RESULTS AND DISCUSSION

8.1 The questionnaire

In general, we can say that all the participants have been found to possess an extended knowledge of depth of word categories, as would be expected of such advanced language learners. On the one hand, the translator trainees gave an impressive array of meaning senses and different word forms, which they are able to use productively. On the other hand, in the case of teacher trainees, there were more instances of collocations. However, in the present study, the aim is to shed light on the associative behaviour of our participants, which will be the focus of the next sections.

8.1.1 Language use

The first research question related to language use yielded very interesting results, which we can see in Table 1. The ratio of responses given by translator trainees in Hungarian for an English prompt word is very high: 392 out of 471, 83%. (See the results with regard to the ten prompt words for the two groups in Appendices 2 and 3.) Furthermore, a significant number of these associations were different meaning senses of the prompt word. For example, for the word paper, the following L1 equivalents were provided as associations: papír, írás, újság, tanulmány, irat, okmány, dokumentum, értekezés, which implies that translator trainees put a lot of emphasis on this aspect of word knowledge. The teacher trainees associated mostly in L2, as would be expected of advanced learners: 90% of their associations (444 out of 494) were in English.

This was in contrast to H1. Whereas it seems to us that for teacher trainees, the use of L2 is separated from that of L1, our group of translator trainees relied heavily on their mother tongue when providing associations, which implies that our findings are in line with the mixed model presented earlier and points to a stronger than usual link between L1 and L2 words. In a research study conducted earlier, even intermediate learners resorted to L2 associations (Dóczi – Kormos 2016), which showed a strengthening of links in L2. Nevertheless, in our case, it was clearly visible that these links not only do not disappear but are actually made stronger due to constant practice.

The rationale behind this model is that even as learners’ level of proficiency increases, the link between L1 and L2 words does not necessarily disappear; in fact, with practice in translation these links may even become stronger and, in this sense, translators are a special group of proficient

language users. This is in line with the finding of Navracsics (2007), who called these responses

“lexical equivalents” or “cross-linguistic synonyms” (p. 32) and stated that learners activate the lexical store of both languages when associating.

Table 1. Differences between associations of translator trainees and teacher trainees

L1 / total L1 paradigmatic L2 / total L2 paradigmatic

translator trainees 83% 65% 17% 50%

teacher trainees 10% 66% 90% 46%

8.1.2 Association patterns in the two groups

In an attempt to confirm or refute H2, we analysed the ratio of the different types of associations in the two groups. In the elicitation process altogether 471 associations were produced by the translator trainees, whereas the teacher trainee group provided 494 target words. In the following, we intend to analyse the two groups, highlighting the similarities and differences.

Table 2. Association patterns and language use of translator trainees and teacher trainees paradigmatic syntagmatic analytical/phono total

L1 L2 L1 L2 L1 L2 L1 L2 total

translator trainees 254 40 135 39 3 0 392 79 471

teacher trainees 33 230 16 207 1 7 50 444 494

In the dataset we observed an overreliance on paradigmatic responses but, interestingly, the ratio of paradigmatic responses does not seem to differ in the two groups, 65% and 66% in L1 and 50%

and 46% in L2 for the two groups, respectively (see Table 1). This finding is in line with Wolter’s research (2001), postulating that paradigmatic responses are present at the centre of the mental lexicon, followed by syntagmatic, collocational relationships. The number of phonological and analytic responses is fairly low in both groups, which points to a well-developed lexical network in the case of both groups. However, if we look at the total ratio of paradigmatic and syntagmatic responses, then the two groups show more differences. Altogether 62% of the total number of responses are paradigmatic and 37% syntagmatic for the translator trainee group, whereas 54% of all the given words are paradigmatic and 45% syntagmatic for the teacher trainee group.

In terms of the ratio of syntagmatic responses in L2, 50% of the L2 words provided by the translator trainee group refer to a sequential relationship, and this ratio is similar, 47% (although the number of words is much higher). We believe that if the translator trainee group had been asked to associate in L2 (and had not used L1 equivalents to such an extent), we would have been able to observe the paradigmatic-syntagmatic shift described by Zareva (2007). The rise in the number of syntagmatic responses in L2 might point to the development of syntax: more advanced second language users may think in terms of longer phrases and sentences.

We have also found that compared to previous studies of both lower and advanced level lear-ners, there were a lot of similar associations, and for our translator trainees a high number of these

were L1 synonyms and different meaning senses or different members of the same word family, which point to strong links between word form and word meaning.

In light of the results, Hypothesis 2 cannot be confirmed. The two groups showed different association patterns. Translator trainees’ and teacher trainees’ mental lexicons seem to store L1 and L2 words in a different manner, presumably due to their training background.

8.2 The translations

The participants in both groups produced comprehensible, acceptable Hungarian translations of the texts. Table 3 shows the evaluation of the two groups’ translations using the analysis described in 7.2. For each challenge, the number of participants per group producing a good solution is shown.

The items marked in bold present the greatest challenges the groups faced. For the translators’

group, the translation of terms related to statistics posed the greatest challenge. Terminological sensitivity also proved to be the teachers’ group’s weakness, besides indicating the sources, which was part of the assignment. Changing the word order of the English sentences that began with connectives (there were four such sentences in the abstract) posed a challenge in both groups. The groups’ total scores were very similar: 47 for the translators, and 46 for the teachers.

Table 3. Challenges in the translation in the two groups translators

(n = 10)

teachers

(n = 10) challenge total

1 omission (difference) 9 8 17

2 paper 8 8 11

3 piracy rate 1 3 4

4 protect more effectively 7 8 15

5 opposite signs 3 < 7 10

6 positive correlation 3 3 6

7 connectives 0 < 2 2

8 omission (of sentences) 9 > 2 11

9 sources 7 > 5 12

group total (max. 90) 47 46

The challenges listed above can be grouped into micro- and macro-level challenges. The translator trainees faced more of the former: they seemed to be in a hurry when translating the text, and gave in to the temptation of using first equivalents, in other words, literal translations in the Hungarian text, which shows that their monitoring skills need improvement so that they think deeper about word meaning, even in stressful situations. The teachers’ group showed a slightly different profile from translators: their weakest point proved to be macro-level challenges, like not omitting anything from the text, and fully complying with the assignment. On some of the challenges, like recognising terminology and finding the right equivalent, and transforming the English word order into natural-sounding Hungarian sentences, there is still need for improvement in both groups.

In the light of the results above, we can only partly confirm H3. For both groups, irrespective of their training background, terminological awareness and sentence-level operations proved to be a challenge.