Linguistic Aspects
5. Sumary of results
According to the investigation, there exist certain tendencies that characterize the difference between English and Hungarian in respect of the parameters exam
ined. These can be summarised as follows.
5.1. Field differences
Sampling a section of the Longman Lexicon of Contemporary English (animal names) and comparing the items with their Hungarian correspondents found the following.
• The main difference in the organisation of hierarchical structures seems to be a frequent lack in English of a superordinate term coprresponding to such a term in Hungarian, and a higher number of hyponyms. The lack of a superordinate term in most cases has to be understood as lack of a primary, unmotivated lexeme. Appar
ently, in certain unpredictable cases English tends to provide a more detailed seg
mentation of the world than does Hungarian. In other words, one could say that ba
sic level terms in English are sometimes at a slightly lower level than in Hungarian.
• The investigation also showed that the status of STs is often uncertain in both English and Hungarian. This fact is reflected in dictionary definitions which some
times seem to be inconsistent.
• It was found that hierarchical structure is typical only of some lexical areas, and inasmuch hierarchical structure (the hyponyny relation) is regarded as synony
mous with field relations (which seems to be the case with many linguists), inter
lingual differences in field relations are not as important as decades of semantic enquiry have suggested.
Pál Heltai
• Where hierarchical organisation does show a difference, it can in most cases be linked to differences in motivation. Such differences arise of course due to the different segmentation of the world, but also interact with it. Unmotivated lexical items tend to be perceived as basic-level (specific) items, while motivated items tend to be perceived as subspecific or generic items.
5.2. Motivational differences
• English in general has a higher number of unmotivated lexical items than Hungarian and this difference is significant both in terms of language system and language use.
• The degree of motivation is higher in peripheral vocabulary in both languages, but even here Hungarian shows a higher preference for motivation.
• Hungarian uses mainly morphological motivation (derivation and compound
ing) while English tends to use word combination (collocation) which is often el
liptic and tends to afford a greater role to semantic motivation.
There are certain patterns in the differences in motivation (e.g. the correspon
dence of phrasal verbs to prefixed verbs in Hungarian), whose study should be im
portant both from a theoretical and a practical point of view.
5.3. Collocational differences
The prevailing view is that collocations are language specific and unpredictable.
However, a study of adjective - noun collocations found that certain general trends can be observed in the differences and these might very well be used for language teaching purposes.
• The number of transferred senses in the spatial adjectives studied was higher in English than in Hungarian. To some extent this can be associated with motiva
tional differences.
• In English there was a strong tendency to use spatial adjectives in A+N collo
cations to denote nonspatial relations such as number, degree, intensity, evalua
tion, etc. The same applies to Hungarian to a lesser extent.
• In A+N collocations transfer of meaning in the adjective is more often based on symbolic features in English than in Hungarian. In the latter transfer of mean
ing tends to be based on experiential features and change (transfer) of meaning beyond a certain point must have overt marking. This, again, is related to differ
ences in motivation.
• A+N collocations in English are to a greater extent elliptical than those in Hun
garian. This difference is closely related to the one described above and also to moti
vational differences.
• In English the possibility of indirect modification (the result of ellipsis) is greater.
The above findings might very well prove to be applicable to other types of col
locations (e.g. the tendency towards elliptical construction may be applicable to V+N collocations), but other collocations may also exhibit characteristics that have to be studied separately.
5.4. Differences in polysemy
Polysemy also operates along different lines in the two languages.
• Conversion and the use of the same form as both a transitive and intransitive verb in English results in polysemy. In many cases the different meanings of the English word correspond to different words (mostly derived verbs) in Hungarian;
thus, the degree of motivation in Hungarian is also higher.
• The use of abstract words to denote specific objects connected with the ab
stract concept (e.g. control) is less productive in Hungarian than in English, while compoundingoccurs more frequently. Again, the difference often manifests itself as a motivational difference.
• Abbreviation of compounds and word combinations in English to a single ele
ment is more extensive than in Hungarian, leading to polysemy on the one hand and motivational differences on the other.
Differences in polysemy show that Hungarian tends to express a new meaning through a new form created through the processes of derivation or compounding.
In other words, transfer of meaning in Hungarian receives overt marking much earlier than in English, which tends to use the same form with multiple meanings.
The differences summarised here ultimately show that the identifiable and sys
tematic differences between the lexical systems of English and Hungarian are to a large extent related to motivational differences. This is not to say that motivational differences are the underlying cause of all other differences; rather, it can be said that most differences manifest themselves in this form. From the practical point of view, however, it might be convenient to summarise English-Hungarian lexical contrasts as motivational contrasts.
6. Error analysis
Ten examination translations by 50 candidates each taking an examination in English at intermediate level at the National Foreign Languages Examinations Board in Budapest were analysed. The candidates were mostly secondary school students. In these examinations, translation is used as a test of reading comprehen
sion.
In this examination the use of the bilingual dictionary is allowed, which places a limitation on the interpretation of the results. The written translations naturally represent the product, and the source of the error and the process leading up to the product cannot be explored directly. In spite of these difficulties, it was decid
ed that attempting an analysis and an indirect interpretation was not a futile effort.
The following procedure was followed. Errors committed by at least two exam
inees were regarded as systematic errors, while single errors were ignored. The items responsible (or thought to be responsible) for the error and the unacceptable translations were recorded and classified. Only the errors identified by the original markers were taken into consideration.
The source of the error was usually indicated by differences in the translations of individual candidates. Admittedly, the source was difficult to identify in some cases, and it is quite possible that the same product from the same source is ob
tained through different processes (e.g one student might have chosen the wrong
Pál Heltai
meaning of a polysemous word, while another might have confused the word with another similar-sounding word), but this did not seem to be the typical case. A mod
erate number of doubtful cases were excluded from the data.
It is outside the scope of the present paper to present the data in full detail. A short summary of the main findings should be sufficient for purposes of illustration.
6.1. Classification of errors
Errors were classified according to the following scheme, developed on the ba
sis of Lado (1964), Arabski (1979) and Laufer (1989).
6.1.1. Errors related to form
• Similar-sounding forms (synforms): modern - modest
• False cognates: actual - aktuális
• False transparency (motivation): friend - barát, but also barátnő [friend-woman]
6.1.2. Errors related to meaning
• Polysemy: story - 1. elbeszélés [ a literary genre], 2. történet [an account], 3. cse
lekmény [plot], 4. anekdota [joke]
• Field differences
Hyponymy: table/desk - asztal Undifferentiated meaning: record - lexical gap: jetlag,pet
- culturally bound terms: pub, drugstore Synonymy (TL): like - szeret!tetszik
• Collocations
- habitual: run smoothly
- creative: green ideas [ideas put forward by the green movement]
- idiomatic: prove the case
6.1.3. Errors not related to linguistic competence
• Lack of translation skills: (translation of proper names, etc)
• Lack of background knowledge: Fleet Street
• Lack of dictionary using skills
- does not find in dictionary and cannot determine from context: vigilante, graffiti
- cannot adjust meaning found in dictionary to context: torch-light 6.2. Discussion of error analysis
As appears from the above classification, most errors were identified as deriving from interlingual contrasts. The study was not intended, however, to prove the as
sumption that in translation most errors are due to interference, neither was it in
tended to define the percentage of interference errors as against non-interference errors.
The results obtained show that the type of problem learners found most diffi
cult to handle in translation was polysemy and collocations. False cognates and idioms, wherever they occurred, generated a high number of mistranslations.
Contrastive analysis did not show which type of lexical contrast would be statis
tically the most significant. The outstanding importance of polysemy and colloca
tional problems is probably due to the fact that it is these lexical contrasts that are statistically the most numerous: practically every word in the central areas of vo
cabulary is polysemous and practically every collocation is idiosyncratic compared to corresponding collocations in another language. Even so, this finding was some
what surprising, since interpreting a polysemous item in context with the added advantage of dictionary use should not be difficult. The high incidence of such problems shows that learners have problems in decoding.
Within polysemy the translation of polysemous function words (some conjunc
tions and prepositions) and the mistakes in the translation of some well-known words as well as the problems of handling homonymy show that the lexical knowl
edge of the learners was very limited, and there was excessive reliance on the dic
tionary, unaccompanied by dictionary using skills.
Within collocations elliptical and creative collocations proved the most difficult.
The translation of collocations showed a tendency for word-for-word translation.
Often, this meant that necessary additions were not performed, which gave rise to foreign sounding collocations.
Field differences seem to have been less important because, with all the inter
lingual differences in hyponymy and synonymy, the dictionary would provide satis
factory working equivalents in most cases. Apparently, finding the right meaning of a polysemous item in the dictionary is more difficult than identifying field differ
ences, and the effects of the wrong choice of meaning are more disastrous in the former case. Thus, if the word meal in a given context could best be translated as reggeli [breakfast] into Hungarian, sticking to the literal translation might still pro
duce an acceptable translation, while choosing the wrong meaning for state (con
dition or country) would render the translation unintelligible. It is very likely that the lower number of such errors was also due to the different criteria used in the evaluation of learners’ translations, where it is only basic comprehension that is required, while stylistic and pragmatic adequacy are of reduced significance. The most difficult type of field relation was general words with a broad undifferentiated meaning, like record (What about your war record?) or lexical gaps (Let’s buy him a pet - with no equivalent for pet in Hungarian).
Motivational differences did not appear to be a significant source of error. How
ever, it should be remembered that motivational differences are often accompany
ing features of field differences, polysemy and collocational differences, therefore they were probably recorded in those categories and not in their own right. It should also be noted that these translations were evaluated from the point of view of read
ing comprehension, and in this way many covert mistakes might have passed un
noticed.
According to the analyses, the effects of background knowledge, dictionary use, dictionary using skills and translation skills were quite considerable. Altogether, these results show that the value of L2-L1 translation as a measure of reading com
prehension is severely limited.
Pál Heltai
7. Conclusions
7.1. Language learning
Semantic literature, preoccupied with sense relations, would seem to suggest that interlingual differences in relations such as hyponymy and antonymy are re
sponsible for most interference errors. The data from this research draw attention to the importance of polysemy and collocational relations at this level. The analysis also identified several specific areas for further investigation of lexical contrasts.
The data in this study suggest that we might have to return to the original idea of Lado (1964) that the greatest difficulty for the learner is one-to-many correspon
dence between L2 and LI. Whatever lexical contrasts there are between English and Hungarian, their effect can be reduced to this simple formula. One-to-many correspondence may be due to several reasons: hyponymy contrasts (H. harang and csengő [churchbell vs. handbell] corresponding to E. bell), motivational differ
ences (H. barát and barátnő [male and female resp.] corresponding to E. friend), or polysemy (H. ellenőrzés and ellenőrzőgomb [control as a general idea and con
trol as a device] corresponding to E. control).
If a hierarchical difference involves a ST in L2 corresponding to two hyponyms in LI, or a polysemous L2 item corresponding to two different words or two words related to each other by derivation in LI, or if an L2 word used in two different collocations corresponds to two different words in LI, difficulty is bound to arise for the learner. However, all this is only true for L2-L1 translation: in L1-L2 trans
lation and in speech production it is one-to-many correspondence in the L1-L2 di
rection that will cause difficulty.
In addition to one-to-many correspondence, there seems to be another factor at work. The most difficult problems for the learner and the translator are those where the difference is difficult to perceive. This is the case with many motivational and collocational differences.
7.2. Pedagogical translation
The findings summarised above suggest that learners at the intermediate level are not prepared for translation, There is a considerable deficit of linguistic compe
tence and a lack of translation skills, including dictionary using skills.
Heavy reliance on the dictionary for decoding and the high number of errors committed mean that the process of learners’ translation is dominated by decod
ing. Learners settle for a kind of semantic translation and have regard only for the referential function of the text. This feature clearly sets apart learners’ translations from professional translation.
L2-L1 translation as a means of testing reading comprehension is of doubtful value, since in evaluating it problems of decoding and encoding are often difficult to keep apart. Neither is it a valid test of the ability to translate, at least at this level of linguistic competence, since learner translators do not even address the prob
lems of that we usually associate with the word translation. The real problems of translation are often eliminated before pedagogical translation begins, through the selection of the text: texts with culturally bound terms and words corresponding to lexical gaps in the TL are avoided. The evaluation of the translation is con
strained by the fact that the translator does not even aim at stylistic or pragmatic adequacy, and it would be unjust to evaluate his product on the basis of such criteria.