What today’s dictionary users expect of a dictionary is to be able to find any word they need and in the shortest possible time. Users of bilingual dictionaries may wish to find information, depending on the character and size of the dictionary, on some or all of the following: spelling; pronunciation; part of speech status; compounds; word forms; syntax; usage and style labels.

But do non-linguist users, who are presumably the majority, need all of that information all the time? Our contention is that users benefit more from finding just some of that information but in an accessible way, and in this article we illustrate that neither is sheer bulk what really matters, nor is a high degree of technical detail necessarily helpful. We show, in particular, that the two attractive-looking volumes recently published by Akadémiai in the ‘Klasszikus Nagyszótárak’ series (reference to which will be made in the form ‘Klasszikus’) are at best thought of as targeting the linguist community rather than a general readership. Also, that they may be called ‘brand new’ only in some promotionalese, since they have not been revised from the time-honoured volumes initially edited by L Országh (Országh 1960, 1963), just enlarged by adding new material.

The ‘Klasszikus’ dictionaries do contain a wealth of data, but finding it may prove a taxing job. This, to our minds, stems from three sources: (a) from an overly hierarchical structure of the entries themselves; (b) from the multitude of (often totally redundant part of speech, specialist, and style) abbreviations; and (c) from the distinction between polysemy and homonymy as an organising principle of the entries. When those three topics have been covered, we will briefly touch on another four: (d) the treatment of compounds; (e) the traditional grammatical point of modal auxiliaries; (f) the ordering of senses within entries, and (g) the illustrative material.

(a) Within entries, the Roman numerals that mark off parts of speech are sometimes on a new line and sometimes appear continuously, in an unpredictable fashion. A random check of pages 1412–1413 of the English–Hungarian volume showed, for example, that in the entries for so and social, different parts of speech start on new lines, while in the soak, soaking, soap, sob, sober and sociable entries, part-of-speech structuring is done continuously. In these latter cases, entries are huge masses of words, making it very hard for the user to locate the desired equivalent. Within verbal entries, marked by upper-case Roman numerals, the separation of transitive from intransitive uses is achieved somewhat redundantly (also an inherited feature from the old editions), both by the letters A. and B. and by abbreviations. The depth of the entries’ hierarchy is also increased by the Arabic numerals for the individual equivalents (these, of course, are hardly dispensable), and the lower-case a), b) etc. letters within them.

(b) The ‘Klasszikus’ dictionaries employ a host of abbreviations: 146 in the English–Hungarian part. Some of those, including part of speech and grammatical abbreviations (over 30 altogether) like fn, mn, hsz  [n., adj., adv.] etc. are self-explanatory. While nobody challenges the need for unambiguous grammatical abbreviations, the abbreviations tsi and tni presented a problem even to many of those advanced dictionary users who we asked. (If your native tongue is English, don’t worry; if you’re Hungarian and still clueless, we’ll tell you: it’s tárgyas ige [transitive verb] and tárgyatlan ige [intransitive verb]. The reason here is that this way of creating abbreviations – first consonant followed by final consonant of a word – is simply non-existent in Hungarian and consequently not used outside the ‘Klasszikus’ dictionaries.

The separation of transitive from intransitive, incidentally, is also problematic. It will suffice to quote one example, the nominal entry for face, where the window faces the garden is found among the transitive uses of the verb, while the house faces north is listed among the intransitive meanings. What often happens in such cases is a confusion between the transitivity of the English verbs and that of their Hungarian equivalents. Here, obviously, both occurrences of face are transitive (and both Hungarian equivalents – a kertre/északra néz – intransitive).

Another abbreviation riddle is the pair htlan ne  and htt ne. It will make the Hungarian-speaking reader’s job easier to know that of the former, there is one in the Hungarian–English and two in the English–Hungarian part, while of the latter, it’s the other way round. Yes, htlan  is határozatlan [indefinite] and this, fortunately for the dictionary, tells the reader what ne may be, since határozatlan and határozott [definite] only collocate with one item: article. 

The pair sztlen i (which signifies személytelen ige [impersonal verb]) and vhi  (which stands for visszaható ige [reflexive verb]) illustrates a bigger problem. It is bad enough to keep the reader in ignorance about English grammar by suggesting that it actually has impersonal and reflexive verbs – you don’t find claims like that in grammars published over the past fifty years. We had an idea as to what such an impersonal verb might be: probably verbs are meant that are only used with dummy subjects, such as rain and snow, i.e. whose subject can only be the pronoun it. Luckily, however, rain and snow are not marked as sztlen i  in the dictionary. As to reflexives, the first three candidates for reflexive-verb-hood that enter your mind – dress, wash and comb – are not marked as being reflexive with the abbreviation sztlen i. What must have happened here is that conscientious editors have removed vestiges of impersonal and reflexive verbs from the dictionary (on their own initiative?), while the old v impers and v refl abbreviations themselves live on, in changed form, in the List of abbreviations of the ‘Klasszikus’ dictionary.

Coming back to abbreviations: a minor but rather disturbing error is that the English pt, pp, and pt/pp  appear among the otherwise uniformly Hungarian ones. In the English–Hungarian part about 70 specialist labels and nearly 20 style labels are used, some of which – including áll and állatorv, for ‘zoology’ and ‘veterinary medicine’ – are rather confusing. 

Let us take one example from the Hungarian–English part as well, which again shows something much graver than the doubtful intelligibility of an abbreviation: here, the word hragos stands for határozóragos [with an adverbial suffix]. Not only does this look like a misprint (haragos? [angry]), but tagged as it is to certain adverbs, it conveys completely redundant information on the morphology of adverbs. An item like lassan is thus tagged as mn hragos, and this would be a rather perverted way of saying ‘adverb’, even if you accepted (as we certainly do not) that lassan ‘slowly’ is an adjective. Lassan  is not a mn.

There is no typographic distinction between the specialist and the style labels, which is sometimes more than confusing (as in the case of rég and régész – these stand for ‘archaic, obsolete’ and ‘archaeology’). On a general plane, specialist labels are often superfluous. Of the style labels used, the abbreviation biz must be singled out as one of the most persistent ones since the days when the first colloq  in the Concise Oxford Dictionary (COD) (1934) of the time, ostensibly the source of many later Hungarian published dictionaries, got translated as biz. The trouble here, of course, is that the English labels colloquial/informal suggest something very different from the Hungarian bizalmas [familiar, intimate]. While anything bizalmas is probably best avoided by the language learner, colloquial/informal  denote a type of usage that ought be encouraged rather than suppressed. 

This general mismatch is compounded by the fact that there are no visible signs of the style labels having been revised. This is very alarming because it means that many items that were marked as colloq in monolingual English dictionaries around 1930 bear the same label disguised as biz in these English–Hungarian dictionaries designed for the year 2000 and upwards. The entry for guy in the English–Hungarian part of the ‘Klasszikus’ dictionary is a case in point. Under meaning 2. you find the labels US, biz and ten Hungarian ‘equivalents’, including some that may be argued to be slang: alak, pofa, pasas, pacák, szivar, ips(z)e, muki, tag, krapek, fej. Most of these are not natural Hungarian translation equivalents for guy, which is probably the most natural word for the concept ‘male human being’ (on both sides of the Atlantic, not just the US side). It is alarming to see that the ‘Klasszikus’ dictionary encourages the view that an English-speaking person using the word guy actually uses a style suggested by these ten Hungarian words.

Especially confusing cases result from sequences of three or more abbreviations (in identical font/typeface), such as narky – mn táj szl – that is, adj dial slang. (It is a question, incidentally, whether ‘dialectal’ and ‘slang’ being used together is a contradiction). A related problem is when the labels and the body of information are printed with the same type (italics), as in the following entry:
 

sgml röv fn infor standard generalised markup language

Apart from the fact that part of speech tags are redundant in the case of all abbreviations (and infor is not indispensable either), the italicised word standard  is easily mistaken for a style label. 

To sum up: the ‘friendly way’ of handling abbreviations within a generally satisfying solution, to our thinking, is to cut down their number to a minimum. This can be done, for instance, by omitting the style label ‘taboo’ or ‘obscene’ wherever the Hungarian translation matches the source language in terms of style.

(c) It is an almost unchallenged method in both mono- and bilingual lexicography to distinguish polysemous words, i.e. words with the same shape, with related meanings and related etymologies – such as break ‘fracture’ vs. ‘interval’ – and homonymous words, i.e. words with more or less obviously differing senses and etymologies – such as bank (of a river) vs. bank (financial institution). While words displaying polysemy are numbered within one and the same entry, homonyms are assigned entries of their own, usually marked by superscripts: bear1 and bear2 . The unwary dictionary user who is looking for a verb  is thus often forced to read through a long entry containing nouns, and in addition to sometimes losing valuable time, they may not find the relevant sense; superscripts may go totally unnoticed. What we feel more important, and more of principle, however, is that most non-linguist readers are not interested in any two words being etymologically related or not – they merely want to look up a word. Consequently, it makes little sense for general-purpose dictionaries to (use superscripts and) distinguish between polysemy and homonymy2. By contrast, distinguishing according to part of speech (i.e., structuring entries in part-of-speech terms), is of direct practical use – but that information is very difficult to gather from the entry structure of traditional-style dictionaries including the ‘Klasszikus’.

To illustrate the three criteria above, we print the entries for tip in the English–Hungarian ‘Klasszikus’ dictionary overleaf. 
 

Ironically, even the etymological basis of the homonymy vs. polysemy distinction may be missing in this particular case, since the different origins of the two senses under the superscripts are not proven (cf. Klein, 1971) – as we suspect is the case with many other such pairs.

(d) The treatment of compounds is a notorious lexicographic problem, especially in English, where decisions concerning the (solid, hyphenated, or separate, ‘open’) spelling of compounds are not easy to make and where (partly following from this) these decisions have important consequences for the length, structure, and ordering of entries. Solid and hyphenated items obviously have entries of their own, but what about open compounds? If you decide that capital punishment ought to belong within the entry capital, then all other relevant capital compounds will have to be treated that way. Alternatively, capital punishment gets a self-contained entry, and words like capitalist, capitalisation, capitalise find themselves wedged between capital and capital punishment. We believe that irrespective of their spelling, these compounds are independent lexical units and thus deserve separate entries.

The ‘Klasszikus’ English dictionary does not treat compounds uniformly. They sometimes turn up within entries, sometimes as independent entries (in the manner that we embrace). Of the capital compounds, for example, the following are listed (among others) in the capital entry: capital letter, capital offence, capital punishment (where capital is an adjective), capital account, capital expenditure, capital surplus (with the noun capital). That this is not a principled decision is shown by capital assets, capital gains, and capital transfer tax being in separate entries. We do realise that a completely uniform system is probably a mere ideal, but a complete lack of principled uniformity is a serious setback even if more than a dozen editors have contributed to the ‘Klasszikus’.

(e) The handling of the modals could, might, should and would shows that in some respects the 1998 ‘Klasszikus’ English dictionary reflects early 20th-century attitudes to language/English. It is hard to believe that these modals should have no entries of their own, only a referring entry to the modals can, may, shall and will, automatically implying that the former are past versions of the latter. There might be a sense in which this is true (at least for can–could), but past reference is hardly the most important feature of could, might, should and would. (It never has been during the history of these modals in Modern English). When the learner comes across sentences like Could you open the door? or You really shouldn’t do that, this dictionary will never guide them to the meanings of could and shouldn’t 3. The (understandably inflated) entries of can, may, shall and will themselves, in turn, contain mixed examples in terms of tense, making complete the implication that might is to may exactly as the past form of any lexical verb is to its present form – a badly mistaken one, we believe.

(f) Depending on dictionary type, the ordering of senses may be based on frequency and go from typical to rare (in a learner’s dictionary), or on chronology, with old senses preceding new ones (in a historically based dictionary). While the ‘Klasszikus’ is between these extremes, it is arguably closer to an ideal whereby new and frequent have priority over old and idiosyncratic. It is a widely known fact about dictionary users that they often stick with the very first sense in an entry. Many dictionaries explicitly boast about placing frequent senses early in the entry for easier access. If now you look at the entry for guy and runway bearing all that in mind, it is striking how the ‘Klasszikus’ dictionary is more like a historically based dictionary and less like an up-to-date bilingual dictionary. Rather than list, e.g. pasas for guy and kifutópálya for runway as the No.1 senses, ‘Klasszikus’ opens the entry with madárijesztõ, nevetséges figura, ijesztõ figura for guy, and with folyómeder, surrantó, visszagurító csatorna and horony  for runway. All of those, we contend, are senses of runway that very few English speakers are aware of, while two are Hungarian words that hardly any speaker knows. Needless to say, this ordering of senses has been with us since the first edition of the old ‘Országh’ dictionary.

(g) Not much has been changed in terms of illustrative material either. Apparently, there have been editors responsible for a particular letter who carefully combed through their lexical material, while most of them just added new headwords without even touching the existing entries. For a dictionary like this to be advertised as coming from ‘one of Europe’s most sophisticated lexicographic centres’, updating the illustrative phrases/sentences would be a basic requirement. Two examples will illustrate this. The first example sentence in the entry for presume in the old Országh dictionaries and in the ‘Klasszikus’ differs in one comma only: 
 

Országh (1960): Doctor Livingstone I ~ ?
‘Klasszikus’ (1998): Doctor Livingstone, I ~ ?
Why just Dr. Livingstone and not Dr. Watson or Mr. Holmes? 

In a sentence exemplifying trial, there isn’t as much as a comma’s worth of difference between the old Országh dictionaries and the 1998 ‘Klasszikus’:
 

Have you read the trial of Warren Hastings?

Examples like these raise the question about the function of illustrative examples in general. Put simply, they either fine-tune some meaning, or provide some syntactic information, that is, they are useful in two cases: when two or three word-level equivalents may not be enough, or when some syntactic information is vital to understanding. In all other cases an overabundance of examples is counter-productive.

What, then, is a dictionary like which comes up to the expectations of the 21st-century user? In addition to the requirement mentioned at the beginning, that it should be comprehensive, it should be user-friendly in the sense of catering to the needs of non-professionals who are rarely great readers of long forewords and complicated introductions even if they suspect that a lot of useful information awaits them there. All the really important information has to be in the body of the dictionary. It is a useful dictionary when the structuring follows clear and psychologically grounded principles, when hierarchies as well as codes and abbreviations are kept to a minimum, and when all of this is apparent in layout and typography. It is convenient to use when each page reminds of the most general conventions. Are the two ‘Klasszikus’ volumes like that? No, they aren’t. Is there a dictionary available like that? Not to our knowledge. Those in favour of a more optimistic ending, however, are referred to  Note 1.
 

Notes

1. The authors are currently working on a new-concept English–Hungarian and Hungarian–English dictionary.(back)
2. Today, the theoretical grounds of the homonymy vs. polysemy distinction are also seen as shakier than ever, and it is not obvious that the linguistic competence of native speakers makes such a difference between different senses of a word shape like tip.(back)
3. Sadly, and quite surprisingly for a learner’s dictionary, the 1984 impression of the Oxford Advanced Learner’s Dictionary of Current English (OALDCE) still has no entry for might. By contrast, the 1st edition of the Longman Dictionary of Contemporary English (LDOCE) in 1978 lists might, with eleven senses, separately from may, which itself has eight.(back)

References

Collins English Dictionary and Thesaurus. (1992). Collins Electronic Dictionary Data. Harper Collins Publishers.
Concise Oxford Dictionary (COD). (3rd ed.). (1934).  Oxford: Oxford University Press. 
Klein, E. (1971). A Comprehensive Etymological Dictionary of the English Language. Amsterdam: Elsevier Publishing Company.
Longman Dictionary of Contemporary English (LDOCE). (1978). Harlow: Longman. 
Longman Dictionary of English Language and Culture (LDELC). (1998). Harlow: Addison Wesley Longman. 
Országh, L. (1963). Magyar–angol szótár. Budapest: Akadémiai Kiadó.
Országh, L. (1960). Angol–magyar szótár. Budapest: Akadémiai Kiadó.
Országh, L. & Futász, D., & Kövecses, Z. (1998). Magyar–angol nagyszótár. Klasszikus Nagyszótárak. Budapest: Akadémiai Kiadó. 
Országh, L., & Magay, T. (1998). Angol–magyar nagyszótár. Klasszikus Nagyszótárak. Budapest: Akadémiai Kiadó. 
Oxford Advanced Learner’s Dictionary of Current English (OALDCE). (3rd ed.). (1984). Oxford: Oxford University Press.