 above. 5 That this is a gross oversimplification can be proved simply by analysing examples provided in the entry. For instance, the entry not move a muscle contains a syntactically non-negative example: She sat without moving a muscle as the nurse injected the anaesthetic. English Negation-Related Phraseological Units 81 4.2. The syntactic structure of the PUs With regard to the type of the negative constituent, the PUs from the data list can be classified into the following categories: 6 1. PUs with the negator, e.g. not have a bean, last but not least; 2. PUs with the negative determiner/ quantifier, e.g. spare no expense/ pains/ trouble, reach the point of no return; 3. PUs with the negative pronominals, e.g. be none the wiser, like nobody‟s business; 4. PUs with the negative adverbs, e.g. lightning never strikes twice, nowhere to be found/ seen; 5. PUs with the negative correlative neither … nor, e.g. be neither fish nor fowl; 6. PUs with the negatives of foreign origins, e.g. persona non grata, je ne sais quoi. In the OIDLE, but not in the CIDI, examples can be found in which two or three negative elements are interchangeably listed for the same PU, for example, never/ not darken your door again, on no account/ not on any account, there‟s nothing/ not much, be none of somebody‟s business/ be no business of sb‟s. The analysis of the PUs in Table 1 shows that the negative constituents under investigation are not equally represented. The vast majority of the PUs contains the negator: 66.85% in the CIDI, and 60.49% in the OIDLE. The next most frequently-used negative constituent is the negative determiner, which amounts to approximately 20% of all examples in both dictionaries. None of the other analysed constituents represents more than 10% of the examples from the database. 6 Due to space limitations, only some examples are provided for the category under discussion. This principle is followed throughout the paper. Gašper Ilc 82 negative constituent CIDI (N= 362) OIDLE (N=529) N % N % negator 24 2 66.85 320 60.49 negative determiner 74 20.44 110 20.79 negative pronominals 22 6.08 47 8.88 negative adverbs 16 4.42 19 3.59 semi-negative adverbs 1 0.28 1 0.19 negative correlatives 3 0.83 2 0.38 negatives of foreign origins 4 1.10 4 0.76 interchangeable negative constituents 0 0.00 26 4.92 Table 1: Frequency of the negative constituents 4.3. Corpus data The corpus data analysis reveals that the PUs from my database can be neatly categorised into the three classes introduced in section 2.2 above. Approximately 40% 7 of all the items from my database belong to CLASS A, the inherently negative PUs. Items belonging to this category may either be propositions or nominations, though the former category is far more typical (approx. 80% of the examples). For instance: ‣ Propositions: if you can‟t beat them, join ‟em; beggars can‟t be choosers; better the devil you know than the devil you don‟t; you can‟t make brick without straw; if it ain‟t broke, don‟t fix it; butter wouldn‟t melt in sb‟s mouth; you can‟t teach an old dog new tricks; don‟t sweat it; you can‟t judge a book by its cover. ‣ Nominations: not with a bang but with a whimper; not be half bad; not know if/ whether you are coming or going; not know whether to laugh or cry; not know which way to turn; not have a minute to call. In addition, many of the PUs belonging to this class, contain a simple lemma with the incorporated negative operator neg in terms of Klima (1964, see section 2 above) rather than the negator not, for example: it‟s now or never; a watched pot never boils; it never rains but it pours; dead men tell no lies; better late than never; long time no see; the point of no return. C LASS B items, i.e., the PUs that are NPIs, represent approximately 30% of the items from my database list. Corpus data reveal that these PUs can also be found in affirmative sentences; however, in such cases 7 Only an approximate number can be given for the three classes, due to different labelling of the PUs with regard to negation in the analysed dictionaries; see section 4.1 above. English Negation-Related Phraseological Units 83 their syntactic well-formedness depends on the presence of a polarity licenser as exemplified in the case of the PU move a muscle in section 2.2 above. Some examples of polarity sensitive PUs from the database include: bat an eye/ eyelid/ eyelash; believe your ears; have a care in the world; know the half of it; budge/ give an inch; know the first thing about it; let the grass grow under your feet; a red cent; turn a hair; sleep a wink. Many of the PUs belonging to CLASS B PUs can be labelled minimizers or maximizers as defined and proposed by Israel (2001). A minimizer is a lexical expression that refers to a minimal scalar degree, whereas a maximizer, predictably, is its opposite, and refers to a maximal scalar degree (Israel 2001: 299-300). Thus, in (5) the minimizer lift a finger emphasizes the lowest possible degree. (5) a. And you don't even mention that the fat bugger hasn‟t lifted a finger in the last eight years to help the cause. b. But I knew that if I lifted a finger against him, he‟d have it his way. Additional examples of minimizers include bat an eye/ eyelash/ eyelid; move a muscle; sleep a wink; budge/ give an inch. Maximizers such as with a ten foot pole; all the tea in China; wild horses; for the life of me; for love or money on the other hand express the high degree of emphasis: (6) a. If you change the bottle, will guys even touch it with a ten foot pole? b. I wouldn‟t do this for all the tea in China. c. On the weekends, only wild horses could get me out. The pragmatic value of maximizers can be best observed in cases in which these lexical items are not fully integrated into the clausal structure but rather function as a speaker‟s comments or evaluations of the proposition i.e., when they have a disjunctive role: 8 (7) a. But I cannot, for the life of me, understand our motivations, thousands of years later, still following the conmen of yesteryear into our gory, bloody, violent end. b. The Big „C‟ Compliance cannot be delivered as a service, nor by Santa Claus, not for all the tea in China. c. I could not think of that one perfect word, not for love or money. Approximately 30% of the PUs from my database belong to CLASS C items, which are freely used with negation (i.e., negation is only an operator syntactically modifying the PUs). Even though these PUs are listed as negative or negation-related in either the OIDLE or the CIDI, the corpus 8 In this function, they typically reinforce or strengthen the negative meaning of the linguistic expression. This type of negation has been traditionally referred to as resumptive negation (cf. Jespersen 1917: 72ff.). Gašper Ilc 84 data reveal that they are inherently non-negative and non-polarity sensitive, as shown by (8) and (9). (8) a. I can‟t be arsed to go cut and paste the exact comment. → negation b. Do you want to be arsed with this? → polarity (interrogative) c. And there are a couple of boxes full of random wonderfulness that will find their way to a charity shop as soon as someone can be arsed. → affirmative, non-polarity (9) a. And today hasn't exactly been a barrel of laughs. → negation b. If you are a barrel of laughs, you may consider marking some funny logos like a right-pointing arrow with letters of “I am yours”, or a wedding image with “game over”. → polarity (conditional) c. Eugene is a barrel of laughs and a hard core conservative. → affirmative, non-polarity It is noteworthy that there seem to be two subgroups of the CLASS C items. The first subgroup comprises PUs that are non-negative and non-polarity sensitive, yet they frequently appear in negative context. Here, it seems that the co-occurrence of the negation and the PU is not syntactically motivated as is the case with CLASS B, but the relation between the two can be described in terms of collocation. A good example is the PU (can) be arsed (8), for which the statistics show that it mostly appears in negative sentences (Table 2). Corpus be arsed total N Negation +be arsed % BNC 16 81 UkWaC 624 63 enTenTen[2012] 1456 78 COCA 0 0 Table 2: Frequency of the PU be arsed A pattern similar to the PU be arsed can be observed with the PUs such as bet on it; in the biblical sense; break the bank; have a clue; be caught/ seen dead; have the faintest/ foggiest (idea); to put too fine a point on it; do things English Negation-Related Phraseological Units 85 by halves; room to swing a cat; say boo (to a goose/ a fly); sit well with sb; by any stretch of imagination; have two pennies/ nickels to rub together; hear the end of it; in all my born days. In the second subgroup, there are PUs that freely occur in the negative and non-negative contexts without any noticeable differences between the different syntactic contexts. An instance of an PU with this property is a barrel of laughs exemplified in (9) with its statistics shown in Table 3. Some other PUs belonging to this subcategory are: go a bundle on sth; be the marrying kind; be a bed of roses; be all fun and games; make bricks without straw; be sb‟s cup of tea; be sb‟s department; be the end of the world; hear yourself think; lose sleep over sth; Mr Nice Guy. Corpus a barrel of laughs total N negated % BNC 6 67 ukWaC 51 20 enTenTen[2012] 84 24 COCA 8 63 Table 3: Frequency of the PU a barrel of laughs My analysis has also shown that the occurrences of PUs vary across corpora; therefore, the selection of a particular corpus can influence the lexicographic treatment of (negation-related) PUs. This can be easily demonstrated just by statistically analysing the frequency values of the PUs be arsed and a barrel of laughs presented in Tables 2 and 3. As table 4 below shows, the difference between the frequency values in question are in most cases statistically significant (with at least p<0.05). Gašper Ilc 86 BNC UkWaC enTenTen [2012] Negation +be arsed Negation +a barrel of laughs Negation +be arsed Negation +a barrel of laughs Negation +be arsed Negation +a barrel of laughs UkWaC p<0.001 p>0.05 enTenTen p<0.01 p<0.0001 p<0.001 p<0.0001 COCA n.a. p>0.05 n.a. P<0.01 n.a p<0.01 Table 4: p-values for the occurrences of the PUs negation+be arsed and a barrel of laughs in the analysed corpora 5. Discussion A detailed analysis of the negation-related phraseological units from my database clearly shows that what specialized idioms dictionaries by and large treat as a uniform syntactic category is more complex. I have observed a gradience from inherently negative PUs to PUs frequently cooccurring with negation via an intermediate stage of negative polarity sensitive PUs (RQ1). This conclusion is also in line with Stantcheva (2006: 400), who observes that in the case of German negation-related PUs, dictionaries “reflect a usage of idioms that suggest no variation in the negative component.” In addition, Stantcheva‟s research (2006: 415) also reveals that many of the PUs classified by dictionaries as negative can also appear in affirmative contexts. The problem of a uniform treatment of negation-related PUs is especially noticeable in the CIDI, which offers only two alternatives: a PU is either inherently negative or inherently non-negative. The OIDLE has a more gradient approach, differentiating between PUs that are inherently negative from those that are not inherently negative but frequently occur with negation. In the latter case, the dictionary provides three options in the concrete lexical entries: (i.) negation appears as part of the PU but in parentheses; (ii.) the PU is followed by the special usage information that the PU is either “often used in negative sentences” or “used with a negative, in a question or in an if-clause”. Comparing this lexicographic system of differentiation and the theoretical classification presented in section 2.2 above, we can first observe that both approaches correctly identify PUs that are inherently negative. The problem arises when differentiating PUs that are NPIs from those that only co-occur/ collocate with negation. For example, it is not clearly explained in the OIDLE what the difference between PUs with negation in parentheses and PUs that are “often used in negative sentences” is. It could equally involve an NPI or an PU that collocates with negation. In addition, it appears that this system of differentiation is not implemented consistently. For example, the English Negation-Related Phraseological Units 87 PUs have a clue and a barrel of laughs, which both belong to CLASS C according to my analysis, are listed in the OIDLE as being different with regard to negation as their constituent (not have a clue and a barrel of laughs, respectively). Similarly, two NPIs, sleep a wink and breathe a word, which are both CLASS B items according to my analysis, are listed as not sleep a wink and (not) breathe a word. Therefore, even though the OIDLE makes an attempt at a more detailed treatment of negation-related PUs, it is inconsistent and does not accurately reflect real language use (RQ2, RQ3 and RQ4). While I believe that the system of classifying negation-related PUs laid out by the OIDLE is a step in the right direction, it calls for a more systematic and consistent approach. Combining the theoretical considerations regarding the negative and negative polarity items with the results of my corpus analysis, I propose the following tripartite model for classifying negation-related PUs. 1. Inherently negative PUs should always be listed together with negation - be it the negator not or a lexical item with the incorporated neg operator in terms Klima (1964), for example, not be half bad and better late than never respectively. In addition, to avoid any misunderstanding these items could also be labelled „negative‟ and/ or followed by a special usage information line „always used in negative sentences‟. 2. Negative polarity sensitive PUs should be listed without negation, for example, move a muscle, give an inch, and ideally followed by usage information „used in negatively-oriented polarity-sensitive contexts‟ in parentheses. This would be a theoretically perfect solution but would have, in my opinion, very little practical value for dictionary users, since it would involve a detailed and linguistically complex definition of polarity contexts (see section 1 and problems listed there). To strike a balance between theoretical correctness and user-friendliness, 9 I propose that the original usage information line from the OIDLE “used with a negative, in a question or in an if-clause” be preserved. This slight overgeneralisation still covers more than 90% of all occurrences of NPIs since negatives, interrogatives and conditionals 10 are by far the most typical licensers of NPIs. 3. PUs that frequently co-occur/ collocate with negation should again be listed without the negation, for example, do things by halves, say boo (to a goose/ a fly), followed by a special usage information line 9 The problem of adapting theoretically adequate descriptions for lexicographic purposes is similarly addressed by Hlebec (2015: 42ff.). 10 Strictly speaking, the term if-clause should be replaced with the term conditional since conditionality is not expressed syntactically by if-clauses only. In addition, not each and every if-clause is a conditional clause. Gašper Ilc 88 „often used in negative sentences‟, as can already be found in the OIDLE. PUs which do not fit into any of the three listed categories above should not be linked to negation in any way, because in those cases negation is simply a free syntactic operator modifying the PU. It must be made clear that the statistical analysis of the negationrelated PUs and their frequency values in different corpora (see Table 4) has shown that the selection of a corpus for lexicographic purposes may influence the final dictionary treatment of (negation-related) PUs. This fact may on the one hand explain the observed differences between the two dictionaries under investigation, since each dictionary is based on a different corpus. On the other hand, however, it also calls for a more complex lexicographic approach to (negation-related) PUs, which would involve the selection and the combination of different corpora. It can be claimed with a high degree of certainty that this is a necessary step, if the lexicographers want to give dictionary users a better insight into the actual language use. 6. Conclusion The analysis presented here has shown that the PUs that are listed as negation-related in the two specialised dictionaries of idioms (the CIDI and the OIDLE) do not represent a homogenous class of PUs. Using a syntactic classification, I have assigned negation-related PUs listed in the two dictionaries to three categories: (i) the PUs that are inherently negative, (ii) the PUs that are negatively-oriented polarity-sensitive, and (iii) the PUs that co-occur/ collocate with negation. The corpus data reveal that only 40% of the PUs that the dictionaries treat as negation-related are really inherently negative. The remaining 60% of the PUs are either negatively-oriented polarity-sensitive (30%) or freely co-occurring/ collocating with negation (30%). These data call for a lexicographic reanalysis of the PUs in question. The paper proposes a user-friendly solution in the form of a tripartite model of classification that can be readily adopted by lexicographers. While inherently negative PUs should be listed together with negation, and followed by a special usage information line „always used in negative sentences‟, negatively-oriented polarity-sensitive PUs and PUs cooccurring/ collocating with negation should be listed without negation, but followed by special usage information lines „used with a negative, in a question or in an if-clause‟ and „often used in negative sentences‟ respectively. I strongly believe that this strategy, if followed, will give any dictionary user a more detailed, adequate and realistic insight into the meaning and usage of such PUs. Moreover, the analysis has also shown English Negation-Related Phraseological Units 89 that a deeper insight into the nature and real usage of (negation-related) PUs can be obtained only if several corpora rather than only one serve as the basis for dictionary compilation. References Biber, Douglas, Stig Johansson, Geoffrey Leech, Susan Conrad & Edward Finegan (1999). Longman Grammar of Spoken and Written English. Harlow: Pearson Education Limited. Chierchia, Gennaro (2006). “Broaden your Views. Implications of Domain Widening and the “Logicality” of Language”. Linguistic Inquiry 37: 535-590. Condoravdi, Cleo (2010). “NPI Licensing in Temporal Clauses”. NLLT 28: 877- 910. Cowie, Arthur P. (1998). “Introduction”. In Arthur P. Cowie (ed.). Phraseology: Theory, Analysis, and Applications. Oxford University Press. 1-20. Fauconnier, Gilles (1975). “Polarity and the Scale Principle”. CLS 11: 188-199. Giannakidou, Anastasia (1998). Polarity Sensitivity as (Non)veridical Dependency. Amsterdam: Benjamins. Ginnakidou, Anastasia & Josep Quer (2013). “Exhaustive and Non-exhaustive Variation with Free Choice and Referential Vagueness: Evidence from Greek, Catalan and Spanish”. Lingua 126: 120-149. Gläser, Rosemarie (1988). “The Grading of Idiomaticity as a Presupposition for a Taxonomy of Idioms”. In Werner Hüllen & Rainer Schulze (eds.). Understanding the Lexicon: Meaning, Sense and World Knowledge in Lexical Semantics. Tübingen: Max Niemeyer. 264-279. Glucksberg, Sam (2001). Understanding Figurative Language. From Metaphors to Idioms. Oxford: Oxford University Press. Hlebec, Boris (2015). “Lexical Definitions of Some Performative Verbs”. Studia Anglica Posnaniensia 50/ 1: 27-44. Hoeksema, Jacob et al. (2000). Perspectives on Negation and Polarity Items. Amsterdam/ Philadelphia: John Benjamins Publishing Company. Hoeksema, Jacob (1997). “Corpus study of negative polarity items.” [online] Html version of a paper which appeared in the IV-V Jornades de corpus linguistics 1996-1997. (http: / / www.let.rug.nl/ hoeksema/ docs/ barcelona.html) (accessed 2 April 2015). Horn, Laurence R. (2001). A Natural History of Negation. Chicago: University of Chicago Press. Howarth, Peter Andrew (1996). Phraseology in English Academic Writing: Some Implications for Language Learning and Dictionary Making. Tübingen: Max Niemeyer. Israel, Michael (2001). “Minimizers, Maximizers, and the Rhetoric of Scalar Reasoning”. Journal of Semantics 18(4): 297-331. Jackendoff, Ray (1995). “The boundaries of the lexicon”. In Martin Everaert, Erik- Jan van der Linden, André Schenk & Rob Schreuder (eds.). Idioms: Structural and Psychological Perspectives. Hillsdale, NJ: Lawrence Erlbaum Associates. 133-165. Jespersen, Otto (1917). Negation in English and other Languages. Copenhagen: A.F. Høst. Jespersen, Otto (1924). The Philosophy of Grammar. London: George Allen&Unwin Ltd. Gašper Ilc 90 Klima, Edward S. (1964). “Negation in English”. In: Jerry A. Fodor & Jerrod J. Katz (eds.). The Structure of Language. Englewood Cliffs, New Jersey: Prentice- Hall. 246-323. Ladusaw, William (1979). Polarity Sensitivity as Inherent Scope Relations. New York/ London: Garland Publishing. Ladusaw, William (1983) “Logical Form and Conditions on Grammaticality”. Linguistics and Philosophy 9: 373-392. Lyons, John (1977). Semantics. Volume I. Cambridge: Cambridge University Press. Mel‟čuk, Igor A. (1995) “Phrasemes in Languages and Phraseology in Linguistics”. In Martin Everaert, Erik-Jan van der Linden, André Schenk and Rob Schreuder (eds.). Idioms: Structural and Psychological Perspectives. Hillsdale, NJ: Lawrence Erlbaum Associates. 167-232. Moon, Rosamund (1998) Fixed Expressions and Idioms in English: A Corpus-based Approach. Oxford: Claredon Press. Paradis, Carita & Caroline Willners (2013). “Negation and Approximation of Antonymic Meanings as Configuration Construals in SPACE”. In Carita Paradis, Jean Hudson & Ulf Magnusson (eds.). The Construal of Spatial Meaning: Windows into Conceptual Space. Oxford: Oxford University Press. 287-311. Richter, Frank, Fabienne Fritzinger and Marion Weller (2010). “Who Can See the Forest for the Trees? Extracting Multiword Negative Polarity Items from Dependency-Parsed Text”. JLCL 25(1): 83-110. Schmidlin, Regula (2007). “Phraseological Expressions in German Standard Varieties”. In Harald Burger, Dmitrij Dobrovol‟skij, Peter Kühn & Neal R. Norrick (eds.). Phraseology. An International Handbook of Contemporary Research. Volume 1. Berlin/ New York: Walter de Gruyter. 221-562. Sprenger, Simone A., Willem J. M. Levelt & Gerard Kempen (2006). “Lexical Access during the Production of Idiomatic Phrases”. Journal of Memory and Language 54: 161-184. Stantcheva, Diana (2006). “The Many Faces of Negation - German VP Idioms with a Negative Component”. International Journal of Lexicography 19(4): 397- 418. van den Wouden, Ton (1997). Negative Contexts. Collocation, Polarity and Multiple Negation. Routledge Studies in Germanic Linguistics. London/ New York: Routledge. Verhagen, Arie (2005). Constructions of Intersubjectivity: Discourse, Syntax and Cognition. Oxford: Oxford University Press. Vrbinc, Alenka & Marjeta Vrbinc (2011). “Creative use of idioms in satirical magazines”. Jezikoslovlje 12.1: 75-91. Vrbinc, Alenka & Marjeta Vrbinc (2014). “Phraseological Units with Onomastic Components: the Case of English and Slovene”. RLA 52 (1): 133-153. Zgusta, Ladislav (1971). Manual of Lexicography. The Hague: Mouton. Zwarts, Frans (1995). “Nonveridical Contexts”. Linguistic Analysis 25: 286-310. English Negation-Related Phraseological Units 91 Corpora BNC: The British National corpus. 2010 (https: / / the.sketchengine.co.uk/ ). enTenTen 2012: English web corpus. 2012 (https: / / the.sketchengine.co.uk/ ). UkWak: web-derived (domain .uk) corpus (https: / / the.sketchengine.co.uk/ ). COCA: Corpus of Contemporary American English (http: / / corpus.byu.edu/ coca/ ). Dictionaries CIDI: The Cambridge International Dictionary of Idioms (2002). Cambridge: Cambridge University Press. OIDLE: Oxford Idioms. Dictionary for Learners of English (2006). Oxford: Oxford University Press. Gašper Ilc University of Ljubljana Ljubljana, Slovenia