Kernerman Dictionary News • Number 12 • July 2004
|
Rav-Milim – a Modern Dictionary for an Ancient but Thriving
Language
1. Hebrew is probably one of the oldest languages in
current usage, and its dictionary-making history goes back more than a
thousand years. It is not our intention here to trace the whole history of
Hebrew dictionaries or lexicographic
compendia, but rather – as a contrastive background to this brief
presentation of the Rav-Milim dictionary of Modern Hebrew (hereafter
MH) – to mention the modern ones, i.e. those in vogue in the twentieth
century before the Rav-Milim publication in 1997. By universal opinion, the scene for the Hebrew dictionaries
in the twentieth century was dominated by three major and influential works.
First and foremost is the Eliezer Ben-Yehuda 16-volume dictionary, a
monumental work of erudition and scholarship by “the reviver of the Hebrew
language”, the publication of which started in 1908 but ended only in 1959.
This is an OED-type of historical dictionary, whose glossary included not
only (though mostly) Biblical and Rabbinical terms but also whatever MH ones
were available then, especially those coined by Ben-Yehuda himself. The Gur
dictionary (1934-36) was really the first general dictionary of MH, quite
popular in the late thirties and forties. Finally, the enormously popular
Even-Shoshan one – first published in 1947-52, then reprinted countless times
in various formats and numbers of volumes (with only one major revision in
1970) – completely dominated the scene in Israel for almost 50 years, being
present, virtually, in most local households. To
complete this picture, one should mention three other dictionaries whose
impact was rather negligible: Cnaani’s 18-volume dictionary (1960-82),
Alcalay (1969-71) and Medan (1954). All
in all, then, just six dictionaries in a whole century, one of them – updated
and revised in only minor ways – exclusively dominating the scene for most of
the second half of that century, and none of them with any computerized
components. Thus, during a period when not only the State of Israel and the
Hebrew language were undergoing extraordinary dynamic cycles of changes and
expansion, but the whole world was – and still is – exploring new frontiers
(and devising new terms and semantic fields to describe them) in technology
and science, and in intellectual and social life, dictionary making in MH was
in practice frozen for some fifty years. This was the state of affairs in late 1992, when it was
decided to compile and publish – both in print and in electronic form – a new
and up-to-date illustrated dictionary of MH, Rav-Milim [Master-Words],
with a shorter companion – richly
annotated and copiously illustrated in color, specially adapted to young
children and teenagers in elementary and secondary schools – Junior
Rav-Milim. Here we restrict ourselves to the description of the
unabridged Rav-Milim printed version, its underlying philosophy and
some of its salient features. 2. Although not a purely corpus-based dictionary, the Rav-Milim
design was deeply influenced by computerized methodologies and techniques of
natural language processing developed since the mid-1980s, not only in its
production and in its extensive cross-checking algorithms, but also in its
very structure and editing method. Indeed, since the late eighties, computers
have altered the way we view dictionaries, their functionality, their aims,
and the degree of thoroughness, coverage, accuracy, precision and methodical
writing we have come to expect from them. These influences were masterly
described by Krishnamurthy in a previous issue of this newsletter (2002). True
enough, the Krishnamurthy paper was about EFL dictionaries; taking into
account, however, that for a great majority of the population of Israel,
immigrants from all over the world, Hebrew is indeed, to a certain extent at
least, a “foreign language”, such insights are highly relevant to a general
dictionary of MH as well. From
its very inception, it was decided that Rav-Milim (RM) will be
developed along completely different – in fact, radically different – lines
than previously published dictionaries of Hebrew (PPDH in short),
constituting an “anti-thesis” – so to speak – to them on almost each and
every methodological issue of dictionary designing and editing. It differs
from PPDH in the list of entries, in the entry’s structure, in the entry’s
“explanation”, in the detailed and fine analysis of the various meanings of
the entry and their order, in the usage examples, in the usage directives, in
the registers’ annotations, in the “etymological” notes, and in the thorough
and detailed processing of collocations and when, where and how to include
them. At the risk of being somewhat simplistic, we can state schematically
that RM is intended to be synchronic and not diachronic, descriptive and not
normative, explanatory and not definitional, contemporary and not archival,
illustrative and not quotation-minded. Furthermore, a maximum of uniformity
and consistency in the dictionary compilation was assured (and continuously
checked by the computer) by having all editorial questions discussed, decided
and recorded formally by the editorial committee, which counted among its
members five prominent professors of Hebrew. In the following we shall briefly present the main
features of RM, most of which were “firsts” in Hebrew lexicography, and some
of which have since been adapted in several Hebrew dictionaries that were
published after it. 3. The written form of Hebrew – as that of other Semitic
languages – is an essentially unvocalized one, vocalization being marked by
diacritical points that may appear below, above, or inside the word’s
letters. Such a vocalization is however rarely used in everyday writing,
except for Biblical texts, poetry or (more recently) for children’s books. To
alleviate some of the annoying ambiguity that would thereby result in many
different “readings” of a given word, it has been customary to add in
appropriate positions of the word some mater lectionis: Vav for
the vowels O and U, Yod for E and I, and Aleph for A, thus
producing the so-called plene script. Still, most PPDH were edited in
the formally vocalized grammatical script, and the entries were also given –
and therefore sorted – in this form. We thought that such a vocalized script
would seem totally out of context to any reader who never encounters such
texts elsewhere, not to mention its childish (on one hand) and somewhat
paternalistic (on the other hand) projection. RM is therefore edited in the
plene spelling, and the headwords are given in that script, since this is
exactly how a user will usually see it in a publication and look for it in
the dictionary. Following the plene headword, its grammatical vocalized form
is given, so as to assist the user in pronouncing it correctly and
recognizing its pattern. Additionally, a pointer is given from that form, in
its alphabetical position, to the plene one, just in case the user encounters
that form or is extrapolating from the given plene one and looking for it in
RM. Incidentally, the number of spelling variants in Hebrew is rather large,
also because of different ways of transcribing loan words from many languages
over the ages, whether from Aramaic in ancient times or mainly from English
most recently; since having pointers from these variants in the main
dictionary page would have hopelessly encumbered it (indeed many pages in
PPHD consist mostly of such pointers!), all pointers pertinent to a given
page were collected and printed in a separate section at the bottom of that
page. The
list of entries in RM is distinguished both by what it contains and by what
it omits. Besides listing virtually every (Hebrew) Biblical word and most terms
from early Rabbinical sources (except Hapax Legomena, whose meaning is
not well understood and is inferred only from the context), the list contains
every word in current usage, from all registers – from the highest literary
ones to the most colloquial and vulgar ones. The only criterion for inclusion
was whether such an utterance can be read or heard somewhere; if so, then we
must help the user understand it, by including it and its meanings in the
dictionary (this was indeed the first time ever that such terms were included
in a general dictionary). On the other hand, the word’s register is always
clearly marked; from the highly literary (to warn the reader against
using such a word in – say – asking directions) to the colloquial, vulgar
or obscene, as well as corrupt form of, etc. We
included as entries also utterances that are not, linguistically, “words” of
the language, but are used in certain ways specific to Hebrew, such as tsvits
tsvits for denoting a bird song, miaou for a cat call, koukourikou
for the rooster call, sha for requesting silence, etc. Special consideration had to be given to the inclusion
of “encyclopedic” terms and knowledge, and of terms from various scientific
and technological domains. A dictionary is neither an encyclopedia nor a
complete guide to the fauna and flora of the world or even of a certain
region of it. As a rule-of-the-thumb, any term that may potentially occur in
a general publication was included, and any term that occurs only in the
relevant professional publications was excluded. For various types of “non-linguistic” terms, the
decision on whether to include them as entries in RM was made by the
editorial committee, and rigorously implemented. Following are some examples of such decisions: ·
No proper name of anyone
(living or dead) is to be included; literary or mythological figures are
mentioned to the extent that they are used metaphorically (Samson, Venus,
Casanova) or in collocations (Richter’s scale, Columbus egg). ·
Country names are included,
along with the language(s), capital and up to three cities, and two
denominations of currency – the minimal one and the main one (cent and
dollar, penny and pound). ·
Places in Israel are
included if they have more than 5000 inhabitants as per the last Israel
census. ·
No specific “creations”
(books, theater, arts, etc) are included, with the exception of the 24 books
of the Bible and the canonical early Rabbinical sources. ·
All elements of the
cyclical table are included, with a uniformly designed explanation. On
the other hand, we omitted from RM thousands of obsolete entries that
appeared in PPHD: words coined from the late nineteenth century and loan
words from other languages that were almost never used, even words officially
coined by the Academy of the Hebrew Language that did not enjoy wide
acceptance, etc. Our policy was that not every word used once or twice by a
writer, as great as he or she may be, should be automatically recorded in the
dictionary. Delicate editorial considerations sometimes have to be applied in
such cases. Another
issue that well illustrates the spirit of RM is the following. Because of the
peculiar history of the Hebrew language, many words have persisted and are in
current usage in certain conjugated or derived forms, while the original
variant is – and was – never in use (hav [give], only in the
imperative; be'etyo [because of him/it/that], only with the
preposition and the pronoun). PPHD used, in such cases, to “extrapolate” and
invent the presumed original form and list it as a dictionary entry. We
refrained from inventing words, and such terms were given as entries “as
are”, which is anyway the form in which the user will encounter such words
and look for them in the dictionary. 4. “The principal reason for the existence of a general
monolingual dictionary is its definitions. All the art and all the
scholarship and all the scientific methods that the editors can command are
required to study meanings and write definitions” (Gove, 1961). Contrary
to Gove’s wise dictum, one cannot but notice that in most PPHD this aspect of
dictionary compilation has been quite neglected, usually with the
justification of offering one or more synonyms of the entry. In RM, however,
we fully endorsed this statement, with all its consequences and
ramifications, except, maybe, for replacing “definitions” by “explanations”,
since our aim was not to give an Aristotelian definition of an entry, but to
explain it completely and precisely. According to the RM concept, the
ultimate test of a good explanation is whether a user who has never
encountered the word before can now understand it as fully and precisely as
possible. On the one hand, we painstakingly analyzed and checked every word
in the explanation to assure its appropriateness and pertinent coverage. On
the other hand, we aimed at detailing explicitly all the nuances and shades
of the basic meaning of the entry, as manifested in the different contexts in
which it actually occurs. Indeed, as stated by Firth (1957), “you recognize a
word by the company it keeps”. One
example should suffice to clarify this approach. The adjective ham [hot,
warm] is defined in Even-Shoshan only as “having a more or less
high temperature”. In RM, this entry details some 11 different meanings or
usages in various contexts (that may well translate into different words in
other languages), which an innocent reader would not be able to guess on
her/his own. Thus, besides the basic meaning as in “hot soup” (vs. “cold
soup”), we have “hot news” (but not “cold news”), “hot temper”, “warm heart”
(the former with a negative connotation, the later with a positive one),
“warm voice” (specific voice texture), “warm clothes” (the clothes themselves
are not warm, they warm the body), “he is hot” (which doesn’t mean he has “a
more or less high temperature”, he is not sick, he just feels hot and would
like to open the window), etc. Even the “Hot! Hot!” call in the hide-and-seek
children’s game deserves and gets its own numbered meaning. Indeed, the fine
analysis of the extremely rich spectrum of the nuances of almost every word,
according to the contexts in which it appears, is one of the greatest
achievements and benefits of the application of computers to the processing
of large corpora, and the lexicographer’s efforts for collecting,
classifying, sorting and adequately explaining these nuances is probably the
most exciting and satisfying part of the dictionary making process. When
the meanings of an entry have changed throughout its history, they were
always ordered, traditionally, chronologically. In RM, which has always had
the user in mind, meanings are ordered by decreasing frequency; the
most frequent sense given first, and adequate period labels attached when
necessary. Finally,
an explanation is almost always followed in RM by one or more example of
usage, which only rarely are quotations from canonical writing. In nearly all
cases, examples were carefully crafted to add interesting and useful details
to the explanation. 5. One of the impacts of large corpora processing on
linguistic studies in general, and on dictionary making in particular, since
the mid-eighties, has been the recognition of the critical importance of
collocations in defining the language elements and structure. If
this is true for European languages, how much more so for Hebrew! Indeed,
with the world dynamically revolving around us, the Hebrew language has
constantly had to acquire and absorb numerous new words from the various
domains of modern life activities. Although some new terms are adapted as
loan-words “as is” and easily become part of current Hebrew, in many cases,
however, Hebrew – being a Semitic language with a structure of 3- (or 4-)
letter roots and derivation patterns – is quite resistant to such assimilation.
A common productive solution is to have a two- (or three-) word Hebrew
sequence to represent a new concept. A large number of single-word nouns in
English, for example, such as school, hospital, lawyer, accountant,
are represented in Hebrew by a two-word sequence. In
spite of that, the treatment of collocations in PPHD has been rather poor, to
say the least. Very few collocations found their way into these dictionaries;
phrasal collocations, idioms and even proverbs (!) were all mixed up; no
clear guidelines were respected in terms of where and how to have the
collocation’s main entry (in fact, in an extreme example, a 4-word
collocation could appear in 4 different entries with 4 different
explanations!), or in terms of how to deal with and uniformly represent the
“empty places” in some of these collocations, etc. Having
researched the problem of collocations already in the eighties (see 1983,
1988), I was strongly biased in favor of a comprehensive, systematic,
rigorous and consistent treatment of the collocational part of RM. A small
sample of the new features introduced in this endeavor now follows: ·
To the question of when
does a sequence of two or more words deserve its own entry in the dictionary
as a collocation, a common answer is: when the meaning of the sequence
is not the total sum of its components’ meanings, and cannot be guessed from
it. This is indeed an important criterion, but it is far from being unique.
We delineated 12 different criteria that can justify such an inclusion, and
every potential collocation was tested accordingly. ·
Almost 10,000 new
collocations were added in RM that never appeared before in PPHD. This is an
extremely high figure when taking into account that the total number of
(single-word) entries in PPHD is of the order of 35,000 entries only. ·
Proverbs (e.g. ‘not all
that glitters is gold’) were completely banned from the dictionary; phrasal
collocations and idioms were sharply separated. ·
Strict rules were set up
and followed on where to introduce the main entry of a collocation and its
explanation. The explanation appears, of course, only once, but pointers to
that occurrence are given from every word of the collocation. ·
Collocations were tagged
by part-of-speech tags: nominal, verbal, adjectival, adverbial, etc. When necessary,
morphological variants were added. ·
Possible additions,
omissions, replacements, etc, in the collocation text were marked clearly, in
a uniform way.
To
sum up: RM was a bold step taken to bring modern methodologies, trends and
techniques to Hebrew dictionary making, applying overwhelmingly a
computerized approach to its compilation and checking procedures. We believe
that it has thus set a new standard of precision, coverage, methodology and
systematization that will be hard to ignore. Dictionaries
Alcalay, Reuven. 1969-1971. Milon 'Ivri Shalem (The Complete Hebrew
Dictionary). 3 vol. Ramat-Gan: Massada. Ben-Yehuda, Eliezer. 1908-1959. Milon
ha-Lashon ha-'Ivrit ha-Yeshana ve-ha-Hadasha (A Complete Dictionary of
Ancient and Modern Hebrew). 16 vol. Cnaani, Yaacov. 1960-1982. Otsar ha-Lashon ha-'Ivrit (Treasure
of the Hebrew Language). 18 vol. Ramat-Gan: Massada. Even-Shoshan, Avraham. 1947-1952. Ha-Milon he-Hadash (The New
Dictionary). 5 vol; 1966-1970, 7 vol; 1997, 5 vol. Jerusalem: Qiryat
Sefer. Grazovsky (Gur), Yehuda. 1934-1936. Milon ha-Safa ha-'Ivrit (Dictionary
of the Hebrew Language). 3 vol; 1947, 1 vol. Tel Aviv: Dvir. Medan, Meir. 1954. Me-'Aleph 'ad Tav – Milon 'Ivri Shimushi
(From A to Z – a Practical Hebrew Dictionary). Jerusalem: Achiasaf. Other References
Choueka, Y., S.T. Klein and E. Neuwitz. 1983. ‘Automatic retrieval of frequent idiomatic and
collocational expressions in a large corpus’. ALLC Journal, 4.34-38. Choueka, Y.
1988. ‘Looking for needles in a haystack or: locating interesting expressions
in large textual databases’. Proceedings of the RIAO Conference (Cambridge,
MA). 609-623. Firth, J.R.
1957. ‘A Synopsis of Linguistic Theory, 1930-1955’. Studies in Linguistic
Analysis, 1-32. Oxford: Blackwell. Gove, P.B. 1961.
‘Linguistic Advances and Lexicography’. Word Study, October 1961, 3-8. Krishnamurthy, R.
2002. ‘The Corpus Revolution in EFL Dictionaries.’ Kernerman Dictionary
News, 10.23-27 (http://kdictionaries.com/newsletter/kdn10-9.html). Rav-Milim
K Dictionaries Ltd |
