A topic to rack our clever brains on: Premodification in Hungarian and English body-part idioms

Barbara Bakó (Debrecen)

http://dx.doi.org/10.13092/lo.89.4270

1 Introduction

One of the most contentious topics receiving the broadest attention and discussion within idiom research concerns the lexical and syntactic variability of idiomatic expressions. In addition to alterations such as passivization or lexical substitution, adding various modifiers into idiomatic strings belongs to the scope of this issue as well. Many previous studies (e. g. Ernst 1980; Nunberg et al. 1994; Langlotz 2006a, 2006b, Stathi 2007; Cserép 2010) have pointed out that premodification in V + NP idioms is a frequently employed phenomenon in actual language use. The adjectives inserted into idioms, however, seem to differ in their function, and, thus, can be classified into several categories (Ernst 1980; Stathi 2007).

The corpus-based study presented in this paper focuses on a specific subclass of verbal idioms, examining the adnominal premodification patterns of expressions that contain body-part terms in their direct object positions. This investigation, which has been conducted in both the Hungarian and the English languages, is mainly concerned with the various categories of adjectival modifiers that can occur with these idioms. However, it also touches upon the semantic autonomy of the body-part constituents in relation with the premodification types. The theoretical background on which this study is based involves Ernst’s (1980) three-way distinction and Stathi’s (2007) five-level taxonomy of modification. Therefore, before a detailed discussion of the results, these two studies will be summarized in section 2.

2 Previous studies

2.1 Ernst (1980)

The first of the most influential contributions to this particular area of idiom research is Ernst’s (1980) taxonomy which – based on the results of an investigation of naturally occurring data in fiction, journalism, the television and radio – distinguishes three main types of premodifiers: external, internal, and conjunction (Ernst 1980: 51–53).

The basic differences between these three groups can be captured along two dimensions: the semantic scope of the modifier on the one hand, and the referentiality of the modified noun on the other hand. The first dimension addresses the question whether a particular adjective has only its head noun in its scope (which is true for the internal and conjunction types) or modifies the whole VP expressed by the idiom (which is true for external modifiers). In contrast, referentiality of the noun is interesting for making a distinction between internal and conjunction modifiers, as this issue concerns the question whether the head noun, which is semantically under the scope of the adjective, has a referent or not. To illustrate how this works in practice, consider the following sentences (Ernst 1980: 51–53):

Ernst (1980) argues that the adjective economic in (1a) belongs to the external category because it – despite of being attached to the noun leg syntactically – semantically modifies the whole idiomatic string in the same way as the corresponding adverb economically does in (1b).¹ Since external modifiers like economic play the same role as adverbials in a sentence, there is a wide agreement among researchers that the main function of this class of adjectives is to contextualize the idiom and to determine how it should be interpreted in the discourse (see also Moon 1998; Burger 2015; Dobrovol’skij 2000; Sabban 2000; and Minugh 2007, 2014). For this reason, Ernst (1980: 52) uses the term “domain delimiter” to refer to them.²

In contrast, the modifier horse-drawn in (2) is considered to be internal as it is attached to the head-noun bandwagon not only syntactically but also semantically. Bandwagon corresponds to the idiomatic interpretation ‘movement’, therefore, the phrase the horse-drawn Reagan bandwagon can idiomatically be understood as “Reagan’s political movement is old-fashioned and behind the times” (Ernst 1980: 52).

Although the adjective cross-gartered in (3) has only the noun leg in its scope, it does not belong to the class of internal modifiers. Given the fact that leg in pull someone’s leg has no referent in the idiomatic sense of the expression, cross-gartered can only modify its literal referent, i. e. Malvolio’s real body part, which is independent of the idiomatic meaning ‘make fun of, fool someone’. Since the interpretation of (3) involves both the idiomatic meaning of the expression and the literal meaning of the idiom constituent leg, Ernst (1980) calls this type of adjectives “conjunction modifiers”.³

2.2 Stathi (2007)

Ernst’s (1980) typology is revised by Stathi (2007), who examined the adjectival premodification patterns of idioms in the German language. Based on her research data, she argues for a five-level hierarchy of adjectival modification that consists of the above-mentioned three types proposed by Ernst and two newly introduced ones which she calls intermediate level and metalinguistic modifications.

Table 1: Stathi’s (2007) hierarchy of adjectival modification⁴

The Stathian (2007) term of internal modification is applied to cases where the modified noun is interpretable at the literal level. It is claimed that adjectives of this type typically occur with transparent metaphorical idioms and their main function is to activate the source domain, i.e. the underlying image of the expression, on the basis of which the modifier-noun sequence is mapped onto the target domain as a whole.

In (4), for instance, the adjective stärkste ‘strongest’ evokes the image of sailing, and receives its interpretation at the literal level. It is known that the strength of the wind influences the speed of a sailing ship: if the wind is strong, the ship moves fast, whereas a weak wind slows the ship down causing it to be unable to go forward. The same concept of intensity applies to the target domain where someone is making a very fast progress in an activity.

Regarding external modification, Stathi (2007) points out that one of the main properties of such modifiers is their semantic incompatibility with the head noun. This semantic clash may be due to the fact that external modifiers usually denote abstract concepts. Although these adjectives tend to function as domain delimiters, there are cases as well where they are used to express location, time, or cause. For instance, nach-saisonale ‘after-season’ in (5) denotes the time when Matthäus brought shame on himself:

As can be seen in Table 1, Stathi’s (2007) taxonomy distinguishes an intermediate level between the internal and external classes. This intermediate type of modification includes adjectives that are interpretable only in the target domain, modifying the noun only in its figurative sense. Stathi (2007) illustrates the behaviour of intermediate modifiers with the German adjective marode ‘rotten’, which usually occurs with abstract nouns denoting institutions and organizations.

Karren ‘cart’ in (6) – being specified by a genitive attribution – obviously stands for Islamic Revolution at the figurative level of interpretation, and the adjective maroden modifies this particular abstract meaning of the noun. Intermediate modification, therefore, implies that the object nominal has some semantic autonomy within the expression. It is important, however, to note that this abstract sense of the noun is available only in the idiomatic string and is not likely to occur outside it.

Conjunction modification is also a part of Stathi’s (2007) typology. Interestingly, this is the only category where the modified noun receives the property of referentiality (see Table 1). That is, as also pointed out in section 2.1, nouns occurring with conjunction premodifiers play a double role: (i) they are used non-referentially as part of the idiom, but (ii) refer to real entities at the literal level of meaning. Conjunctive premodifiers, thereby, are seen as context-embedding tools for idioms, which are used to provide as much background information about the referent of the noun as possible in the most economic way.

According to Stathi (2007), this category of modification is typical for (but not restricted to) idioms containing body-part terms. In such cases, the premodifier characterizes the body-part of the person mentioned in the discourse; however, it is also possible that the adjective-noun sequence, by means of metonymical shift, refers to the person rather than to his/her body-part. In (7), for example, the adjective unverschämtes ‘brazen’ does not supply additional information about the mouth literally, but metonymically describes its bearer (der Herr ‘man’) as brazen.

In addition to those discussed so far, adjectival modifiers also have a fifth type which is referred to as metalinguistic modification. Adjectives of this class (such as proverbial, metaphorical, and literal) are inserted into idiomatic strings in order to serve two closely related functions: (i) “to highlight and draw attention to the use of the idiom in the text”, and (ii) “to signal that the expression, which could also be understood literally, is to be interpreted idiomatically” (Stathi 2007: 102):

Stathi’s (2007) five classes of adjectival modification can be arranged hierarchically in a similar fashion as Fraser’s (1970) seven-level hierarchy of idiom transformation. The order of the levels is presented above in Table 1 and should be read from the top to the bottom: if an idiom permits a certain level of modification, it is also open to all modifications that are below it in the hierarchy. That is, if an expression allows external modification, it also allows the conjunction and metaphorical types, but not intermediate level or internal modification.

3 Adjectival modification in idioms with body-part terms

The study presented in this paper examines the adjectival premodification patterns of V + NP idioms containing body-part terms, in both Hungarian and English corpora. Moreover, it also concerns the semantic autonomy of the body-part constituents and its potential relationship with the various modification types. It is primarily based on the above-discussed five-class typology introduced by Stathi (2007), who found that all of the three German body-part idioms selected for her research occurred only with the conjunction and the hierarchically lower-level metalinguistic types of modification. Therefore, this investigation addresses the following three research questions:

3.1 Research data and methods

For both the English and the Hungarian parts of the study, 10 V + NP idioms with body-part constituents have been selected from various idiom dictionaries (Spears 2000; Siefring 2004; O. Nagy 1985; Bárdosi 2003). Both the Hungarian and the English groups of the expressions contained 5 idioms that could be found only in that particular language, while the other 5 idioms occurred in both languages.⁶

The adjectival premodification properties of the 20 idioms have been studied in large corpora. For the English part, the Corpus of Global Web-Based English (GloWbE) has been used, which contains 1.9 billion words of text from 20 different countries. The texts in this corpus consist of web-based materials such as personal blogs, company websites, magazines, and newspapers.

The premodified alterations of the Hungarian idioms have been checked in the Hungarian National Corpus (HNC) (Oravecz et al. 2014). It is the largest freely-available Hungarian corpus, containing more than 1 billion words from texts of five genres: official, press, spoken, personal, and academic. However, since only a small number of occurrences of premodified body-part idioms were found in this corpus, some search in the Google has also been done in order to collect more data for my study. Sentences that have occurred multiple times have been filtered out in both parts of the investigation.

3.2 Results

The results of this corpus-based study are presented and discussed in two subsections. Section 3.2.1 focuses on the first research question, hence, deals with the types of adjectival premodifiers that have occurred with the selected idiomatic expressions in the corpora, whereas section 3.2.2 aims to answer Questions 2 and 3, accounting for semantic autonomy and its role in premodification. Before a detailed look at these particular issues, however, two important facts should be noted with respect to the results:

3.2.1 Types of adjectival modification

Consistent with Stathi’s (2007) assumptions, conjunction modification showed the highest rate of frequency with the selected body-part idioms in both languages. Except for one English (lose one’s head) and two Hungarian ones (elveszti a fejét [lit. ‘lose one’s head’; fig. ‘lose self-control’] and otthagyja a fogát [lit. ‘leave one’s tooth somewhere’; fig. ‘to die’]), all expressions occurred with this particular type of adnominals.⁷

It has been pointed out in section 2.2 that conjunction modifiers can serve two functions when they modify body-part constituents; i. e., they can (i) describe the body part only, or (ii) metonymically supply information about the person whose body part is referred to by the head noun. Both possibilities were represented in my research data:

The sentences in (9) illustrate the case when the adjective (long and csokibarnára sült ‘chocolate-brown-tanned’ in these particular instances) modifies the literal body-part referent of its head noun (nose and láb ‘leg’, respectively). As opposed to this, a metonymic shift takes place in both (10a) and (10b), and not the body parts feet and segg ‘ass’, but their bearers are characterized as talented and városi ‘urban’ by the inserted adjectives. The role of conjunctive adnominals, however, is not always as clear as in these four examples. A not so unequivocal occurrence of this class of premodification is illustrated in (11) where a three-way ambiguity can be detected with respect to the possible functions and meanings of the adjective fat.

That is, similar to the sentences in (9) and (10), fat can refer to the physical property of either the body part or the people (bankers) who have been mentioned previously. A third – perhaps the most likely – interpretation, however, is implied by the larger context in which various financial institutions and their unfair policies against poor countries are discussed. In this sense, the adjective fat might metaphorically be understood as ‘rich’, emphasizing the huge opulence of banks and the distance between their and the poor countries’ power and financial situation. Therefore, the phrase kiss their fat asses in (11) may metaphtonymically be interpreted as ‘flatter the rich financial institutions/banks’.

In addition to conjunction modification, metalinguistic adjectives like proverbial or metaphorical have been found by Stathi (2007) to be the other category that can potentially be inserted into idiomatic strings with body-part terms. Interestingly, no occurrences of such premodifiers could be identified in this present study for either the English or the Hungarian expressions. Nevertheless, not conjunction modification was the only type available for the idioms in our research data: the other three categories, i.e. external, intermediate level and internal, were shown to be compatible with them as well.

Two unambiguous examples of external modification can be seen in (12a) and (13a). Both the domain delimiter politikai értelemben vett ‘taken in the political sense/political’ inserted into the Hungarian idiom otthagyja a fogát (lit. ‘leave one’s tooth there’; fig. ‘to die’) in (12a) and the adjective academic in (13a) can be paraphrased with their corresponding adverbials politikai értelemben véve/politikailag ‘(taken) in the political sense/politically’ and academically, as shown in the (b) sentences:

Due to limitations of space, it is not possible to include all instances of external modification in this paper, but some of them are listed below in the form of concordance lines.

(14), (15), and (16) contain domain delimiters, while first-team in (17) denotes location and can be understood as ‘the player found his confidence/stability in the first team’. Similarly, the premodifier központi ‘central’ inserted into the Hungarian idiom felnyitja valaki szemét valamire (lit. ‘open someone’s eyes to something’, fig. ‘make someone realize the truth about something) in (19) can be paraphrased as the locative adverbial a központban ‘in the centre’. In (18), the adjective unwelcome is somewhat ambiguous: it can function as an external modifier expressing the meaning ‘it was not gladly received that the Anglo Saxons and the Normans interfered’, or as a conjunction modifier metonymically characterizing these two peoples as unwanted.

The next level in Stathi’s (2007) hierarchy constitutes intermediate modification which is claimed to be semantically compatible only with the abstract meaning of the noun. Only the following instance has been found in our data to which this function may be assigned:

The Hungarian adjective becses ‘precious/respectable’ is generally used to characterize entities and abstract concepts that have an outstandingly high value (either in the material or in the emotional sense), as in the literal phrases becses drágakő ‘precious gem’ or becses hagyomány ‘precious tradition’. It is, however, not a regular collocation of the noun fog ‘tooth’; therefore, it can function neither as a conjunction nor as an internal modifier in (20).¹⁰ Moreover, external modification can also be ruled out in this case as becses does not modify the idiom holistically. Following from the figurative meaning ‘to die’ of the expression otthagyja a fogát (lit. ‘leave his tooth there’; fig. ‘to die’) the most likely explanation for the use of this adjective in (20) is that the noun fog is assigned a meaning that corresponds to one of our greatest values, i. e. life. Hence, the string otthagyni a becses fogát(lit. ‘to leave his precious tooth there’) becomes interpretable in a similar way as the expression lose his precious life. This assumption is also supported by the fact that the phrase becses élet ‘precious life’ as well as the co-occurrences of the noun élet ‘life’ with the synonyms of becses (drága ‘precious’, értékes ‘valuable’, tiszteletre méltó ‘honorable’) are used relatively frequently in the Hungarian language.¹¹

In contrast, consistent with Stathi’s (2007) definition of internal modification, the adjectives in (21)–(23) modify the noun literally, and they are mapped onto the target domain with the noun as a unit. For both the idiom felnyitja valaki szemét valamire and its English equivalent open someone’s eye to something, the premodifiers leragadt ‘stuck’ and blind in (21) and (22) evoke the same underlying mental image. That is, based on our general background knowledge, we know that if someone’s ability of seeing is impaired, that person is unable to perceive the world in its full details. If he gets his eyesight (back), information that can only be accessed via vision becomes also available for him. This image relates to the target domain where someone is provided with previously unknown facts, by which he is able to approach something from a new perspective. Enyves ‘gluey’ behaves according to the same principle in (23) where the activated image involves a gluey hand to which every single item that the person touches sticks. This is in accordance with the often-used meaning of the idiom ‘to steal something’.

Cserép (2010: 107) points out that there are cases in which both the adjective and the noun have their own figurative senses and contribute to the overall meaning separately rather than as a unit. This assumption is supported by (24) and (25) where the adjectives mocskos ‘filthy’ and dirty can be interpreted as ‘unethical/unfair’, while the noun hand (kéz in Hungarian) may refer to ‘influence/possession’. Since the unethical nature of the act is conveyed even by the canonical form of the idiom, these adjectives can be seen as tools that intensify this particular feature of the event.

The research data collected for the current study also contained some instances that did not really fit into any of Stathi’s (2007) five categories. These adjectives did not give any information about the head noun or the VP at either the idiomatic or the literal level of interpretation. Consider the following sentences:

All adjectives in (26)–(30) can be considered semantically “empty” in the sense that they do not contribute to the propositional meanings of the sentences. Instead, their role is to provide some extra emotional content, expressing the speaker’s feelings and attitude about the event denoted by the VP. A closer look at the dictionary meanings and the general uses of the adjectives supply evidence in favour of such an interpretation. According to the Oxford Advanced Learner’s Dictionary (2005: 154, 627), both bloody and fucking can be defined as “a swear word that [...] is used to emphasize a comment or an angry statement”. To the Hungarian word átkozott ‘cursed’, the online version of A Magyar Nyelv Értelmező Szótára [The Hungarian Language’s Explanatory Dictionary] assigns the function of referring to things that cause someone an extreme extent of annoyance. By analogy with these dictionary senses, bloody, fucking and átkozott in the above examples can all be treated as intensifiers that express the speaker’s anger about the American interference in (26), the lack of his own knowledge in (27), and the addressee’s act of losing his self-control in both (28) and (29). Although the intensifier role is not listed for redvás ‘carious’ in the dictionary, its occurrences in ordinary phrases prove the existence of this function. That is, all of its 41 instances in the Hungarian National Corpus behaved in the same way as the above-mentioned three adjectives. Adapting McClure’s (2011) term, I will use the name expressive to refer to this type of modification.

3.2.2 Semantic autonomy of the noun

In addition to investigating the function of the adjectives that occurred with the 20 idioms, this study also concerns the semantic autonomy of the body-part constituents in these expressions. As Langlotz (2006b) points out, the concept of semantic autonomy refers to the phenomenon when an idiom component develops a lexicalized figurative meaning that is available not only within that particular idiom but also outside it. For example, both the verb swallow and the noun phrase bitter pill in the expression swallow the bitter pill can be used with the meanings ‘accept’ and ‘unpleasant fact’ in contexts other than the idiomatic phrase.

This should be distinguished from the term “relative semantic autonomy” which applies to cases where the constituents acquire abstract senses that can be accessed only when they occur in the idiom. The components of rock the boat, for instance, have such phrase-induced figurative senses since neither rock nor boat can be interpreted as ‘spoil’ and ‘comfortable situation’, respectively, when they are not part of the idiomatic string.

In our study, there were only two idioms whose nominal constituents had a lexicalized figurative sense: lay one’s hands on something and its Hungarian equivalent ráteszi a kezét valamire. In these expressions, the noun hand (kéz in Hungarian) can figuratively be interpreted as ‘influence/occupancy’. This meaning of the noun, however, is not restricted to this idiom, but can be detected in some other phrases as well; for example, in the English have someone/something in hand or in the Hungarian rossz kezekbe kerül (lit. ‘wrong hands-into get’; fig. ‘get into wrong hands’).

In contrast, the nouns in the pair of lose one’s head and elveszti a fejét, as well as palm in grease someone’s palm have phrase-induced figurative senses, i.e. they can metaphorically and metonymically be identified as ‘self-control’ and ‘person’, respectively, only in these particular expressions.

These were the only idioms in our investigation whose body-part constituents could be considered as showing some degree of semantic autonomy. In the case of the other 15 strings, no correspondences could be established between the individual idiom components and (parts of) the overall figurative meanings. These expressions, therefore, should be treated as semantically opaque or non-decomposable idioms.

With respect to the relationship between semantic autonomy and adjectival premodification, we could earlier see that only two types, intermediate level and internal modifications are able to indicate that the nominal constituent of the idiom has its own (either lexicalized or phrase-induced) figurative sense. Intermediate level modification has been claimed to operate only at the abstract level, which obviously requires the head noun to have an independent meaning that contributes to the overall idiomatic interpretation. I agree with Stathi (2007: 104) in that her internal modification is not necessarily a signal of the semantic autonomy of the noun, since it constitutes the mapping of the modifier-noun sequence onto the target domain as a whole. Nevertheless, the literal-and-figurative type of internal modification proposed by Cserép (2010: 107), which refers to cases where both the adjective and the noun individually contribute to the meaning of the idiom, does presuppose that the nominal head is semantically autonomous.

In light of these, my prediction was that only those five idioms would be compatible with the intermediate level and literal-and-figurative types of adjectival modification whose nouns could be assigned an independent abstract meaning. Two expressions, lay one’s hands on something and ráteszi a kezét valamire behaved according to the expectations. As could be seen in (24) and (25), both of them occurred with the adjectives dirty and mocskos ‘filthy’, which internally modified their head nouns in a way that they also contributed their own figurative senses to the meaning of the string. Although the other three transparent expressions had no instances in the corpora with either the literal-and-figurative or the intermediate class of premodifiers, it does not necessarily mean that they are not open to such adnominals.

A lot more interesting and surprising finding of the current study is that the Hungarian idiom otthagyja a fogát ‘to die’ occurred with the intermediate level adjective becses ‘precious’, as has been shown in (20). Since the image evoked by its literal meaning (i.e. leaving your tooth somewhere) seems to have nothing to do with the idiomatic interpretation ‘to die’, and no correspondences can be detected between the meaning and the idiom components, this expression – at least in principle – should be considered as semantically opaque, hence, unable to combine with intermediate premodifiers. In this particular case, however, the speaker managed to remotivate the idiom, as a result of which, the constituent fog ‘tooth’ received some degree of semantic autonomy. That is, possibly on the basis of the meaning ‘to die’, he analyzed the individual idiom parts in accordance with another death-related expression életét veszti (‘lose one’s life’). Consequently, the verb otthagy ‘leave something there’ corresponds to veszt ‘lose’, while the noun fog ‘tooth’ is assigned the abstract meaning élet ‘life’, as illustrated in (34). Since the adjective becses ‘precious’ is a regular collocation of élet ‘life’, it operates in (20) only at the figurative level, modifying the ad hoc abstract meaning ‘life’ of the constituent fog.

Although it is not possible to draw general conclusions on the basis of one example, it should be noted that the above-discussed case supports Stathi’s (2007: 105) assumption that “idioms of the kick-the-bucket-type [i. e. opaque/non-decomposable idioms] may be treated as analytical sequences by the speaker”. Following from this, a very interesting question arises. Do opaque idioms really show no degree of decomposability? I agree with Cserép (2010: 111) in that in lack of sufficient psycholinguistic testing, this possibility should not be excluded yet. Some research focusing on this particular issue (also in light of other syntactic processes) may bring us closer to the answer.

4 Conclusion

To sum up, premodification seems to be a very common phenomenon in verbal idioms with body-part constituents. Although the majority of the adjectives belonged to the conjunction category, all of the higher-level types of Stathi’s (2007) hierarchy were represented in our research data. In addition to the five classes of adjectival modifiers proposed by Stathi (2007), however, we also argued for two other categories. On the one hand, there have been adjectives that did not contribute to the proposition at either level of interpretation but functioned as tools to express the speaker’s attitude and feelings about the event. Following McClure (2011), this class of premodifiers has been labeled as expressives. On the other hand, some evidence has been found in favour of Cserép’s (2010) literal-and-figurative type of modification as well, which is applied to cases where both the adjective and its head noun can be assigned their own figurative senses. These two categories are missing from Stathi’s (2007) taxonomy. With respect to semantic autonomy, it has been found that it may be possible in certain cases to assign ad hoc abstract meanings to the constituents of opaque idioms.

Corpus

Davies, Mark (2013): Corpus of Global Web-Based English: 1.9 billion words from speakers in 20 countries (GloWbE). https://corpus.byu.edu/glowbe/ [03.10.2017].

Oravecz, Csaba/Váradi, Tamás/Sass, Bálint (2014): “The Hungarian Gigaword Corpus”. In Proceedings of LREC 2014. 1719-1723. http://mnsz.nytud.hu/ [03.10.2017].

References

Bárczi, Géza et al. (1959): A magyar nyelv értelmező szótára [‘The Hungarian Language’s Explanatory Dictionary’]. http://mek.oszk.hu/adatbazis/magyar-nyelv-ertelmezo-szotara/elolap.php [03.10.2017].

Bárdosi, Vilmos (2003): Magyar szólástár [‘Hungarian Idiom Collection’]. Budapest: Tinta Könyvkiadó.

Burger, Harald (2015): Phraseologie: Eine Einführung am Beispiel des Deutschen. 5. Aufl. Berlin: Schmidt.

Cserép, Attila (2010): “Premodification in idioms”. Argumentum 6: 100–112.

Dobrovol’skij, Dimitrij (2000): „Zu semantischen und pragmatischen Effekten kreativer Idiom-Modifikationen“. In: Gréciano, Gertrud (ed.): Micro et Macrolexèmes et leur figement discursif. Actes du colloque international CNRS URA 10-35 (Langue-Discours-Cognition) 6-7-8 décembre 1998 Saverne. Louvain/Paris, Peeters: 217–230.

Ernst, Thomas (1980): “Grist for the linguistic mill: Idioms and ‘extra’ adjectives”. Journal of Linguistic Research 1/3: 51–68.

Fraser, Bruce (1970): “Idioms within a transformational grammar”. Foundations of Language 6/1: 22–42.

Hornby, Albert Sidney (2005): Oxford Advanced Learner’s Dictionary of Current English. Oxford: Oxford University Press.

Langlotz, Andreas (2006a): “Occasional adnominal idiom modification – A cognitive linguistic approach”. International Journal of English Studies 6/1: 85–108.

Langlotz, Andreas (2006b): Idiomatic Creativity. A Cognitive-Linguistic Model of Idiom-representation and Idiom-variation in English. Amsterdam/Philadelphia: Benjamins.

Minugh, David C. (2007): “The filling in the sandwich: internal modification of idioms”. In: Fachinetti, Roberta (ed.): Corpus Linguistics 25 Years on. Amsterdam, Rodopi: 205–224.

Minugh, David C. (2014): Studies in Corpora and Idioms: Getting the Cat out of the Bag. Stockholm: Universitetsservice AB.

Moon, Rosamund (1998): Fixed Expressions and Idioms in English: A Corpus-Based Approach. Oxford: Clarendon Press.

Nagy, Gábor O. (1985): Magyar szólások és közmondások [‘Hungarian Idioms and Proverbs’]. Budapest: Gondolat.

Nicolas, Tim (1995): “Semantics of idiom modification”. In: Everaert, Martin et al. (eds.): Idioms: Structural and Psychological Perspectives. Hillsdale/NJ, Lawrence Erlbaum Associates: 233–252.

Nunberg, Geoffrey/Sag, Ivan Andrew/Wasow, Thomas (1994): “Idioms”. Language 70/3: 491–538.

Quirk, Randolph et al. (1985): A Comprehensive Grammar of The English Language. London/New York: Longman.

Sabban, Annette (2000): „Okkasionelle Variationen von Phrasemen im Spannungsfeld zwischen Zeichenbeschaffenheit und Kontextbezogenheit“. In: Greciano, Gertrud (ed.): Micro et Macrolexèmes et leur figement discursif. Actes du colloque international CNRS URA 10-35 (Langue-Discours- Cognition) 6-7-8 décembre 1998 Saverne. Louvain/Paris, Peeters: 201–215.

Schenk, André (1995): “The syntactic behaviour of idioms”. In: Everaert, Martin et al. (eds.): Idioms: Structural and psychological perspectives. Hillsdale/NJ, Lawrence Erlbaum Associates: 253–271.

Siefring, Judith (2004): The Oxford Dictionary of Idioms. Oxford: Oxford University Press.

Spears, Richard A. (2000): NTC’s American Idioms Dictionary. New York: McGraw-Hill.

Stathi, Katerina (2007): “A corpus-based analysis of adjectival modification in German idioms”. In: Fellbaum, Christiane (ed.): Idioms and Collocations: Corpus-based Linguistic and Lexicographic Studies. London/New York, Continuum: 81–108.

Appendix

The twenty idioms investigated in our research are listed below. For the Hungarian expressions, a morpheme-by-morpheme, an interlinear transcription as well as a translation into English are provide.

el-veszt-i a fej-é-t

away-lose-3SG the head-POSS.3SG-ACC

’to lose one’s head/to lose self-control’

bele-üt-i az orr-á-t valami-be

into-hit-3SG the nose-POSS.3SG-ACC something-into

’to poke one’s nose into something/to interfere with something’

rá-tesz-i a kez-é-t valami-re

onto-put-3SG the hand-POSS.3SG-ACC something-onto

’to lay one’s hands on something/to obtain something violently/unethically’

fel-nyit-ja valaki szem-é-t valami-re

up-open-3SG someone eye-POSS.3SG-ACC something-onto

’to open someone’s eye to something/to cause someone to be aware of something’

nyal-ja valaki segg-é-t

lick-3SG someone ass-POSS.3SG-ACC

’to kiss someone’s ass/to flatter someone’

ott-hagy-ja a fog-á-t

there-leave-3SG the tooth-POSS.3SG-ACC

’to die’

be-ad-ja a derek-á-t

in-give-3SG the waist-POSS.3SG-ACC

’to bow to something’

mereszt-i a segg-é-t

stiffen-3SG the ass-POSS.3SG-ACC

’to idle’

lógat-ja a láb-á-t

hang-3SG the leg-POSS.3SG-ACC

’to idle’

össze-akaszt-ja a bajsz-á-t

together-hang-3SG the moustache-POSS.3SG-ACC

’to fall out with someone’

English idioms

lose one’s head

’to lose self-control’

poke one’s nose in(to) something

’to interfere with something’

lay one’s hands on something

’to obtain something’

open someone’s eyes

’to cause someone to be aware of something’

kiss someone’s ass

’to flatter someone’

cool one’s heels

’to wait’

dig one’s heels in

’to refuse to alter one’s course of action or opinions’

pull someone’s leg

’to fool, trick someone’

grease someone’s palm

’to bribe someone’

find one’s feet

’to gain confidence and experience in a new situation’

Notes

¹ As Langlotz (2006a: 97) points out, this phenomenon is not the exclusive peculiarity of idiomatic expressions. External idiom premodifiers equal with a subclass of noninherent adjectives, called adverbial adjectives (Quirk et al. 1985), which are used in the attributive position syntactically, but can be interpreted as adverbials semantically (e. g. my former friend = Formerly, he was my friend). back

² Minugh (2014) uses the concept of “textual anchorings”. back

³ Schenk (1995), Nicolas (1995), and Langlotz (2006a; 2006b) consider conjunction modification as a form of wordplay. Langlotz (2006b: 213), for example, argues for this claim as followings: “This form of variation is highly conspicuous, fully context-dependent and non-predictable. Moreover, it communicates its own creativity. For these reasons it is different from inconspicuous literal-scene manipulation, which systematically exploits the idiom’s inherent figurative potential irrespective of the usage-context”. back

⁴ It is based on the table that can be found in Stathi (2007: 103). back

⁵ The translations of German examples in this paper are adopted from Stathi (2007). back

⁶ The list of both the Hungarian and English idioms (with morpheme-by-morpheme interpretations for the Hungarian expressions) can be found in the Appendix. back

⁷ It does not mean at all that they are not compatible with conjunction modification. By analogy with the conjunction adnominals occurring with the other expressions, I can imagine examples like elvesztette vén fejét ‘lost his old head’ or otthagyta a lyukas fogát ‘lost his decayed tooth there’. back

⁸ It is interesting to see how the premodifier changed the dictionary meaning of this idiom. Since we are not talking about the real death of Primakov, but the end of his political career, it would be more reasonable to interpret the idiom as ‘to fail’ rather than as ‘to die’. back

⁹ The figurative interpretation of the part amely felnyitotta a központi szemeket is ‘which made the people in the centre realize the truth’ (lit. ‘which opened the central eyes’). back

¹⁰ I can imagine the use of the phrase becses fog (‘precious tooth’) only in ironical or very special contexts such as referring to a tooth that is honoured as a relic. back

¹¹ Some frequency numbers from the Hungarian National Corpus are: becses élet: 14; drága élet: 49; értékes élet: 31; tiszteletre méltó élet: 8. back