1 Introduction

This paper aims at describing a Hungarian construction which has been quite disregarded in the literature so far. We intend to do so in the spirit of the authors' teacher, László Kálmán: following Kálmán (2001a), we give a non-transformational analysis, presuming that the linguistic structure of an utterance can be discussed in terms of overlapping constructions. The construction in the focus of this investigation provides an especially good example of those linguistic expressions whose characteristics are not derivable from the characteristics of their constitutive elements. We also follow Kálmán's principle according to which the number of theoretical assumptions must be kept at the minimum in linguistic descriptions. Undoubtedly, some assumptions are inevitable – e.g. the presumption that some kind of categories, labeled as Adjectives and Nouns, exist.¹ We intend to describe the given construction similarly to the descriptions of various patterns in Kálmán (2001b).

Kálmán always stressed that linguistic research must be empirically grounded. In light of this advice, the authors feel it important to indicate the character of the linguistic examples in a more sophisticated way than is usual in the literature. Examples taken from the corpus (see below) or already published in the literature and considered as completely grammatical by the authors of the current article are unmarked. (These examples are sometimes simplified, but the authors consider that these modifications do not influence grammaticality.) Examples constructed by the authors for demonstration of agrammatical word sequences are marked by an asterisk (*). Those constructed examples whose grammaticality is uncertain or debated are marked with a question mark (?), which is sometimes doubled to indicate a higher degree of questionability. Grammatical but semantically or pragmatically odd examples are marked with a hash (#). Grammatical examples constructed by the authors are marked with a wrench (). Finally, examples taken from the corpus or the literature whose grammaticality is questioned by the authors is marked by an inverted question mark (¿).

The construction under investigation consists of an adjective, the word egy, and a noun.

(1)

furcsa	szerkezet
strange	construction
‘strange construction’

(2)

egy	furcsa	szerkezet
art.indef/one	strange	construction
‘a strange construction’

(3)

furcsa	egy	szerkezet
strange	emph	construction
‘a really strange construction’

The “A egy N” exemplified in (3) belongs to the constructions Kálmán (2001a, 70–77) calls syntactic idioms: they require certain well-defined syntactic structures. In certain respects, it resembles lexical idioms as well (Kálmán 2001a, 68–70), as it is partially “filled in”, that is, a lexical slot in it is not variable (i.e. the word egy). However, the typical examples of lexical idioms are words with complements or existential sentences.

The meaning of this construction resembles, on the one hand, that of a noun modified by an adjective (1), and on the other hand, that of an indefinite noun phrase (2). Apparently, the function of egy in (2) and in (3) is somewhat different. By default egy is the numeral ‘one’ in Hungarian. In (2) egy can be replaced by other numerals,² while in (3) it cannot. This could be an argument for the supposition that in (3) egy must be identified with the indefinite article ‘a, an’ (developed from the numeral). Nonetheless, a definite article a/az cannot occur in the same position of a noun phrase (*furcsa a szerkezet³). Additionally, by default, when a noun is modified by both an adjective (or several ones) and an article, the adjective is preceded by the article, as in (2). Thus it seems that the function of egy in (3) resembles that of an emphatic clitic; therefore, in what follows, we gloss it as emph. Despite the glossing, the authors do not want to state that the emphatic nature of the construction emerges from the clitic: it is rather a peculiarity emerging from the whole construction.

(4)

Ez	furcsa	szerkezet.
this	strange	construction
‘This is a strange construction.’

(5)

Ez	a	furcsa	szerkezet.
this	art.def	strange	construction
‘This is the strange construction.’

(6)

Ez	egy	furcsa	szerkezet.
this	art.indf/one	strange	construction
‘This is a/one strange construction.’

(7)

Ez	furcsa	egy	szerkezet.
this	strange	emph	construction
‘This is a really strange construction.’

As examples (4) and (6) show, there is no significant difference between the meaning of a construction without an article and one with an indefinite article. However, the two constructions with egy ((6)–(7)) considerably differ from each other. As a consequence, it is reasonable to consider that the words egy in the two instances are distinct homophones synchronically (belonging to the same etymon, as well as the numeral ‘one’). However, it seems that the emphatic clitic cannot be used with articles, it is not completely detached from the indefinite article ((8)–(9)).

The “A egy N” construction is not thoroughly analyzed in the literature. Traditional academic and university grammars solely mention the phenomenon that the indefinite article can stand between the adjectival modifier and the noun when the adjective is stressed, but they do not go further (Imre & Temesi 1962, 280; Velcsov 1968, 74; Rácz 1968, 341, 343). Another reason proposed in the literature for the emergence of egy in this position is simply emotional emphasis (Kugler 2000, 287). However, Deme (1962, 473) ignores even the possibility of this word order. The prescriptive literature does not show much interest in the phenomenon: Elekfi & Grétsy (1980, 463) remark that the indefinite article is completely unstressed when it stands between the modifier and the noun, but otherwise, they ignore the construction. According to Jakab (1998, 144), putting the indefinite article between the modifier and the modified for emphasis is “magyaros”, that is, typical Hungarian. Jakab (2000, 48–49) states that the construction can be applied even when the use of the indefinite article is incorrect. According to him, while (10) is correct, (11) is not. In fact, sentences like (11) are widely used in Hungarian, but stigmatised and labeled as “Germanisms” by prescriptive linguists.⁴

In early generative approaches to Hungarian, such as Szabolcsi & Laczkó (1992), Kenesei, Vago & Fenyvesi (1998), or É. Kiss (2002) we could not find instances of the “A egy N” construction. However, more recent comprehensive generative studies on Hungarian do register it: Dékány (2011) gives an instance of the construction under discussion, but she does not provide an analysis treating it as “a special construction that does not represent garden variety DPs in Hungarian” (p. 34). The newest comprehensive grammar of Hungarian, Farkas & Alberti (2018) also gives an instance of the “A egy N” construction, stating that it “is only allowed in order to trigger some emphatic or funny effect” (p. 126). The same volume contains Szabó’s analysis of the use of the “A egy N” construction in exclamatives (Szabó 2018, 186–187). Both Dékány (2011) and Szabó (2018) refer back to den Dikken & Lipták (1997) as an already existing analysis of the structure. Den Dikken & Lipták (1997) provide a generative approach based on cross-linguistic data mainly from Germanic languages, analyzing the “A egy N” construction as a nominal-internal predication. While their approach is theoretical in nature, deriving comparative data, our investigation is rather descriptive, as atheoretical as possible, and restricted to Hungarian. We focus on the attested uses of the construction, relying on corpus data, and treating linguistic phenomena as gradual if possible, not as binary (Boolean).

The overall aim of the present study is the detailed and precise description of the “A egy N” construction from morphological, syntactic, and semantico-pragmatic aspects, based on linguistic data extracted from the Hungarian Gigaword Corpus (Oravecz, Váradi & Sass 2014). Such a data-based description can contribute to the clarification of misleading overgeneralizations/explanations on the construction under discussion, and to a better understanding of how the levels of morphology, syntax, and pragmatics interact in constructions.

In Section 2, a methodological description is sketched. In Section 3, constructions having similar structures (that is, non-emphatic “A egy N”) are presented, and it is explored how they can be distinguished from the construction with emphatic egy. In Section 4, the lexical inventory of words used in the given construction, and the purposes of their use are examined. In Section 5, the morphosyntactic features of the construction are overviewed, specifically what kind of morphemes can occur inside it and where, how complex its structure can be, and where it can occur inside a sentence. In Section 6, similar constructions in which we find a noun instead of an adjective are addressed. In Section 7, the important findings are summarised.

2 Methodological remarks

This study was conducted in a corpus-driven manner (Tognini-Bonelli 2001, Chapter 5), that is, we investigated available corpus data without any preconception or preliminary hypothesis. To have an overall picture of the “A egy N” construction, we made a significant effort to select all available data from a de facto standard, balanced Hungarian corpus: the Hungarian Gigaword Corpus (Oravecz, Váradi & Sass 2014), version 2.0.5. We excluded the social media subcorpus, as it often contains multiple copies of the very same texts.

In order to get preferably all real corpus examples for the use of the construction with emphatic egy, a corpus search was carried out for word sequences in which an adjective is immediately followed by the word egy and a noun in a corpus.

In other words, we formulated a deliberately broad query which covers all examples we need with a good chance at the price that the result may contain many irrelevant hits. This is the usual approach to follow when all data for a phenomenon is to be collected, not just some random examples (Sass 2022, principle #2). Clearly, this approach often requires a large amount of manual work to filter out irrelevant hits. Accordingly, the following CQL (Corpus Query Language) query was run:

[msd="MN\.NOM"] [word="egy"] [msd="FN.*"]

To make this CQL query more understandable, we break it down into parts. A condition of the form X="Y" surrounded by square brackets represents a word. In the case of the first and third word, we refer to the morphosyntactic description (msd) of the word, while we refer simply to the surface form (word) in the case of the second one. MN is the code for adjectives, FN is for nouns. MN\.NOM means an adjective in nominative case, i.e. without any suffixes and FN.* means a noun with any possible combination of suffixes.⁵

Runnig the above query resulted in 28,232 raw corpus hits, which we processed partly manually and partly automatically in the following way to obtain the dataset to base our study on.

To reduce the amount of manual work needed, firstly, we investigated the 17,719-element raw frequency list of the constructions without context and divided this list into two groups manually in an intuitive manner based solely on the construction itself: “A” containing possible good examples and “B” containing word sequences having very little chance of being a good example.

Secondly, we queried all hits for constructions for both lists automatically together with context and checked the 5,195-element list coming from group “A” one-by-one and, additionally, a small sample from the much longer list coming from group “B” as well to collect good examples. This is how the final, curated 1158-element dataset came about.

The dataset is made available for investigation and further research at https://github.com/sassbalint/strange-a-construction. Basic statistics of the dataset are presented in Section 4.1 after reviewing the formally similar constructions in Section 3 which were excluded from the scope of our study.

Undoubtedly, the linguist's judgement on the data is always subjective to some extent. Corpus data help researchers to face linguistic facts they did not think of, e.g. constructions which are similar on the surface and can be different semantically or pragmatically (cf. Section 3). Additionally, corpus data can draw attention to marginal cases, which otherwise could be thought impossible by the researchers (see Section 5.1.1). We intend to be cautious with labelling such marginal cases, as they may be results of performance errors, reflections of individual or dialectal features, or typical for certain social groups.

Moreover, the exact interpretation of an expression, even in a wider context, can always be questioned. It is not always straightforward whether a statement is meant ironically or not, whether an assertion is intended to be praise or disapproval. (It is also a real problem in everyday communication, and sometimes ambiguity is intentional.)⁶

3 Similar constructions

Apparently, “A egy N” sequences may occur in various constructions, such as a predicate and a subject (12), or by chance, not forming a construction on their own ((13)–(14)).

There are cases when it is very difficult to identify the construction without the knowledge of the wider context and/or the intonation pattern. This is especially true for sentences that can be interpreted in two ways. (1) if the noun is the subject, or (2) if the subject is omitted, and the “A egy N” sequence is a manifestation of the construction described in this study, i.e. a phrase containing emphatic egy ((15)–(17)).⁷

If the sequence did not represent unambiguously the construction we were interested in, we excluded it. Nonetheless, it is possible that in some cases the text sample given by the search engine of the corpus (five words both before and after the “A egy N” sequence) suggested that the sequence represents the construction we are looking for, although it could have been clear from a wider context that in fact, it did not.

In addition to the predicate–subject structure, there are cases in which the “A egy N” sequence forms a noun phrase but a different one. Two main types can be distinguished. In one of them, egy is the numeral ‘one’ (3.1), in the other one, egy is an adjective ‘sole, only’ (3.2). However, the distinction is mostly problematic when we work with written texts: as the emphatic clitic egy is never stressed, while numerals and adjectives usually are, the construction under investigation could more easily be identified in spoken texts, but such a study goes beyond the scope of our current work.

In marginal cases, the “A egy N” sequence, in all probability, represented the construction we are interested in, yet was ignored. For example, the most frequent sequence was kedves egy lány ‘really nice girl’, but it always occurred as a nickname in an internet forum. As nicknames are much more typically noun phrases than complete sentences, and there seems to be no reason to interpret egy here as a numeral or adjective, it is most probably the emphatic clitic. Nonetheless, such cases were ignored, as they do not represent the typical use of the construction. Additionally, taking them into consideration would distort the statistical data on the frequency of adjectives in the construction with emphatic egy.

3.1 egy as a numeral

In constructions in which egy is a numeral, it can be replaced by other numerals as két ‘two’, három ‘three’, fél ‘half’, etc. Although the denotation of the construction varies according to the numeral, its function does not change: it expresses measure (cf. Footnote 2). In these constructions, nouns usually express time (perc ‘minute’, óra ‘hour’, nap ‘day’, hét ‘week’, hónap ‘month’, év ‘year’, évtized ‘decade’), distance ((centi/kilo)méter ‘(centi/kilo)meter’), volume (deci ‘decilitre’, liter ‘litre’), currency (korona ‘crown’, dollár ‘dollar’, schilling ‘schilling’) or sporadically some other words which can be used as units (megapixel ‘megapixel’, mondat ‘sentence’, gombóc ‘scoop (of ice cream)’, pont ‘score, point (e.g. in games)’).

Such constructions can be divided into two subgroups depending on whether the numeral and the adjective can be reordered without a considerable change in meaning or not. The difference is less salient with sequences containing egy ‘one’, and much clearer with phrases containing some other numerals. In examples (18)–(20), the time of two months or three hours is said to be short or long, and the amount of hundred forints is said to be symbolic.

(18)

rövid	két	hónap
short	two	month
‘short two months’

(19)

hosszú	három	óra
long	three	hour
‘long three hours’

(20)

szimbolikus	száz	forint
symbolic	hundred	forint
‘symbolic one hundred forints’

Placing the adjective before the numeral would be understood to mean that the months, the hours, and the forints are short, long, and symbolic distributively, i.e. the adjective refers to each unit individually. It is especially strange with adjectives meaning smallness or negligibility combined with a relatively high number. Example (21) is semantically odd, because even if the individual months are felt short, it does not count when there are so many of them. Speaking about a currency as in (22), the construction may be used when the currency was withdrawn from circulation or completely depreciates due to hyperinflation (in progress).

(21)

?száz	rövid	hónap
hundred	short	month
‘one hundred short months’

(22)

egymilliárd	szimbolikus	forint
one.billion	symbolic	forint
‘one billion symbolic forints’

Some adjectives are special in similar constructions. The adjective jó ‘good’ means rather ‘about, at least’, or even ‘a bit more’ when it is put before the numeral ((23)–(24)).

(23)

Jó	egy	percet	töltöttem	ott.
good	one	minute	spend;pst.1sg	there
‘I spent there at least one minute.’

(24)

Egy	jó	percet	töltöttem	ott.
one	good	minute	spend;pst.1sg	there
‘I spent there one good minute.’

Similarly to jó, bő ‘loose-fitting, abundant’ can be used meaning ‘a bit more’ and szűk ‘fitting, narrow’ as ‘a bit less’. Their interpretation does not vary with the word order, but they sound odd when preceded by any other numeral than egy ((25)–(26)).

(25)

?Két	bő	órát	töltöttem	ott.
two	abundant	hour;acc	spend;pst.1sg	there
‘I spent there more than two hours.’

(26)

?Három	szűk	perc	alatt	kész	lett.
three	tight	minute	in	ready	become.pst.3sg
‘It was ready in not more than three minutes.’

The situation is similar with relational adjectives, but in these cases, the difference is quite salient in sequences with egy as well. Constructions like (27) are used when one girl is expected from each city, and we are talking about the one from Budapest; more generally speaking: when the given number of things, people, etc. have already been introduced into the discussion. This type of construction with egy is semantically quite close to the one discussed in Section 3.2, but egy can be replaced by other numerals. On the contrary, constructions like (28) denote a given amount of girls who are from Budapest.

(27)

budapesti	egy	lány
Budapest;adjz	one	girl
‘one girl from Budapest’

(28)

egy	budapesti	lány
one	Budapest;adjz	girl
‘one girl from Budapest’

However, the change in the word order does not cause radical semantic difference when the adjective denotes quality (e.g. (29)–(30)), although the “A Num N” construction suggests that A characterises the Ns as a whole while the “Num A N” construction suggests that A characterises each N individually. In any case, these problems are out of the scope of our study.

(29)

izgalmas	két	hét
exciting	two	week
‘exciting two weeks’

(30)

szerencsétlen	három	kutya
unfortunate	three	dog
‘unfortunate three dogs’

These kinds of constructions with egy are typical only with words presented at the beginning of this section. These words usually occur as arguments ((31)–(32)), but rarely also as classifiers (33).

(31)

A	pályaudvaron	kemény	egy	schillinget
art.def	railway.station;ADE	hard	one	schilling
költöttem	vécére.
spend;pst.1sg	toilet;sub
‘I spent one precious schilling on the toilet at the railway station.’

(32)

Csak	ezzel	a	szerencsétlen	egy	témával
only	this;ins	adj.def	wretched	one	subject
foglalkozik
deal.with;prs.3sg
lt. ‘S/he is engaged only in this one wretched subject.’8

(33)

Az	ajtó	is	tetszett	neki:	szép	egy	darab
art.def	door	also	be.liked;pst.3sg	3sg.dat	beautiful	one	piece
keményfa.
hardwood
‘S/he also liked the door: (it was) a beautiful single piece of hardwood.’

These kinds of structures are possible when the adjective can also be interpreted collectively for the given quantity of entities as a unit. In the opposite case, the “A NUM N” sequence cannot be interpreted as one construction ((34)–(36)).

(34)

*apró	ezer	tárgy
tiny	thousand	thing
‘thousand tiny things’

(35)

*általános	három	iskola
general	three	school
‘three elementary schools’

(36)

*háztartási	nyolc	keksz
household;adjz	eight	biscuit
‘eight pieces/kinds of budget biscuits’

3.2 egy as an adjective

As it was mentioned above (3.1, 11), the “A egy N” in which egy is an adjective is very close to some examples in which egy is a numeral. Examples (31)–(33) can be interpreted as sentences in which egy is synonymous with the adjectives egyedüli, egyetlen ‘sole, single, only’. In such sentences, egy can be replaced with other numerals. However, there are certain cases when this is impossible or highly atypical ((37)–(40)).

(37)

Jó,	hogy	jössz,	édes	egy	komám.
good	that	come;prs.2sg	sweet	one	mate-poss.1sg
‘It is good that you come, my dear only mate.’

(38)

Kossuth	Ferenc	a	nagy	Kossuth	Lajos	édes	egy
Kossuth	Ferenc	art.def	great	Kossuth	Lajos	sweet	one
fia.
son;poss.3sg
‘Ferenc Kossuth is the beloved sole son of the great Lajos Kossuth.’

(39)

Megvédelmezem	édes	egy	hazánkat.
defend;prs.1sg	sweet	one	homeland;poss.1pl.acc
‘I (will) defend our beloved homeland.’

(40)

A	termékekkel	semmi	baj	nincs	az
art.def	product;pl.ins	nothing	trouble	neg.be.prS.3sg	art.def
égadta	egy	világon.
heaven.given	one	world.sup
‘There is no problem with the products at all.’

It is noteworthy, that in similar structures with egyetlen the order of the two adjectives is reversible ((41)–(42)), although the two strings are semantically different: when egyetlen is in the first place, the construction does not mean anymore that there is only one entity (e.g. mate), which is additionally of the given quality (e.g. dear), but that there is only one entity of the given quality (and there can be entities without the quality, e.g. mates who are not dear). Nonetheless, the adjective egy ‘sole, single, only’ can only be used when it is preceded by another adjective, that is, a noun modified by the adjective egy always has to be modified by another adjective as well.

(41)

édes	egyetlen	komám
sweet	sole	mate;poss.1sg
‘my dear only mate’

(42)

egyetlen	édes	komám
sole	sweet	mate;poss.1sg
‘my only dear mate’

The phrase égadta egy világon (40) is idiomatic, its idiomaticity certainly played a role in the development of its variant égegyadta világon, morph-by-morph ‘heaven-one-given world-sup’. The morphological structure of the latter seems to be unique in Hungarian, although, according to Google search, this latter variant is about twice as frequent as its original.

4 Lexicon, semantics, pragmatics

The set of adjectives that can and actually do occur in the construction under discussion is strictly determined by the semantics and pragmatics of the construction. As the construction denotes a subjective derogatory (or, less frequently, appreciative) value judgement based on the high degree of a quality, the adjective must express a gradual quality. In Section 4.1, the words that can be and actually are present as an adjective in the construction are discussed. In Section 4.2, the ways of interpreting the construction are presented.

4.1 Authorised entry only

As for the “A egy N” construction with the emphatic clitic, the set of adjectives occurring in it is restricted. This fact is strongly related to the semantics and the pragmatics of the construction. As the construction expresses a high degree of the quality expressed by the adjective, the adjective has to indicate a gradual quality. The construction also expresses subjective value judgement and strong personal commitment to it, thus, one can expect that it can hardly be combined with talán ‘perhaps, maybe’. According to our interpretation, in examples like (43), talán does not express the uncertainty of the speaker but the opposite: it is irony over the possibility that someone can question the importance of the new data.

(43)

Talán	nem	érdektelen	egy	adalék	az
perhaps	neg	uninteresting	emph	complementing.data	art.det
irodalmi	utókornak.
literature;adjz	posterity;dat
‘It might be a really interesting additional piece of information for the posterity interested in literature.’

The use of the “A egy N” construction expresses the speaker's attitude: his or her surprise, outrage, anger, or delight towards the degree of the quality denoted by the adjective.

We could collect 1,158 realisations of the construction in our gigaword corpus, which suggests that it is relatively rare. In these examples, 394 different adjectives are used. The most frequent one of these, furcsa appeared 120 times, which is more than 10% of the cases. Thirteen adjectives occurred in more than 1% of the cases (Table 1): one of these thirteen adjectives occurs in 37% of all the attested cases.

Table 1.

The most frequent adjectives in the emphatic “A egy N” construction

Adjective	Meaning	Tokens	Proportion (%)
furcsa	strange, odd	120	10.3
érdekes	interesting	61	5.2
fura	strange, odd	46	3.9
különös	strange, odd	32	2.7
hülye	stupid, idiotic	30	2.5
ronda	ugly	26	2.2
szép	beautiful	19	1.6
csodálatos	wonderful	15	1.2
gyönyörű	beautiful, magnificent	14	1.2
szörnyű	horrible, terrible	13	1.1
fantasztikus	fantastic	13	1.1
randa	ugly	12	1.0
csúnya	ugly	12	1.0

Additionally, most of the thirteen adjectives belong to one of three groups of synonyms. In almost 17% of the cases, the adjective means ‘strange, odd’ (furcsa, fura, különös). Moreover, érdekes ‘interesting’ can also be interpreted many times as ‘unusual, odd’ and some of the adjectives less frequently occurring in the construction also have a similar meaning: sajátos ‘peculiar’ (3), szokatlan ‘unusual, uncommon’ (2), különleges ‘special’ (2), etc. Even some other adjectives with a different primary meaning are used in this meaning. E.g. in (44), ritka ‘rare, infrequent’, generally occurring in the same meaning in the constructions with emphatic egy as well, can be understood only as ‘strange, unusual’.

(44)

Kozmánénak	hívták,	ritka	egy	alak	volt.
Kozmáné.dat	call;pst.3pl	rare	emph	figure	be.pst.3sg
‘She was called Mrs. Kozma, she was a really strange character.’

Another important semantic group contains adjectives meaning ‘ugly’, most frequently ronda, randa, csúnya, but also ocsmány (9), rusnya (4), csúf (2), etc., all meaning ‘ugly’ (although in some cases used rather as a negative moral than an aesthetic judgement). An additional considerable meaning is the opposite, ‘beautiful’, represented by szép and gyönyörű among the most frequent ones, and in some of the cases csodálatos ‘wonderful’ is also used as an aesthetic judgement.

However, all this does not mean that only a narrow set of adjectives can occur in the construction with emphatic egy. 229 adjectives occur just once in our material, and it is sure that there are many other adjectives which could easily have occurred and are only missing by chance. It is also clear that many frequent adjectives cannot occur in such a construction, and if they do occur, the construction must have a very specific reading. For example, adjectives denoting colours, nationality or ethnicity are quite frequent in the corpus, but in our material, they never occur as an adjective in the construction under discussion. However, this fact does not mean that it is impossible to use them in such cases as in the following constructed examples. (45) could be used when someone was specifically asked to buy a red apple, and s/he bought a green one.⁹ (46) could be used when someone was stated to have a French girlfriend, and it turns out that the girlfriend is not French at all. (Examples like (47) are exceptional because the name of the ethnic group denotes something else, e.g. a way of lifestyle.)

(45)

Piros	egy	almát	vettél!
red	emph	apple;acc	buy;pst.2sg.do
‘You bought a red apple indeed!’

(46)

Tényleg	francia	egy	barátnője	van!
really	French	emph	girlfriend;poss.3sg	be.prs.3sg
‘S/he really has a French girlfiend!’

(47)

Elég	cigány	egy	élet	ez.
fairly	gipsy	emph	life	this
‘It is a fairly carefree/nomadic/chaotic life.’

As colour and origin are not gradual qualities,¹⁰ these constructions can be interpreted ironically: the construction suggests a high degree of a quality which is not gradual at all. Such examples are really rare in our material. (48) is a noteworthy exception, the peculiarity of which is difficult to reflect in the English translation. The emphatic clitic co-occurs with kálvinista ‘Calvinist’, which is a non-gradual quality. Moreover, without a comma between ősi and kálvinista, ősi in the phrase must be understood as a subsective adjective (similarly to as it is defined in (Partee 1995, 324), however, here it is a subsective adjective related to another adjective, not to a noun), that is, the family is not simply ‘ancient and Calvinist’, but ‘Calvinist for a long time’.

(48)

Ősi	kálvinista	egy	család	sarja.
ancient	Calvinist	emph	family	descendant;poss.sg3
‘She is a descentant of a really ancient Calvinist family.’

The sentence comes from István Szilágyi's novel, Kő hull apadó kútba [Stone falling into receding well], and the context unambiguously shows that the sentence is somewhat intentionally ill-formed.

És bátya azt se felejtse el, hogy én egy ősi kálvinista család sarja vagyok. Ugye milyen jól hangzik. Hát csodálatos. Egy ősi kálvinista család sarja. Egy sarja ősi kálvinista család. Egy család ősi kálvinista sarja. Egy kálvinista ősi család sarja. Sarja egy ősi kálvinista család. Kálvinista sarja ősi egy család.

Te gyermek, megint ég a szemed.

Ősi kálvinista egy család sarja.

Ilka. Csillapodj, Ilka.

Család kálvinista ősi egy sarja.

And, uncle, do not forget either that I am a descendent of an ancient Calvinist family. It sounds good, doesn’t it? Really wonderful. The descendent of an ancient Calvinist family. Of an ancient Calvinist family descendent. An ancient Calvinist descendent of a family. The descendent of a Calvinist ancient family. Calvinist descendent of ancient a family.

You, my child, your eyes are daggering again.

A descentant of a really ancient Calvinist family.

Ilka, relax, Ilka.

A descendant of family Calvinist ancient.

The permutation in the word order is quite mechanical, and sometimes it results in ungrammatical or semantically strange sentences. In all probability, the sentence with the word order especially interesting for us is not standing in a featured position by chance. It emphasises the ancient Calvinism of the family, although the untypical use of the construction reflects the sarcasm of the character about her own pride in her origin.

Some non-gradual adjectives, e.g. nonsubsective ones (Partee 1995, 325), could hardly occur in the construction. Occasionally, examples (49)–(51) can be used even sarcastically, i.e. when the poet is alive, the director still fills his/her position and the reason is genuine (contrary to someone's statements).

(49)

*Néhai	egy	költő!
late	emph	poet
‘S/he is a really dead poet!’

(50)

*Ő	aztán	egykori	egy	igazgató!
s/he	then	one-time	emph	director
‘S/he is really a former director!’

(51)

*Ez	tényleg	állítólagos	egy	ok!
this	really	ostensible	emph	reason
‘This is a really ostensible reason in fact!’

As the adjective is always emphasised in a way, it is no wonder that in some cases more expressive adjectives are more frequent than their basic synonyms. E.g. the most general adjective rossz ‘bad’ occurs only once, while the synonymous szörnyű ‘horrible’, rémes ‘awful, terrible’, cudar ‘grim, foul’, pocsék ‘shoddy, atrocious’, szánalmas ‘pitiable; pathetic’, keserves ‘painful, troublesome’, ócska ‘rubbishy, paltry, shabby, dilapidated’, gyalázatos ‘shameful, disgraceful, atrocious’, vacak ‘shoddy, cheap’, rettenetes ‘terrible, horrible’ all occur at least four times. The adjective ronda ‘ugly, repulsive’ is more frequent than csúnya ‘ugly, unsightly’, hatalmas ‘huge, gigantic’ than nagy ‘big, large’, mocskos ‘scrotty’ than koszos ‘dirty’, etc. However, it is not always the case, e.g. the moderate szép ‘beautiful’ is a bit more frequent than the more suggestive equivalent gyönyörű. However, as more than 95% of the adjectives occur less than 10 times in the corpus, it is quite difficult to prove statistically that more expressive adjectives are preferred. Even in this case, we should compare their frequency in the construction with emphatic egy to their frequency in the whole corpus. Additionally, in many cases the semantic network is complicated, e.g. nehéz ‘heavy, hard, difficult’ can be a synonym of súlyos ‘heavy, weighty, serious’, bonyolult ‘complicated, difficult, tricky’, kemény ‘hard, tough’, etc. in different cases.

4.2 Deprecation and approbation

The construction tends to denote a negative value judgement in general. This tendency is reflected by the fact that adjectives occurring in the construction denote a negative attribute by default. According to our calculations made on our material,¹¹ about 40–45% of the adjectives used in the construction denote negative ethic or aesthetic qualities, stupidity, unluckiness, discomfort, or misery. On the contrary, only 20–27% can be considered as positive. All the other adjectives are neutral in the sense that either they do not reflect any positive or negative value, or they reflect some value (negative, positive, or both) only under given circumstances.

As language users usually come across the construction in a deprecating context, it is expected that they tend to interpret (and use) it as a negative judgement even when the adjective itself does not denote negative characteristics. The pragmatic interpretations of the positive and negative adjectives are different. Although the constructions with positive adjectives can be interpreted as praise and tribute, often they express sarcasm or irony. In cases like (52), only the intonation pattern would show whether the utterance expresses recognition or mockery.

(52)

Okos	egy	férje	van.
smart	emph	husband;poss.sg3	be.prs.3sg
‘S/he has a really smart husband’

However, the opposite is not true: a construction with a negative adjective can never be interpreted as an approbation (53).¹²

(53)

Hülye	egy	férje	van.
stupid	emph	husband;poss.sg3	be.prs.3sg
‘S/he has a really stupid husband’

This irreversibility of the judgement by sarcasm is a specificity of the given construction and not of negative adjectives. In other cases, negative adjectives can express appreciation in sarcastic use: (54), with a special intonation pattern, indeed, can be used against someone criticising Chomsky, indicating that Chomsky is smart, just the critic cannot admit it.

(54)

Mert	Chomsky	hülye.
because	Chomsky	stupid
‘Because Chomsky is stupid.’

5 Morphosyntax

This section presents the morphosyntactic characteristics of the construction. The morphological peculiarities are treated in Section 5.1, while the syntactic behavior is discussed in 5.2.

5.1 Morphology

The morphological features of the noun and the adjective in the construction are analyzed separately, in Section 5.1.1 and 5.1.2, respectively.

5.1.1 The morphology of the noun

In general, a Hungarian noun can take a plural marker, a possessive suffix, an anaphoric possessive suffix and a case suffix. The use of the case and possessive suffixes are not restricted, they are used as it is expected in the “A egy N” environment ((55)–(59)). However, both categories of suffixes are relatively rare in constructions with emphatic egy. Their combination is extremely rare, although completely grammatical instances can be constructed (60).

(55)

Szép	egy	ordast	lőttél.
beatiful	emph	wolf;acc	shoot;pst.2sg
‘You shot a really beautiful wolf.’

(56)

Csúnya	egy	disznóságnak	tűnik	nekem	ez	az
ugly	emph	shame;dat	seem;prs.3sg	1sg.dat	this	art.def
egész.
whole
‘That all seems like an ugly dirty trick for me.’

(57)

Elég	büdös	egy	csizmáról	van	szó.
fairly	stinky	emph	boot;del	be.prs.3sg	word
‘It is about a quite stinky boot.’

(58)

Szörnyű	egy	éjszakám	volt.
terrible	emph	night;poss.1sg	be.prt.3sg
‘I had a really terrible night.’

(59)

Keserves	egy	napunk	volt.
troublesome	emph	day;poss.1pl	be.prt.3sg
‘We had a really troublesome day.’

(60)

Keserves	egy	napunkról	kell	beszámolnom.
troublesome	emph	day;poss.1pl;del	have.to.prs.3sg	report;inf;1sg
‘I have to report about a really troublesome day of ours.’

Similarly, there are no examples for the use of the anaphoric possessive suffix -é in the “A egy N” environment. As for the authors' intuition, it can be used indeed as the nominal part of the predicate (61), but it cannot be combined with a case suffix (62).

(61)

Hülye	egy	tanáré	lehet	ez	a	példa.
stupid	emph	teacher;anp	be.mod.prs.3sg	this	art.def	example
‘This example must be a really stupid teacher's (one).’

(62)

*Hülye	egy	tanárét	hallottam.
stupid	emph	teacher;anp	hear;pst.1sg
‘I heard a really stupid teacher's (one).’ or
‘I heard a teacher's really stupid one.’

The question of the use of the plural marker in the “A egy N” construction is more complex. Although den Dikken & Lipták (1997, 63, 68, 70–72) discuss constructions in which the noun is plural, e.g. (63), according to the authors' native intuition, these constructions are ungrammatical or at least extremely odd.

(63)

¿Véres	egy	napokat	élünk.
bloody	emph	day;pl.acc	live;prs.1pl
‘We live really bloody days.’

In our corpus material, we can find only a handful of examples with plural nouns ((64)–(70)). Their low number may suggest that they are just slips of the pen. Additionally, some of them are odd for other reasons. In (64), the noun is written with a capital letter, which may suggest that it is a proper name (e.g. of a theatrical company or a band).¹³ In (68) netovább ‘ne plus ultra’ should be agreed in number with építmény ‘building’, but it is singular. In (70), egy may also be interpreted as a numeral, in that case, the sentence means that such dramatic sole hits occurred elsewhere as well.

(64)

¿Szép	egy	Bogarak	vagytok.
beautiful	emph	bug;pl	be.prs.2pl
‘You are really beautiful bugs.’

(65)

¿Érdekes	egy	szülemények	ezek.
interesting	emph	creature;pl	these
‘These are really interesting creatures.’

(66)

¿Nyomorult	egy	emberek	ezek.
miserable	emph	human;pl	these
‘These are really miserable people.’

(67)

¿Ügyes	egy	népek	vagyunk.
skilful	emph	people;pl	be.prs.1pl
‘We are really skilful peoples.’

(68)

¿Szörnyű	egy	építmények,	a	giccs	netovábbja.
terrible	emph	building;pl	art.def	kitsch	ne.plus.ultra;poss.s3
‘These are really terrible buildings, the pinnacle of kitsch.’

(69)

¿Furcsa	egy	szokások	voltak.
strange	emph	custom;pl	be.prt.3pl
‘They were really strange customs.’

(70)

¿Nem	csak	nálunk	voltak	ilyen	drámai	egy	tusok.
no	only	1pl.ade	be.pst.3pl	such	dramatic	emph	hit;pl
‘It was not only at our place where such dramatic hits occurred.’

We have to conclude that even if these examples are grammatical for anyone at all, they are very marginal. In all probability, plural is severely dispreferred in this construction for historical reasons: the emphatic clitic egy emerged from the indefinite article ‘a(n)’, which cannot be attached to plural nouns. As egy in the emphatic constructions does not denote singleness anymore, it becomes possible to use it with plural forms, and, maybe, the change in that direction has begun, but has not gone far yet.

Additionally, we have to mention that in some cases it seems that plural is avoided. In the constructed example (71), plural marking is more usual, and in that case, the sentence can be interpreted as ‘we are the specimens of a strange animal race’. However, singular marking is also possible, suggesting a meaning ‘we, as a community, belong to a strange animal race’. In the corpus example (72), singular is selected, and the plural form would sound quite strange in this construction. In the corpus example (73), agreement rules require a singular form of fajzat ‘breed’, but if kettő‘ ‘two’ is changed to ketten ‘two of them’ (74), a plural form must be used: in this case, the use of emphatic egy would be strange. In the corpus example (75), a motivation for using a collective noun can be that this way plural agreement is not necessary at all.

(71)

Furcsa	állat/állatok	vagyunk.
strange	animal/animal;pl	be.prs.1pl
‘We are strange animals.’

(72)

Furcsa	egy	állat	vagyunk.
strange	emph	animal	be.prs.1pl
‘We are really strange animals.’

(73)

Ocsmány	egy	fajzat	mind	a	kettő.
ugly	emph	breed	all	art.def	two
‘Both are ugly breeds.’

(74)

Ocsmány	(?egy)	fajzatok	mind	a	ketten.
ugly	emph	breed	all	art.def	two.of.them
‘Both are ugly breeds.’

(75)

Nem	vagyunk	mi	barom	egy	nemzet?
neg	be.prs.1pl	1pl	idiot	emph	nation
‘Are not we a really idiot nation?’

5.1.2 The morphology of the adjective

As adjectival modifiers never agree with the noun either in number, case, or possessor in Hungarian, it is no wonder that the adjective in the construction with emphatic egy is never suffixed with any of them. Although den Dikken & Lipták (1997, 69–71) claim that the emphatic “A egy N” construction is “an inverted predicate of a small clause”, they find it disturbing (although not unexplainable) that while in predication the predicate adjective (or noun) agrees with the subject in number, in this construction this requirement might be overridden. As it has been argued in Section 5.1.1, constructions with plural nouns are extremely marginal. Even if they occur, most probably they emerge by the analogy of the existing constructions with singular nouns by the substitution of the singular noun with a plural one. In this case, no agreement can be expected.

The most important issue here is the comparability of the adjective. Adjectives in the construction with emphatic egy are never comparative or superlative ((76)–(77)).

(76)

*furcsább	egy	szerkezet
strange;comp	emph	construction
‘a really stranger construction’

(77)

*legfurcsább	egy	szerkezet
strange.sup	emph	construction
‘really the strangest construction’

As superlative adjectives denote an extremity of quality and thus they are never gradual, their absence is not surprising in a construction where adjectives are expected to express gradual qualities (cf. 4.1). However, comparative can be gradual at least in the sense that something can be a bit or much more characteristic of something – e.g. (76) could mean something like ‘a very much stranger construction’.

The oddity of sentences like (78) suggests that the “A egy N” construction itself expresses a comparison (to the expectation of the speaker, cf. 4.2); therefore, it cannot express a comparison to something else.

(78)

*Ez	furcsa	(egy)	szerkezet,	de	az	még	furcsább
this	strange	emph	construction	but	that	even	strange;comp
egy szerkezet!
emph construction
‘This is (really) a strange construction, but that is even stranger of a construction.’

5.2 Syntax

Here the morphosyntactic characteristics of the construction are presented. In Section 5.2.1, the inner structure of the construction is discussed. The problem how the construction can be integrated into a sentence is treated in Section 5.2.2.

5.2.1 The inner syntax of the construction

The adjective in the construction can be modified by an adverb. This is not a rare case, there were about 200 such cases attested in the corpus material. The most frequent modifier is elég ‘fairly, enough’ (45 examples, plus 4 examples of eléggé ‘id.’) and nagyon ‘very’ (30), but adverbial pronouns as milyen ‘what (kind of), what a(n)’ (21), amilyen ‘such as’ (5), ilyen ‘like this, such, so’ (5), olyan ‘like that, such, so’ (6), mily ‘(arch.) what (kind of), what a(n)’ (1), bármilyen ‘any (kind of), whatever’ (1), etc. are also popular.¹⁴ Additionally, igen ‘very’ (11) and micsoda ‘what a(n)’ (16), igazán ‘really’ (5), ritka ‘extremely, unusually’ (5) also occur often ((79)–(85)).

(79)

Elég	ramaty	egy	honlapja	van.
fairly	shoddy	emph	home.page;poss.3sg	be.prs.3sg
‘His/her home page is shoddy enough.’

(80)

Palics	valamikor	nagyon	pletykás	egy	hely	volt.
Palić	some.time	very	gossipy	emph	place	be.pst.3sg
‘Palić used to be a really very gossipy place.’

(81)

Milyen	ocsmány	egy	rendszer!
what.a	repulsive	emph	system
‘What a repulsive system!’

(82)

Amilyen	gusztustalan	egy	állat,	még	le	is	pisilt.
such.as	disgusting	emph	animal	even	down	also	pee;pst.3sg
‘Being such a disgusting brute, s/he even peed on me.’

(83)

Miért	is	olyan	eszement	egy	ország	ez?
why	also	so	insane	emph	country	this
‘Why is this country so insane?’

(84)

Hát	micsoda	érzéketlen	egy	pasit	fogtam	ki?!
well	what	emotionless	emph	guy;acc	grab.pst.1sg	vm
‘Well, what an emotionless/insensitive guy I got!’

(85)

Ritka	gusztustalan	egy	alak	vagy.
unusually	disguisting	emph	chap	be.prs.2sg
‘You are an unusually disguisting character.’

Additionally, some other, mostly intensifying adverbs are also attested in the same position: átkozottul ‘damned(ly)’, bitangul ‘damn, bloody’, egészen ‘entirely, totally’, eszméletlen ‘mind-blowingly’, igazán ‘really’, iszonyú(an) ‘terribly, awfully’, kegyetlenül ‘terribly’, kibaszottul ‘fucking’, kurva ‘damned, fucking’, marhára ‘fecking’, meglehetősen ‘fairly, considerably’, őrült ‘madly, extremely’, rettenet ‘terribly, awfully’, rohadt(ul) ‘damn, bloody’, roppant ‘enormously’, teljesen ‘completely, entirely’, tök ‘totally, completely’, túl ‘too, overly’. Theoretically, we could expect the use of adverbs which weaken the adjectives. In practice, only one contradictory case was attested: in (86), tetű ‘louse’ is an intensifier used exclusively with lassú ‘slow’, and this is moderated by enyhén ‘mildly’.¹⁵

(86)

Képekkel	böngészni	enyhén	tetűlassú	egy	dolog.
image;pl.ins	browse;inf	mildly	louse.slow	emph	thing
‘Browsing with images is a bit damned slow thing.’

Rarely, the adverb can denote epistemic circumstances (87), point of view (88) or reason (89).

(87)

A	hobbitok	köztudottan	ínyenc	egy	népség.
art.def	hobbit;pl	well-known.that	gourmet	emph	people
‘Hobbits are known as really gourmet people.’

(88)

Motorikusan	kellemes	egy	autó.
motor;adjz;advz	pleasant	emph	car
‘As far as the engine is concerned, it is a pleasant car.’

(89)

A	történelem	attól	bonyolult	egy	struktúra,	…
art.def	history	that.abl	complicated	emph	structure
‘History is made such a really complicated structure by …’

There are cases, in which it is difficult or impossible to determine whether the adverb (or rather a discourse particle) modifies the adjective or the whole phrase ((90)–(92)).

(90)

Az	valóban	rémes	egy	látvány.
art.def	really	horrible	emph	spectacle
‘That is a really horrible spectacle.’ or  ‘That is really a horrible spectacle.’

(91)

Kétségkívül	dögös	egy	bige.
undoubtedly	kicking	emph	chick
‘She is an undoubtedly kicking chick.’ or
‘She is undoubtedly a kicking chick.’

(92)

A	szemezés	tényleg	fura	egy	dolog.
art.def	ogling	indeed	strange	emph	thing
‘Ogling is a really strange thing’ or
‘Ogling is a strange thing, indeed.’

Marginally, the adjective can be negated ((93)–(94)). In these cases, the adjective is negated, the negative particle is interpreted similarly to a prefix deriving an antonym.

(93)

Nem	rossz	egy	csomag	az.
neg	bad	emph	packet	that
‘Really, that packet is not a bad one.’

(43)

Talán	nem	érdektelen	egy	adalék	az
perhaps	neg	uninteresting	emph	complementing.data	art.det
irodalmi	utókornak.
literature;adjz	posterity;dat
‘It might be a really interesting additional piece of information for the posterity interested in literature.’

(94)

Nem	igazán	fiatalos,	nem	rugalmas	egy	autó.
neg	really	youngish	neg	flexible	emph	car
‘It is not really a youngish, not a flexible car.’

Example (94) shows another peculiarity of the construction: there can be two (or even more) adjectives in it. In some cases, it is difficult to determine whether a word is an adjective or an adverb. In example (95), szuper can be an adjective (‘super, great’), or an intensifying adverb (‘very, super-’).¹⁶

(95)

Szuper	jó	egy	oldal!
super	good	emph	site
‘It is really a super, great / very great site.’

In example (96) it is not clear whether there is only one but coordinated adjective compound or two juxtaposed adjectives.

(96)

Zsugori-fukar	egy	ember	volt.
stingy-mingy	emph	human	be.pst.3sg
‘S/he was really a tight-fisted wo/man.’

However, in about fifteen cases, two adjectives are separated by a comma ((97)–(99)). As example (99) shows, the adjectives themselves can be derived even from collocations. The two adjectives are usually synomyms, or at least similarly negative or positive words. When they are not, it is rather questionable whether it is an “A egy N” construction (100).

(97)

Hihetetlen,	őrült	egy	futam	volt.
inbelievable	crazy	emph	race	be.pst.3sg
‘It was a really unbelievable, crazy race.’

(98)

Sunyi,	lusta	egy	ember	volt.
sneaky	lazy	emph	human	be.pst.3sg
‘S/he was really a sneaky, lazy wo/man.’

(99)

Gyönyörű	szabású,	tetszetős	egy	kard.
beautiful	cut;adjz	good-looking	emph	sword
‘It is a really beautifully-cut, good-looking sword.’

(100)

Nem	szól	semmit	semmihez,	oly	közönyös,
neg	say.prs.3sg	nothing;acc	nothing;dat	so	stolid
csodálatos	egy	lány.
wonderful	emph	girl
‘She does not comment anything on anything, she is really so stolid, wonderful a girl.’
or rather: ‘… she is so stolid, she is a really wonderful a girl.’

In constructions with multiple adjectives, emphatic egy tends to occur after the last adjective. In these cases, it is impossible to tell whether the emphasis applies to all the adjectives or just to the last one(s). We could identify only one example in which emphatic egy does not occur after the last adjective (101).

(101)

Amilyen	álnok	egy	kurva	természete	van…
such.as	treacherous	emph	damn	nature;poss.3sg	be.prs.3sg
‘Having a such really treacherous damn character she …’

Such examples are easy to construct if we choose relational adjectives to stand after the emphatic egy ((102)–(104)). Relational adjectives are closely connected to the modified noun, resembling compounds. In these cases it is impossible for the emphatic egy to stand after the relational adjective, as it cannot be attached to a relational adjective in general (cf. Section 4.1).

(102)

borzalmas	egy	általános	iskola
terrible	emph	general	school
‘a really terrible elementary school’

(103)

száraz	egy	háztartási	keksz
dry	emph	household;adjz	biscuit
‘a really dry budget biscuit’

(104)

érdekes	egy	történelmi	esemény
interesting	emph	historical	event
‘a really interesting historical event’

In some cases, the adjectives are separated by a conjunction (105–107). In our corpus, no convincing examples could be found for the alternative coordination with vagy ‘or’. The only example, presented in (107), is difficult to interpret. On the one hand, the adjectives can qualify the thousand different men who spent a lot of time with a woman-myth; in that case egy is the indefinite article modifying the compound nő-mítosszal ‘with a woman-myth’. On the other hand, the sequence of alternative properties expressed by the adjectives, with the mysterious phrase sercegő bábos ‘with scrapping puppet(s)’¹⁷ among them, qualify the woman-myth. In that latter case, egy can be interpreted as an emphatic clitic.¹⁸

(105)

Elég	durva	és	szomorú	egy	könyv.
fairly	crude	and	sad	emph	book
‘It is a really crude and sad book.’

(106)

Az	egy	olyan	forró,	meg	olyan	szenvedélyes	egy	nyelv.
that	art.indf	so	hot	and	so	passionate	emph	language
‘That is really such a hot and such a passionate language.’

(107)

a.

Ezer	órákat	töltött	ezer	férfi
thousand	hour.pl	spend;pst.3sg	thousand	man

b.

nagyagyú	kicsinyes	nagyvonalú
big.brained	small-minded	generous

c.

sercegő	bábos	vagy	rosszindulatú
scrape;prs.part	puppet;adjz	or	malevolent

d.

egy	nő-mítosszal.
??	woman-myth-ins
‘Thousand men – being clever, small-minded, generous, scraping-puppeted (?), or malevolent – spent thousand hours with a woman-myth.’ or: ‘Thousand men spent thousand hours with a clever, small-minded, generous, scraping-puppeted (?) or malevolent woman-myth.’

Example (108) is also interesting, because the whole phrase is used as a modifier of a noun (barátnő ‘girlfriend, female friend’).

(108)

egy	igazi	kötött	pulcsis,	SZTK	keretes
art.indf	real	knit;pst.part	sweater;adjz	NHS	frame;adj
szerencsétlen	egy	hárpia	barátnővel	együtt
wretched	emph	harpy	girlfriend-ins	together
‘together with a girlfriend who is a real harpy wearing a knitted sweater and an NHS-financed (i.e. cheap, crippled) (eyeglass) frame’

Example (108) is also strange in that it is preceded by an indefinite article. However, it does not belong to the “A egy N” construction but to barátnő ‘girlfriend, female friend’. As it will be shown in 5.2.2, “A egy N” constructions cannot be preceded by an article belonging to them. Nonetheless, there are eight examples in our material in which an indefinite article occurs before the “A egy N” construction. We consider these constructions ill-formed. A strong argument for this decision is that although constructions with adjectives modified by adverbs or with multiple adjectives represent a minority of the cases, in half of the examples with an indefinite article we find either an adverb or multiple adjectives or both ((109)–(111)). However, it is not always the case, cf. (112).

(109)

¿Ez	egy	nagyon	fura	egy	helyzet.
this	art.indf	very	strange	emph	situation
‘This is really a very strange situation.’

(110)

¿Én	egy	szerencsés,	boldog	egy	ember	vagyok.
1sg	art.indf	lucky	happy	emph	human	be.prs.1sg
‘I am a really lucky, happy person.’

(111)

¿Ez	kérem	egy	roppant	egyszerűen	bonyolult	egy	kérdés.
this	you.know	art.indf	enormously	simply	complicated	emph	question
‘This is an enormously simply complicated question, you know.’

(112)

¿Milyen	egy	érdekes	egy	nép	vagyunk.
what.kind.of	art.indf	interesting	emph	people	be.prs.1pl
‘We are such an interesting people.’

5.2.2 The outer syntax of the construction

In most cases the “A egy N” construction forms a bare predicate with a covert pronoun subject. The schematic structure of (113) is shown in (114) following the common analysis of Hungarian matrix clauses (Kálmán 2001b, 24–35). In Hungarian, copular van ‘be’ has to be omitted in 3sg in present tense, but has to be overt in first and second person, in plural, in past or future tenses, and in non-indicative moods (115).

(113)

Furcsa	egy	helyzet –	vélekedett	Bob	Barr.
strange	emph	situation	opine.pst.3sg	Bob	Barr
‘What a strange situation, opined Bob Barr.’

(114)

[SUBJECT(Ez)]	[PREDICATEfurcsa	egy	helyzet]
this	strange	emph	situation
‘This is a strange situation’

(115)

Furcsa	egy	helyzet	volt.
strange	emph	situation	be.pst.3sg
‘It was a really strange situation.’

In (113) the subject pronoun is deictic, its reference is the present situation of the discourse. However, the subject can also be a referential expression as in (116).¹⁹

(116)

Szóval	furcsa	egy	hely	[volt]	Japán.
well	strange	emph	place	be.pst.3sg	Japan
‘Well, Japan is (was) a really strange place.’

In ordinary predication, the resulting declarative denotes a proposition (p) that can be negated (¬p), see (117)–(118). Despite the fact that we found above that syntactically the “A egy N” construction behaves like a predicate, (120) shows that it cannot be negated; note that (122) is not the negated form of (119), but that of (121).

(117)

Budapest	szép.
Budapest	beautiful
‘Budapest is beautiful.’

(118)

Budapest	nem	szép.
Budapest	neg	beautiful
‘Budapest is not beautiful.’

(119)

Japán	furcsa	egy	hely.
Japan	strange	emph	place
‘Japan is a really strange place.’

(120)

*Japán	nem	furcsa	egy	hely.
Japan	neg	strange	emph	place
‘Japan is not a really strange place.’

(121)

Japán	egy	furcsa	hely.
Japan	art.indf	strange	place
‘Japan is a strange place.’

(122)

Japán	nem	egy	furcsa	hely.
Japan	neg	art.indf	strange	place
‘Japan is not a strange place.’

The ungrammaticality of the negative declarative in (120) can be an argument for the pragmatic marker status of egy in “A egy N” construction, which does not only have a semantic contribution, but a pragmatic one. (This would also be in accordance with what we stated above in 5.2.1 about the marginal examples (93)–(94).)

Compared to its function as a predicate, quite rarely, the “A egy N” construction can also be the complement of the verb in the sentence. In (123) the function of the construction under discussion is the sentence subject. (124) is seemingly similar to (123), but (124) is a possessive sentence. In (84) the function of the construction is the sentence object.²⁰ In (125) the function of the “A egy N” construction is a locative complement.

(123)

Ronda	egy	idő	van.
ugly	emph	weather	be.prs.3sg
‘We have really ugly weather.’

(124)

Furcsa	egy	ízlésed	van.
strange	emph	taste;poss.2sg	be.prs.3sg
‘You have a really strange taste.’

(84)

Hát	micsoda	érzéketlen	egy	pasit	fogtam	ki?!
well	what	emotionless	emph	guy;acc	grab.pst.1sg	vm
‘Well, what an emotionless/insensitive guy I got!’

(125)

Érdekes	egy	országban	élsz.
interesting	emph	country;ine	live;prs.2sg
‘You live in a really strange country.’

Although some examples of den Dikken & Lipták (1997, 65) suggest that the “A egy N” construction can occur postverbally if the preverbal focus position is filled by another element, there are no more than three occurrences of “A egy N” in our corpus sample where the construction is not immediately preceding the verb ((126)–(128)), of which only one (in (128)) is a postverbal occurrence. According to our native intuition, these examples are extremely odd.

(126)

¿Dica	előbb	mellre	vette	a	labdát,
Dica	first	breast;supess	take;pst;3sg;do	art.def	ball;acc
majd	szép	egy	fordulással	a	jobb	alsó
later	beautiful	emph	turn;ins	art.def	right	bottom
sarokba	küldte.
corner;ill	send;pst;3sg;do
‘Firstly, Dica took the ball on his breast, then with a very beautiful turn, he sent it to the right bottom corner.’

(127)

¿Széles	egy	világ	az	enyém	lett.
wide	emph	world	art.def	mine	become;pst;3sg
‘A wide world became mine.’

(128)

¿Biztos	van	szegény	egy	lánynak	egy	csomó	dolga.
sure	be;prs;3sg	poor	emph	girl;dat	art.indef	lot	thing;poss.3sg
‘This poor girl has surely a lot of things to do.’

Corpus examples in (123)–(125) illustrate that, in fact, the complement realised by the “A egy N” construction in most cases occurs preverbally. Furthermore, (84) shows clearly that the construction under discussion occupies the immediate preverbal position, that is, the so-called focus position (see É. Kiss 2002, 83–104).²¹ The ungrammaticality of (129) reflects that, at least according to our native intuition, the construction cannot occur in the postverbal domain, and (130) presents that the construction under discussion cannot occupy a preverbal position other than the preverbal focus position either (inversion of the VM-V sequence).²²

(129)

*Kifogtam	(egy)	érzéketlen	egy	pasit.
vm;grab.pst.1sg	art.ind	emotionless	emph	guy;acc
‘I got a really insensitive/emotionless guy.’

(130)

*Érzéketlen	egy	pasit	kifogtam.
emotionless	emph	guy;acc	vm;grab.pst.1sg
‘I got a really insensitive/emotionless guy.’

According to É. Kiss (2002, 90), the function of the preverbal focus position is exhaustive identification. The predicate of the sentence holds only for the entity/entities expressed by the constituent in the preverbal focus position. Despite its similar syntactic position, the “A egy N” construction does not express “exhaustive identification”. In (84), the érzéketlen egy pasit phrase does not identify the person I got, rather it states that the person I got is emotionless/insensitive. Note also that (84) can only be felicitous in contexts where flirting or having a significant other is already a question under discussion.

If we take a closer look at the syntactic behavior of sentences containing the “A egy N” construction, the question of their sentence type arises. Firstly, let us begin with sentences in which “A egy N” is a predicate, like the one in (116).

(116)

Szóval	furcsa	egy	hely	[volt]	Japán.
well	strange	emph	place	be.pst.3sg	Japan
‘Well, Japan is (was) a really strange place.’

At first sight, (116) seems to be an ordinary declarative sentence. However, the corresponding polar interrogative matrix clause sounds somewhat strange ((131)–(132)); its embedded version is also weird (133). On the contrary, a biased negative polar interrogative with an outside negation reading can be felicitous (134). It can also be the anchor, so the declarative part of a tag question (135). As for wh-interrogatives, constructed examples (136)–(138) seem to be possible (at least as echo questions), although our grammaticality judgement is uncertain, as these kinds of structures are so rare that they do not occur at all in our corpus sample.

(131)

?Furcsa	egy	hely	Japán?
strange	emph	place	Japan
‘Is Japan a really strange place?’23

(132)

??Furcsa	egy	hely-e	Japán?
strange	emph	place-qprt	Japan
‘Is Japan a really strange place?’

(133)

*Péter	nem	tudja,	hogy	furcsa	egy	hely-e	Japán.
Peter	neg	know.prs.3sg	that	strange	emph	place-qprt	Japan
‘Peter does not know whether Japan is a really strange place?’

(134)

Hát	nem	furcsa	egy	hely	Japán?
well	neg	strange	emph	place	Japan
‘Well, isn't Japan a really strange place?’

(135)

Furcsa	egy	hely	Japán,	ugye/nem?
strange	emph	place	Japan	ugye.tag/nem.tag
‘Japan is a really strange place, right?’

(136)

Mi	furcsa	egy	hely?
what	strange	emph	place
‘What is a really strange place?’

(137)

Milyen	egy	hely	Japán?
what.like	emph	place	Japan
‘What kind of place is Japan?’

(138)

Furcsa	egy	MI	Japán?
strange	emph	what	Japan
‘Really strange a what is Japan?’

In our corpus sample, we can hardly find sentences that seem to be interrogatives. The only example is (139), taken from the poem Bárány Boldizsár (Balthazar, the lamb) by Magda Szabó, where the construction is used as a doubtful repetition of what has recently been said.²⁴

(139)

Szorgalmas	egy	menyecske? –	duzzog	Boldizsár
hardworking	emph	maiden	sulk.prs.3sg	Boldizsár
‘Is she really a hardworking maiden?, sulks Boldizsár.’

All the other cases when the construction with emphatic egy and a question mark appear together belong to the following types: (i) tag questions with sentence-initial ugye (‘is that so?’) (140), sentence-final nem (‘not?’) (141), or nem igaz (‘isn't that right’) (142); (ii) questions with tag-like elements: mi? (‘what?’) (143), hallod-e? (‘do you hear it?’) (144), or tudod-e? (‘do you know it?’ (145); (iii) negative polar interrogatives with outer reading (146); (iv) self-addressed/rhetorical questions (147).

(140)

Ugye	épületes	egy	ügy?
tag	instructive	emph	case
‘It is a really instructive case, isn't it?’

(141)

Eszement	egy	ötlet,	nem?
crazy	emph	idea	not
‘It is a really crazy idea, isn't it?’

(142)

Elég	uncsi	egy	magyarázat,	nem	igaz?
enough	boring	emph	explanation	not	true
‘It is a really boring explanation, isn't it?’

(143)

Furcsa	egy	világ	ez,	mi?
strange	emph	world	this	what
‘It is a really strange world, isn't it?’

(144)

Nevetséges	egy	figura	vagy	néha,	hallod	-e?
ridiculous	emph	guy	be;prs.2sg	sometimes	hear;prs.2sg	qprt
‘You are sometimes a really ridiculous guy, don't you think?’

(145)

Te	eszméletlen	hülye	egy	ökör	vagy
you	unconsciously	stupid	emph	bullock	be;prs.sg2
tudod	-e?
know;prs.2sg	qprt
‘You are extremely stupid bullock, don't you know?’

(146)

Nem	vagyok	én	fura	egy	nő?
Not	be;prs.1sg	strange	emph		woman
‘Am I not a really strange woman?’

(147)

Vajon	sértődékeny	egy	elnök?
prt	touchy	emph	president
‘I wonder whether he is a really touchy president.’

Secondly, let us examine also marginal cases where the “A egy N” construction is a complement, like the one in (125). In our corpus, there are no interrogatives in which the “A egy N” construction occurs; although we can construct examples ((148)–(152)).

(148)

?Érdekes	egy	országban	élsz?
interesting	emph	country;ine	live;prs.2sg
‘Do you live in a really interesting country?’

(149)

??Érdekes	egy	országban	élsz-e?
interesting	emph	country;ine	live;prs.2sg-qprt
‘Do you live in a really interesting country?’

(150)

#Nem	tudom,	hogy	érdekes	egy	országban
neg	know.prs.1sg	that	interesting	emph	country;ine
élsz-e.
live;prs.2sg-qprt
‘I do not know whether you live in a really interesting country.’

(151)

*Nem	élsz	érdekes	egy	országban?
neg	live;prs.2sg	interesting	emph	country;ine
‘Don't you live in a really interesting country?’

(152)

Érdekes	egy	országban	élsz,	ugye/nem?
interesting	emph	country;ine	live;prs.2sg	ugye.tag/nem.tag
‘You live in a really interesting country, right?’

It has already been demonstrated that sentences containing the “A egy N” construction cannot be transformed to a corresponding interrogative; our corpus sample also lacks these types of data. However, we do have some sentences realising non-canonical, biased questions in which “A egy N” constructions can occur. What is common in these types of questions is that their sentence type is not interrogative. Thus, the construction under discussion cannot occur in interrogatives. It has also been demonstrated above that sentences containing the “A egy N” construction cannot be negated (120); and negative sentences hardly ever occur in our corpus sample.

The speaker-related evaluative and emotive nature of utterances containing the “A egy N” construction suggests that their sentence type may be exclamative. In Hungarian, some constructions are similar to wh-exclamatives in English. These include micsoda-exclamatives (see Section 6 below and Lipták 2006) and de-exclamatives. While micsoda ‘what’ may be followed by a noun or an adjective, de ‘literally: but’ may be followed by an adjective or an adverb. The “A egy N” construction occurs not only in wh-exclamatives, but also in de-exclamatives in our corpus sample: we found 18 instances like the ones in (153)–(156).

(153)

De	furcsa	egy	válság	ez, …
de	strange	emph	crisis	this
‘What a strange crisis this is!’

(154)

De	szörnyű	egy	eset.
de	awful	emph	case
‘What an awful case!’

(155)

De	cudar	egy	anyag!
de	bad	emph	stuff
‘What a bad stuff!’

(156)

Jajde	etikus	egy	vitapartner	vagy!
exclam.prt	ethical	emph	debate;partner	be.prs.2sg
‘What an ethical debate partner you are!’

Sentences containing the “A egy N” construction (without micsoda, mekkora, or de) do not show all the characteristics of exclamatives (Lipták 2006, 374). It seems that they can be embedded under non-factive predicates, e.g. úgy tudom ‘I have the information that’ ((157)–(158)).

(157)

Úgy	tudom,	furcsa	egy	helyzet	volt.
so	know.prs.1sg	strange	emph	situation	be.pst.3sg
‘As far as I know it was a really strange situation.’

(158)

Úgy	tudom,	érzéketlen	egy	pasit	fogott	ki.
so	know.prs.1sg	emotionless	emph	guy	get.pst.3sg	out
‘As far as I know she got a really emotionless guy.’

Scalarity, another characteristic feature of exclamatives, characterises sentences containing the “A egy N” construction, as well. This is due to the fact that A in the construction should denote a gradual quality (see 4.1).

6 Nouns instead of adjectives?

According to den Dikken & Lipták (1997), Hungarian also has a “N/A egy N” construction. Up to this point, we used the terms noun and adjective as if they denoted well-defined and clearly distinguishable categories. In this section, it is demonstrated that it is not always easy to distinguish the two categories: some words show characteristics of both nouns and adjectives, and each individual word can be transitional in its own way. In any case, it will be argued that “N egy N” constructions do not exist in Hungarian.²⁵

Before looking at the problems with nouns and adjectives, let us consider another problem. In their paper, den Dikken & Lipták (1997, 69–71) argue that the “N/A egy N” construction is “an inverted predicate of a small clause” (den Dikken & Lipták 1997, 69). They support their claim by demonstrating the different syntactic behavior of two adjectives meaning ‘small’, kis and kicsi. Both can be used in attributive positions (159), but kis cannot occur in predicative positions.

This demonstration cannot be accepted as a decisive argument, because kis cannot occur either when the noun is not present after it in such sentences as (162). In the same way, one could argue that there is an ellipsis in the “N/A egy N” construction. In any case, we do not want to argue either against or in the favour of any of these analyses.

The test provided by den Dikken & Lipták (1997) is interesting for us because it is worth conducting similar tests for alleged nouns standing before egy in the supposed “N egy N” construction.

The statement that “N egy N” constructions exist is surprising, as most of the nominal predicative phrases cannot be associated to emphatic constructions with egy, despite the fact that the predicate classifies or identifies the subject ((163)–(166)).

(163)

A	cselló	egy	hangszer.
art.def	cello	art.indf	instrument
‘Cello is an instrument.’

(165)

A	cselló	az	a	gordonka.
art.def	cello	that	art.def	violoncello
‘Cello is the (same as) violoncello.’

Den Dikken & Lipták (1997, 62–63, 65, 68, 70–71) cite only three nouns occurring as the first noun in “N egy N” constructions: csoda ‘wonder’, kutya ‘dog’ and fene ‘hell’.

The case of csoda seems to be the most convincing, although even this is problematic. It is used as a noun undoubtedly, but it is also used in a position not so typical for nouns, as an attributive. According to Juhász et al. (1992), in that position it means ‘wonderful, extraordinary’. (Additionally, it is also used as an adverb, meaning ‘extraordinarily, remarkably, very’.) However, in many cases csoda is used in a prefix-like way, i.e. as a first constituent of a compound, basically in the very same meaning: csodacsatár ‘wonderstriker’, csodacsapat ‘wonderteam’, csodafegyver ‘wonderweapon’, csoda(gyógy)szer ‘wonder drug, magic pill’, csodamasina ‘wonder machine, wonder gadget’, etc. In these cases, the constituent rather means ‘extraordinary’ than anything ‘supernatural’. The difference is not striking, especially if we consider that csoda as an independent noun is also used in a similar meaning (a művészet/természet csodái ‘the wonders (extraordinary products) of art/nature’). However, the difference can be noticed if we compare these compounds with the rarer type of csodaország ‘wonderland’ or csodaszarvas ‘wonder deer (mythical animal)’, in which csoda really means something ‘supernatural’. In the supposed “N egy N” construction, it is also interpreted as extraordinary in some respect, not as anything supernatural ((167)–(169)).

(168)

Egy/A	nyelv	(egy/a)	csoda.
art.ind/def	language	art.ind/def	wonder
‘A/The language is a/the wonder.’

(169)

egy	csoda	nyelv
art.indf	wonder	language
?‘a wonderful language’, rather ‘an extraordinary language’

Den Dikken & Lipták (1997, 68) state that *csoda kocsi ‘wonder car’ is agrammatical, which is basically true, but csodakocsi is perfect in the same meaning. Actually, it is very difficult to find examples in which csoda does not form a compound with the noun it modifies. Juhász et al. (1992) gives the example Csoda szerencséje volt ‘s/he had wonderful luck’: csoda szerencse ‘wonder(ful?) luck’ is certainly better than csodaszerencse. The behavior of csoda needs further investigation, here we can only point out two facts. On the one hand, while the usual second constituents in compounds with csoda denote objects or persons, csoda and szerencse are both abstract notions, which can also be almost synonymous: Csak a csoda/szerencse segíthet ‘Only a wonder/luck can help’. On the other hand, the behavior of adjectival modifiers and adjectival first constituents in the “A egy N” construction seems to be very similar. For example, szuper hős ‘super (terrific) hero’ can be contrasted to szuperhős ‘superhero, a hero with supernatural skills’. However, the constructed example (170) can also be used when we acknowledge that someone is a real superhero (or sarcastically, when someone was called a superhero, but s/he is not a hero at all).

(170)

Ő	aztán	tényleg	szuper	egy	hős!
S/he	then	really	super	emph	hero
‘S/he is really what I call a super hero.’ or ‘…superhero.’

This kind of usage is metalinguistic, which can be demonstrated by another constructed example. In Hungarian, jegesmedve ‘polar bear’ is a compound consisting of the adjective jeges ‘icy, iced’ (derived from the noun jég ‘ice’) and the noun medve ‘bear’ (although the bear's environment is usually icy, but the bear itself is not). It can be contrasted to jeges medve ‘icy bear’, a bear (even a brown bear, a panda, or a koala) covered by ice. However, a sentence like (171) can be expected when someone sees a photo of an icy polar bear and refers to the fact that this particular polar bear is icy in the literal sense of the word, as its name suggests. Even if the bear on the picture is not a polar bear, it will be interpreted as a reference to the name of the polar bear.

(171)

Ez	aztán	tényleg	jeges	egy	medve!
This	then	really	ice-adjz	emph	bear
‘This is really what I call an icy bear.’ or ‘…a polar bear.’

Moreover, a similar metalinguistic use can occur even with compounds with a noun as their first constituent. The noun látványpékség ‘exhibition bakery’ denotes bakeries where the customers can see at least some parts of the manufacturing process (especially the ovens). The compound consists of látvány ‘sight, view, spectacle’ and pékség ‘bakery’. The constructed sentence (172) can be used for admitting that a bakery deserves its name, namely it is really spectacular.

(172)

Ez	aztán	tényleg	látvány	egy	pékség!
This	then	really	spectacle	emph	bakery
‘This is really what I call an exhibition bakery.’

Based on the latter fact, one could argue that “csoda egy N” constructions belong to this type, that is, csoda is a noun in them. However, it could hardly be accepted. Examples as (172) are very marginal, we could not find any examples in the corpus, and maybe such constructions are not acceptable for all native speakers of Hungarian – while “csoda egy N” constructions are widely used. A process in the opposite direction is much more probable: as csoda can be a noun and “csoda egy N” is usual, constructions as (172) can emerge by the analogy of the constructions with csoda. The behaviour of csoda can be explained by its distribution, i.e. that it occurs in positions typical for both nouns and adjectives in the very same meaning.

Moving back to the examples of den Dikken & Lipták (1997), the case of kutya ‘dog’ is much more straightforward ((173)–(175)). As a noun, it is used in several positions. However, in the attributive position, it cannot mean ‘dog’ or anything ‘related to dogs’, but simply a swearword which can be generally translated as ‘damn’; except for some idiomatic phrases, such as kutya idő ‘nasty weather’, it is outdated and sounds strange. Otherwise, if it occurs at all, it is interpreted as something negative, depending on the meaning of the modified noun. If it is used with nyelv ‘language, tongue’, it can be interpreted as ‘difficult, complicated’, but also can be used as a simply defamatory attribute. An important point is that kutya as a predicate noun can only be interpreted as ‘dog’. Consequently, the interpretation of kutya is completely different as a predicate than as the first noun in the supposed “N egy N” construction. As the word has a different meaning in the attributive position as generally otherwise, it is reasonable to analyze the attributive kutya as an adjective. This analysis is supported by the fact that in this position, kutya can be compared similarly to other adjectives: kutyább/legkutyább idő ‘nastier/nastiest weather’. As a consequence, the phrase kutya egy nyelv must be analyzed as an “A egy N” construction.

(174)

Egy/A	nyelv	(egy/a)	kutya.
art.ind/def	language	art.ind/def	dog
‘A/The language is a/the dog.’

(175)

egy	kutya	nyelv
art.indf	“dog”	language
‘a ?difficult language’

It has to be remarked that kutya is often used as a euphemic expression instead of kurva, which means ‘whore’ as a noun (and it cannot be interpreted any other way as a predicate). However, as an attributive, it is used in a similar way as the attributive kutya ((176)–(178)).

(177)

Egy/A	nyelv	(egy/a)	kurva.
art.ind/def	language	art.ind/def	whore
‘A/The language is a/the whore.’

(178)

egy	kurva	nyelv
art.indf	“whore”	language
‘a ?difficult language’

The case of fene is even more striking ((179)–(181)). Den Dikken & Lipták (1997, 62) gloss fene as ‘hell’, which is clearly perfect in some cases, as in the phrases menj a fenébe ‘go to hell’, hol a fenében van? ‘where the hell is it?’, mi a fenét gondolsz? ‘what the hell is on your mind?’, etc.; but it does not mean in general ‘underworld’ or any other ‘terrible place’. Originally it meant ‘ulcerated, purulent wound’, but it is not used in this meaning anymore. It just occurs in idiomatic expressions as (182)–(184). As a predicate, it cannot be interpreted in any way. On the contrary, it is interpreted similarly to an attribute in the “A egy N” construction: in this case, its meaning is negative and vague as in the case of kutya or kurva.

(180)

Egy/A	nyelv	(egy/a)	fene.
art.ind/def	language	art.ind/def	???
‘A/The language is a/the ???.’

(181)

egy	fene	nyelv
art.indf	damn	language
?‘a damn language’

There are some cases in our corpus that suggest that the “A egy N” construction is closer to the attributive structure than to predication, and it is not a “N egy N” construction at all ((185)–(190)).

(186)

Egy/A	szarvas	(egy/a)	suta.
art.ind/def	deer	art.ind/def	???
‘A/The deer is a/the roe/clumsy (one).’

(187)

egy	suta	szarvas
art.indf	clumsy	deer
‘a clumsy deer’

(189)

A	szőr	tiszta.
art.def	hair	clean
‘The hair is clean.’

(190)

tiszta	szőr
clean	hair
‘full of hair’ or ‘clean hair’

Examples with csurom ‘soaking’ and merő ‘pure, full, complete’ ((191)–(193)) are especially interesting with respect to our discussion, as they can be used in the “X egy N” construction, as well.

Items occupying the position of A in these structures cannot be used as predicates ((194)–(195)), only as intensifying modifiers.

Before concluding that the phrases with csoda, kutya, fene (and also kurva) are the representations of the “A egy N” construction, we have to consider one more fact: these words can also be used as adverbs, more precisely, in an “adadjectival” position. All of them function as intensifiers (196). Although csoda is basically positive while the others are negative, this is manifested only in the fact that csoda is not combined with adjectives bearing negative connotations.

(196)

csoda/fene/kutya/kurva	jó/nagy/érdekes
damn	good/big/interesting
‘damn good/big/interesting’

This fact allows us to raise the possibility that “csoda/kutya/kurva/fene/etc. egy N” constructions are in fact “ADV egy N” constructions, which developed from “ADV A egy N” constructions with omitting the adjective.

This possibility is supported by another observation. E.g. in example (44), it was argued that ritka, instead of its general meaning ‘rare, infrequent’, must be interpreted as ‘strange, unusual’. In Section 5.2.1 it has been demonstrated that ritka is also used as an adverb modifying the adjective of the “A egy N” construction. Based on these facts, we can interpret ritka egy alak as a “ADV egy N” construction, from which the adjective was omitted.

This can be supported by the fact that although egy usually does not occur between an adverb and an adjective, we find several similar cases in the corpus. These constructions are emphatic, similarly to the “A egy N” constructions ((197)–(200)).

Similar examples can also be found with csoda and fene ((201)–(202)) (however, just one of each), although not with kutya and kurva.

The emergence of the forms in which egy occurs between the adverb and the adjective instead of between the adjective and the noun (which is also an unusual position) can easily be accounted for by the modification of the surface forms. The unexpected word order adds expressiveness to the collocation, which can be institutionalised (grammaticalised). Similarly, omitting the adjective from the “ADV A egy N” construction gives an emphasis to the unusuality and extraordinality of the thing denoted by the noun, leaving out the specification of its exceptionality.

This assumption is supported by another type of the “X egy N” constructions, distinguished by den Dikken & Lipták (1997, 67–68), which they call micsoda exclamatives. As it was discussed in Section 5.2.1, micsoda is a relatively frequent adverbial-like modifier of adjectives in “A egy N” constructions. Additionally, “micsoda egy A N” constructions also occur in the corpus ((203)–(205)).

In these cases, it does not make a real difference that the emphatic clitic follows the adverb or the adjective. In addition to micsoda, other question words, milyen ‘what kind of’ and mekkora ‘what size of’ are used in the very same way ((206)–(210)). Den Dikken & Lipták (1997) do not discuss these together with micsoda, supposedly because they analyze micsoda as a noun, while milyen and mekkora as adjectives. According to us, in the constructions with emphatic egy, their adverbial (more precisely, “adadjectival”) behaviour plays a key role.

Similarly to the cases with micsoda, the adjective(s) can be completely omitted after milyen and mekkora ((211)–(215)). In most of the cases ((211)–(213)), the noun in the phrase is a word that could be used in the position of the adjective as well, so these can be analysed supposing that the noun was dropped. This latter interpretation is supported by the fact that the phrases with and without an adjective meaning great size are synonymous: milyen/mekkora egy féreg! ‘what a scum’ ∼ milyen/mekkora hatalmas egy féreg! ‘what a great scum’, etc. However, examples (214)–(215) show that there are cases when it is sure that the adjective was omitted. Although mekkora appears in both of them, it can be changed to milyen in any case. Nonetheless, examples with milyen are more difficult to identify: milyen is a question word with quite a vague meaning. Therefore, exclamations starting with this question word, which resemble rhetorical questions, can be always interpreted as actual open-ended questions. On the contrary, exclamations with mekkora are much easier to identify, as not all the nouns in the structure denote things that have actual size.²⁶

Among the pronouns used in the “X egy N” construction semmi ‘nothing’, more exactly nem semmi ‘not nothing’ should also be mentioned. Although in Hungarian double negation is the norm in some structures (216), in this case, double negation is used as something positive (meaning ‘uncommon, unusual, extraordinary; significant, considerable’). It is used in predication, as an attribute, and in emphatic constructions with egy as well ((217)–(219)).

(216)

Nem	csináltam	semmit.
neg	do;prt.s1	nothing;acc
‘I did not do anything.’

(217)

Ez	a	feladat	nem	semmi!
this	art.def	task	neg	nothing
‘This task is quite extraordinary.’

(218)

Nem	semmi	feladatot	adott	a	tanár.
neg	nothing	task;acc	give;pst.3sg	art.def	teacher
‘The teacher gave an extraordinary task.’

Before concluding that there are no “N egy N” constructions at all, we have to consider a specific group of nouns. A specific feature of these nouns is that they denote people (less typically animals, objects, etc.) who (which) can be characterised by some quality. For example, hős ‘hero’ is basically a bátor ember ‘brave, courageous person’, zseni ‘genius’ is a ‘intelligent, smart person’. These nouns can also be used in an attributive position, where their meaning is basically the same: hős ‘brave, courageous’, zseni ‘intelligent, smart’. (This case is completely different from cases like that of kutya, which has completely different meanings when used as a noun (‘dog’) and when used as an adjective (‘damn’).) Moreover, words like hős ‘hero’ and zseni ‘genius’ can be compared, although not in a way typical for adjectives, but with the comparative and superlative forms of the adjective nagy: nagyobb hős ‘greater hero; braver, more courageous person’, legnagyobb zseni ‘greatest genius; the most intelligent, the smartest person’. However, these compared forms cannot be used when the noun is in an attributive position. (On the contrary, kutya can be compared in the regular way (kutyább, legkutyább), and these forms can be used in the attributive position.) In any case, these words are rather nouns than adjectives due to their distribution, but they can occur in emphatic constructions with egy (220–221). However, these nouns can have no attributive adjectives modifying them (222–223).²⁷

(220)

Hős	egy	szomszédod	van.
hero	emph	neighbour-poss.2sg	be.prs.3sg
‘Your neighbour is a real hero.’

(221)

Zseni	egy	férje	van.
genius	emph	husband-poss.3sg	be.prs.3sg
‘Her husband is a genuine genius.’

Additionally, there are also some nouns denoting persons (or less tipicallly animals, things) by quality and used in attributive position as well, which are not comparable, and can be used in emphatic constructions with egy. For example, vak ‘blind’ is such a word, but in (224) it can rather refer to someone visually impaired (but not completely blind, although the phrase refers to the fact that the problem is serious); or in a metaphorical sense, suggests that the wife is unsuspicious and careless despite some evident signs (e.g. of cheating).

(224)

Vak	egy	felesége	van.
blind	emph	wife-poss.3sg	be.prs.3sg
‘He has a really blind wife.’

Based on the examples of possible and impossible “X egy N” constructions, we have to conclude that the possible candidates for the position of X are not selected by their syntactic category, but rather based on their semantics. Words denoting gradual quality despite being adjectives, nouns, or specific elements like (nem) semmi, can occur before egy. Additionally, some adverbial-like elements modifying X in the construction can occur even when X is omitted, practically occupying the position of X, resulting in constructions that express that the referent of N is particular, and extraordinary due to some unidentified reason.

7 Conclusion

This paper presents a detailed description of the emphatic “A egy N” construction in Hungarian. We followed a corpus-driven methodology paying attention to collecting all relevant data from the corpus exhaustively and basing our investigation exclusively on linguistic data. The main findings of the investigation can be described as follows.

The “A egy N” construction is quite rare in written texts. In most cases, its form is composed of an adjective followed by egy followed by a noun. We argued that egy in this construction can be identified neither as a numeral nor as an indefinite article; in this paper, it is classified as an “emphatic clitic”. It was demonstrated that the form of the construction can be generalised as “X egy N”, where X is an expression denoting a gradual property. This construction itself has an emphatic nature that does not arise from one or other part of the construction.

Utterances containing an “X egy N” express subjective value judgement, which is most frequently derogatory, and less frequently appreciating, in a speaker-related way: its use expresses that the characteristic expressed by X of a certain property (specified by N), according to the speaker, surprisingly, unexpectedly differs in measure from what he/she expected. Due to this pragmatic function, the “X egy N” construction does not occur in neutral interrogative sentences, but it is frequently used in exclamatives.

The morphology of the elements of the construction shows some variation. The form of the adjective is fixed: it cannot be comparative, or superlative. The noun can be suffixed by possessive and case suffixes if it is necessary due to the syntactic position of the phrase. According to the native intuition of the authors, the occurrence of plural nouns in the given construction is extremely marked.

The internal syntax of the construction is somewhat fixed with little variation. Instead of one adjective, in some cases, two or more coordinated adjectives occur. More frequently, the adjective is modified by an adverb, almost exclusively intensifying the adjective. In these cases, egy can also occur between the adverb and the adjective. It is not an exceptional case when the adjective is omitted after the adverb: in this case, the construction expresses the unusuality of the denotation of the noun without specifying it.

The “X egy N” construction itself can be the nominal part of the predicate (frequently with a covert pronoun subject); or as a complement of the verb, it has to be placed into the immediate preverbal position (i.e. the focus position), although it does not get exhaustive interpretation.

Further research needs to be done on the prosodic patterns of the construction, as prosody was out of the scope of the current study since it was based on written corpus data. It would be important to describe the typical intonation patterns of the construction with special attention to the differences between the prosody of the very same phrases when used for appreciation and sarcasm. As it is probably impossible or extremely difficult to find enough occurrences of the pattern in speech corpora, it would be reasonable to carry out experiments with elicitation tasks to get examples proper for a thorough analysis.

Additionally, very little is known about the evolution and development of the construction. It is not yet clear when and how the construction emerged. A key issue can be whether the development of the construction was influenced by language contact. It must be emphasised that any cross-linguistic investigation, irrespective of whether it is historically or synchronically (typologically, theoretically) oriented, makes sense after a comprehensive and thorough description of the compared constructions in all the languages studied.

Contribution

Most of the paper was written by László Fejes and Cecília Sarolta Molnár, except for Section 2 which was written by Bálint Sass, who also contributed to other sections with his remarks. Sections 3, 4, 5.1, 5.2.1 were predominantly written by László Fejes and Section 5.2.2 by Cecília Sarolta Molnár, but the final versions took shape through collaboration. The remaining sections are the result of joint work.