Page 1LIETUVIŲ KALBOTYROS KLAUSIMAI XXXIX (1998)
Vida ŽILINSKIENĖ
Matematikos ir informatikos institutas, Vilnius
A FREQUENCY DICTIONARY OF CONTEMPORARY WRITTEN LITHUANIAN
(first research results)
A computerized frequency dictionary of contemporary written Lithuanian is already in operation, compiled from material representing four functional styles (fiction, journalism, science, and technical writing). For more information about the dictionary’s methodology and principles, see the article (Grumadienė 1995).
The lexical diversity of a language, a particular functional style, a writer, or a work is assessed by the so-called iteration coefficient (the ratio of word occurrences to words), which also represents the average frequency of a word in the text. Calculations show that the iteration coefficients for the styles mentioned above and for the Lithuanian language as a whole are:
fiction – 9.53,
journalism – 10.51,
science – 12.63,
technical writing – 20.78,
the language as a whole – 20.06.
The lower this coefficient, the more different words are found in texts containing the same number of word occurrences (in our case, 300,000 word occurrences from texts of each functional style were examined), and the more diverse the vocabulary of that style. The figures given clearly show that: 1) the vocabulary of fiction is twice as diverse as that of technical writing; 2) the lexical diversity of fiction, journalism, and even scientific writing is fairly similar.
We can compare the diversity of word forms found in the texts examined by comparing the ratio of word occurrences to word forms in each style:
fiction – 4.34,
journalism – 4.48,
science – 4.83,
technical style – 7.29,
the entire language – 7.16.
These figures also clearly show that: 1) a greater variety of grammatical forms is used in literary, journalistic, and scientific styles; 2) in this respect, the technical style also differs considerably from the other styles.
219
Page 2Since we have a frequency dictionary of Lithuanian journalism prepared using the same methodology (Žilinskienė 1990), calculated from newspaper and magazine texts from 1973–74, and a frequency dictionary of Latvian journalism (Jakubaitė 1969), calculated from texts from 1966–67, we can compare their iteration coefficients:
Lithuanian journalism, 1993–95 – 10.51; Lithuanian journalism, 1973–74 – 15.38; Latvian journalism, 1966–67 – 18.60.
These coefficients clearly show that the vocabulary of contemporary Lithuanian journalism is more diverse than that of Lithuanian journalism from twenty years earlier or Latvian journalism from almost thirty years earlier. Let us compare the ten most frequent words in the individual styles and in the language as a whole (see Table 1; in this and the other tables, the number next to a word indicates its frequency).
Table 1. The ten most frequent words in individual styles
fiction | journalism | science | technical style | entire language and* 11442 | and* 9743 | and* 11140 | and* 10990 | and* 43315 he 9634 | be 5825 | be 6294 | be 5138 | he 24613 I 4580 | he 5824 | he 4692 | he 4463 | be 21366 to 4216 | that 3487 | that 2783 | that 3286 | that 13528 be 4109 | that* 3235 | this 2138 | which 2273 | to 10715 that 3972 | to 2719 | which 2054 | Lithuania 2195 | that* 9877 or* 2822 | I 2602 | to 1941 | this 2135 | I 9025 that* 2811 | which 1985 | that* 1844 | that* 1992 | which 7610 how 2219 | or* 1904 | with 1838 | whether 1903 | from 7162 what 2136 | with 1897 | from 1647 | to 1839 | with 7034
The frequencies of the words marked with an asterisk (*) in Table 1 differ from those of the same words given in the dissertation (Žilinskienė 1995), because here the overall frequency of the particular lexeme is given, whereas the dissertation gives the frequency of the corresponding lexeme when it occurs in the studied texts only as a conjunction.
Table 1 clearly shows that: 1) the first ten most frequent words include function words, pronouns, which from a statistical point of view generally behave like function words, and the verb to be, which is often used as an auxiliary verb in texts; 2) in the first ten most frequent words—
220
Page 3In the top hundred, six words (and, he, be, that, that, into) are the same; only their ranks (position numbers in the frequency lists) differ, except, of course, and. 3) There is only one word with a specific meaning, Lithuania, in the top ten most frequent words of the business style.
Let us compare the twenty most frequent words in contemporary Lithuanian journalism, earlier Lithuanian journalism, and Latvian journalism (see Table 2).
Table 2. Most frequent words in Lithuanian and Latvian journalism
1993–95 journalism | 1973–74 journalism | Latvian journalism | and 9743 | and 11890 | un 10401 | be 5825 | he 5733 | būt 6395 | he 5824 | be 4167 | that 5244 | that 3487 | that 3939 | or 3631 | that 3235 | into 2614 | par 3546 | into 2719 | I 2461 | what 3119 | I 2602 | that 2455 | viņš 2610 | which 1985 | which 2394 | no 2342 | but 1904 | with 2046 | ka 2325 | with 1897 | this 2024 | ari 2304 | from 1890 | all 1825 | this 2275 | Lithuania 1839 | work 1759 | darbs 2059 | this 1715 | from 1602 | mēs 2030 | all 1517 | how 1594 | gads 1789 | own 1463 | year 1563 | bet 1643 | but 1273 | but 1538 | uz 1630 | person 1271 | person 1421 | kā 1609 | or 1213 | not 1389 | viss 1601 | year 1302 | only 1277 | padome 1386 | about 1165 | own 1263 | savs 1349
Comparing Lithuanian journalism from the two periods, we see that the frequency of the most frequent words is very similar; only three words (but, or, about), which appear
221
Page 4in the top twenty of contemporary journalism, have shifted lower in earlier journalism (but is in 22nd place, or in 39th, and about in 26th) (Žilinskienė 1990), while the word Lithuania was encoded as a country name in earlier journalism, where only the total frequency of separate groups of proper nouns (e.g., first names, surnames, country names, hydronyms, etc.) was counted; therefore it is not listed in the dictionary of words ordered by decreasing frequency. When processing contemporary journalism, savo was treated as the genitive of the pronoun savas, and, naturally, the lexeme savas was then recorded; recalculating earlier journalism in this way, savas would have had a frequency of 1461 and ranked 17th. In addition, it should be noted that nouns have already appeared in the second ten most frequent words (year, person); only work has moved into the third ten in contemporary journalism.
Comparing with Latvian journalism, we can see that most of the same function words appear in the top ten, as do the same two most frequent nouns; only the Latvian padome (council) is still very frequent (in earlier Lithuanian journalism, taryba (council) occurred 455 times, and in contemporary journalism 137 times).
Let us compare the nouns, adjectives, numerals, verbs, and adverbs in the first hundred most frequent words in all four styles and in the language as a whole (see Table 3).
Table 3. The most frequent nominal words, verbs, and adverbs in individual styles
fiction
journalism
scientific
business
entire language
person
Lithuania
work
Lithuania
Lithuania
eye
person
system
work
work
hand
year
Lithuania
document
person
head
work
person
article
year
day
country
research
republic
law
time
day
part
law
document
house
period
period
person
time
word
state
process
company
day
man
earth
element
law
state
time
Vilnius
material
chairman
republic
work
time
water
state
part
child
life
time
institution
law
father
school
year
organization
article
leg
Russia
culture
language
company
222
Page 5fiction time woman life window big good journalism child Lithuanian company time bank law world house economy Europe part city big new good important scientific data Lithuanian condition method number meaning case language science form group surface method result activity problem place property big various main new important specialized part word deputy court commission question case resolution order goods council country employee member person case government violation measure day project bank meeting form contract time price state high administrative main important all styles word language country person system time earth time place life big new high
223
Page 6fiction journalism scientific specialized all styles general social one two one first two three one first two one first two one first two be be able not be have say know go want take need say look be unable see seem be have be able not be say need be be able have not be constitute use determine be have be able need not be envisage want be be able have not be need know determine want constitute say receive be unable so very much where here there well then very now yes therefore here still already very much therefore often also much so now therefore often very much still so already now here also where therefore only
224
Page 7As might be expected, the first hundred most frequent words show that fiction differs first and foremost from non-fiction styles: 1) fiction has the fewest nouns (18) and adjectives (2), and twice as many verbs (15); 2) journalism is closer to fiction both in the proportions of parts of speech and in terms of specific lexemes (e.g., house, time, child, life, good, say, here occur among the first hundred most frequent words only in these styles). Several lexemes are shared by the first hundred most frequent words in all styles: person, work, time, be, be able, not be, have. From the first hundred most frequent nouns alone, it would already be easy to tell which style the words were selected from. For example, it is impossible to imagine that words such as research, process, element, data, condition, meaning, form would be frequent in fiction, just as it is impossible to imagine that eye, hand, head, child, father and the like would recur frequently in scientific-style texts. Such words can be frequent only in some particular sample in which the writing is specifically about that subject. Moreover, it is quite obvious that the verbs in the first hundred most frequent words have very general meanings and function mainly as auxiliary verbs in texts.
Let us compare the nouns, adjectives, numerals, and verbs occurring in the first hundred most frequent words in contemporary and earlier Lithuanian journalism and Latvian journalism (see Table 4).
Table 4. The most frequent nouns, adjectives, and verbs in Lithuanian and Latvian journalism
1993–95 journalism
1973–74 journalism
Latvian journalism
Lithuania
work
work
person
year
gads
year
person
council
work
school
person
country
country
time
day
time
party
occasion
republic
nation
state
economy
state
land
day
day
Vilnius
life
republic
time
film
earth
life
district
place
school
time
debt
Russia
city
factory
child
science
district
Lithuanian
earth
life
225
Page 81993–95 publicistic writing
1973–74 publicistic writing
Latvian publicistic writing
enterprise generation bank law world house economy Europe part city
child people council question world place time teacher production organization sport student woman
city question organization occasion committee enterprise world word struggle task lesson strength worker
large new good important
new large good Soviet important young high socialist real
young large good high
one first two three
one first
first one two second
to be have be able not be say need
to be have be able need not be know work speak want say
to be not be be able become happen work begin give
This comparison shows precisely which lexemes are most characteristic of publicistic writing, that is, which lexemes are frequent both in the Lithuanian publicistic writing of the two periods and in Latvian
226
Page 9in the press. Finally, certain frequent lexemes (e.g., only in the current press have bankas, įstatymas already entered the list of most frequent words, whereas in the earlier press these were tarybinis, socialistinis) reflect the political and economic aspects of the corresponding period.
This study used material from the computerized Frequency Dictionary of Contemporary Written Lithuanian, compiled with the support of the Lithuanian Science and Studies Foundation.
Received 1998 06 26
LITERATŪRA
Grumadienė L. 1995: Dabartinės rašomosios lietuvių kalbos dažnumų žodynas: sandara ir darbo eiga. – Lietuvių kalba: tyrėjai ir tyrimai. Konferencijos pranešimų tezės, Vilnius: Lietuvių kalbos institutas, 18–19.
Jakubaitė T. 1969: Jakubaitė T., Gulėvska D., Ozola V., Prūse R., Rubina A., Sika N. Latviešu valodas biežuma vārdnīca. 2, Rīgā: Zinātne.
Žilinskienė V. 1990: Lietuvių kalbos dažninis žodynas, Vilnius: Mokslas, 176.
Žilinskienė V. 1995: Rašomosios lietuvių kalbos dažninis žodynas. – Lietuvių kalba: tyrėjai ir tyrimai. Konferencijos pranešimų tezės, Vilnius: Lietuvių kalbos institutas, 33–34.
HÄUFIGKEITSWÖRTERBUCH DER LITAUISCHEN SCHRIFTSPRACHE
(erste Forschungsresultate)
Zusammenfassung
Dieser Beitrag bietet die Einschätzung der Häufigkeit der Wörter in der Lexik der verschiedenen funktionalen Stile der gegenwärtigen litauischen Schriftsprache nach Iterationskoeffizienten, wobei auf das Verhältnis der Wortformen in den Stilgattungen hingewiesen wird. Verglichen werden, die gebräuchlichsten Wörter in den Stilgattungen und die gebräuchlichsten in der litauischen Publizistik 1993–1995 bzw. 1973–1974 sowie in der lettischen Publizistik. Der Vergleich zeigt, welche Lexeme für die Publizistik besonders kennzeichnend sind und wie sie die politischen bzw. wirtschaftlichen Aspekte einer gewissen Zeitperiode widerspiegeln.
227
