Substitutable Defining Formats : A Study of Translation Equivalents

The present article reports on the results of a study which investigated the quality of translation equivalence for mental verbs as guided by two monolingual dictionaries namely, the Oxford Advanced Learner’s Dictionary of Current English and the COBUILD Dictionary, which employ two distinct defining strategies, Substitutability and Full-Sentence strategies respectively. Two parallel tests were administered to a group of Persian translation trainees. In one of the tests, the unknown verbs were defined by Substitutability and in the other, by FullSentence Strategy. The results showed that the participants performed significantly better on the test the items of which were defined by Substitutability strategy. The analysis of the test items revealed that both definition formats are conducive to kidrule strategy which guided them to select familiar words from the entry as a substitute for the search term. The application of this strategy was also found to depend on the position of information content, information load, in the definitions.


INTRODUCTION
Dictionaries play a major role wherever a language is used, taught and learned, especially in language education (Kirkness 2004).In the education system in general and in language teaching in particular, no book has been used as widely as the dictionary.It is the major source of information on language for all members of a literate society (Kirkness 2004).Lexicography, then, as the art and science of compiling, writing, editing and publishing of dictionaries, is viewed as an integral part of applied linguistics and its constituent subject areas.The last few decades have witnessed studies which devoted much attention to dictionary users (Kirkness 2004).This focus on dictionary users, as one perspective among others, and the need for further studies on what different groups of users do in real look-ups, connect lexicography with applied linguistics.Much as foreign and second language teaching and learning is one of the core activities of applied linguistics, there are other areas in which applied linguists make investigations.One such area is translation.Dictionaries of different types are employed by professional as well as novice translators.
General bilingual dictionaries, known as translation dictionaries, are evidently instruments translators and translation trainees resort to.They usually do not have trouble with conceptual meanings of words.However, due to different cultural and linguistic environments, misunderstandings and errors of interpretations occur with non-conceptual meanings (Yong & Peng 2007).Some experts (e.g.Adamska-Sałaciak 2006, Baker & Kaplan 1994), therefore, have asserted that bilingual dictionaries are not sufficient in finding contextually adequate equivalents for some words in two languages and some others contend that there is a basic lack of equivalence or anisomorphism between languages (Adamska-Sałaciak 2014).Realising that the problem at hand requires their reference to a dictionary, the translators may decide to consult a monolingual one.However, extracting the relevant information so that an adequate equivalent is derived is a complex step.It demands that the translator recognises and interprets the information correctly, in direct relation to the context of the translation task.The extent to which the whole operation of meaning extraction is successful is judged by the quality of the equivalents the translators provide.Part of this success depends on the obligatory changes they make in the part of speech of the words, i.e. shifts as Newmark (1988) puts it.Meeting the requirements of the context in which the words have been used is also an important factor in the quality of equivalents.Therefore, the success of dictionary look-ups inevitably depends on the format and quality of dictionary definitions.
Traditionally, monolingual dictionaries have different strategies for meaning explanation.To explain word meaning, four strategies namely, illustration, exemplification, discussion, and definition have generally been used by lexicographers (Ilson, as cited in Pearson 1998).
Apart from illustration, exemplification, and discussion, meaning specification basically involves definition.Meaning specification through definition is the major task for lexicographers.As the most common method of meaning explanation, definitions, according to Hanks (1987), may follow two styles, or explanatory strategies: the Substitutable Defining Strategy (SDS, hereafter) and Full-Sentence Defining Strategy (FSDS, hereafter)."A substitutable or analytic definition is assumed to be substitutable for its definiendum in any context in which the definiendum does or can appear" (Ilson, as cited in Pearson 1998 p. 218).The format of the definiens, the defining part of a definition, in an analytical definition is an incomplete sentence, a phrase constituent.This strategy is followed by some dictionaries, such as Oxford Advanced Learner's Dictionary of Current English (OALDCE).The substitutability principle entails the headword of the definition being at least in the syntactic form of the word to be defined: a verb for a verb, a noun for a noun and so on (Geeraerts 2003).The strategy was criticised by some lexicographers, like Landau (1989), who believe that although attempts are made to satisfy the principle, there are still definitions where clarity is impaired when lexicographers try to use a substitutable definition.There are also cases where, as Landau states, the principle is impossible to apply.Moreover, using analytic definition in dictionaries does not always specify the meaning of entry word, thus, the users may fail to provide an appropriate equivalence (Burkhanov 2004).
In the FSDS, employed by the COBUILD learners' dictionary, words "appear in their normal full spelling forms and the explanations are written in real sentences" (Sinclair 1987, p. 16).The definition utilises vocabularies and grammatical structures that occur naturally with the word being explained.This definition enables the lexicographers to give a lot of information about the way a word is used by speakers of the language (Sinclair 1995), allows additional information to be added in the sentence and gives a much fuller picture of the target lexical item.It does not make unreasonable demands on users nor does it require them to know any special conventions (Hanks 1987, Rundell 2008).
The COBUILD strategy for defining the words is flexible (Ndlovu, as cited in Khumalo 2002).To define verbs, which are the focus of the present paper, the following formats are employed: As a basic tool in the process of foreign language learning and translation activities, the usefulness of dictionaries and the significance of their definition formats are uncontroversial assumptions among lexicographers as well as language instructors.For example, it goes without saying that, no matter how the translation trainees approach the task at hand, they encounter problems related to the meaning of lexical items.The main difficulty encountered while translating is "lexical, not grammatical" (Newmark 1988, p. 32), at least for those who have already acquired the ability of source text analysis and comprehension as components of translation competence.It is, though, not unlikely that they do not either know the meaning of some lexical items, or they find them hard to translate.Speaking about word meaning, however, may indicate that words are viewed as possessing autonomous identity.This can be characterised as the decontextualised meaning of words, intended by general statements in dictionaries, by capturing the essential attributes of word meaning (Hutton 2014) through analytic substitutable strategy, and operationalised by single lexical item equivalence in some studies (e.g.Dziemianko & Lew 2013, Chan 2013).Nonetheless, the translation trainees are taught to adopt sentence as the unit of translation (Newmark 1988) and approach the texts sentence by sentence.Words are, therefore, encountered within a context where their senses are diffused across a sentence or the whole text.For translators, the problem is to extract the meaning of a lexical item within that particular context, i.e., to derive the contextual meaning of the word.The format of the contextual definition, known as Full-Sentence Definition in COBUILD dictionaries, is expected to be more effective in this regard.The claim is that, COBUILD dictionaries explain word sense in real texts by putting the users in actual situation of word use, conveying the illocutionary force of expressions, and providing a fuller picture of the lexical items (Rundell 2008).Paragraph explanation puts word meaning and its contexts into an integrated unit and connects the meaning and context to each other.This connection enables the COBUILD dictionary to provide a different kind of information for the user (Yong & Peng 2007).
In virtue of the significance of definition formats in meaning extraction for translators, the main purpose of the present study was to examine which definition format, substitutable or full sentence, might more effectively assist the Iranian translation trainees in extracting contextual meaning of unknown lexical items.

REVIEW OF LITERATURE
Studies on how English learners perceive and use dictionaries have focused mostly on students or learners of second languages (Nesi 2000).The scope of the studies ranges from the way dictionaries and their linguistic context might be effective in vocabulary acquisition (e.g.AL -Mahbashi, Mohd Noor & Amir 2015;Sadeghy & Nobakht 2014) to how dictionary labels are organised to assist the users (e.g.Vrbinc & Vrbinc 2015).
A significant number of attempts have been made to examine whether bilingual or monolingual dictionaries are to be recommended to second language learners.Laufer and Hadar (1997), for instance, examined the effectiveness of monolingual, bilingual, and bilingualised dictionaries.Whereas bilingual, or translation dictionaries, usually provide just L1 equivalents, bilingialised dictionaries usually repeat the definitions and examples, translate L2 definitions literally or give translation equivalents in L1 (Kirkness 2004).The authors set a number of tasks for different groups of advanced learners and concluded that the bilingualised dictionaries tend to be more effective for receptive and productive tasks.
Aiming to study and improve dictionary consultation skills, Wingate ( 2004) conducted his study on 17 intermediate Chinese learners of German.Examining the subjects' look-up process, he identified two problems.First, the subjects failed to pay attention to the whole information included in the entries.Second, they did not possess proper strategies for dictionary consultation.The author found that the subjects used a familiar part of the entries as an equivalent for the unknown word.This behaviour, called kidrule strategy, was previously introduced as a negative strategy in dictionary consultation by Miller and Gildea (1987).
Still, another topic debated with regard to dictionary use is the presence of different grammatical indications in the dictionary.Bogaards and Kloot (2001), for instance, compared the usefulness of the grammatical information that the learners' dictionaries of English, namely, LDOCE3, COBUILD2, and CIDE provided for verb complementation.They concluded that the COBUILD definitions tended to yield better results with respect to correctness, measured by completing the missing verbs in translation tasks at the level of sentence.Yet as to the user-friendliness of the syntactic information no significant main effect of the dictionaries was found.
In another study aimed at assessing the user-friendliness of noun and verb coding system in learners' dictionaries of English, Dziemianko (2006) found that, as far as information on verb syntax is concerned, full-sentence definitions save the learners' conscious effort to look for the information.Dziemianko and Lew (2013) also compared the usefulness of the single-clause when-definitions with that of analytic substitutable definitions in recognising the part of speech of nouns for Polish native speakers.The participants were to provide Polish equivalents and compose English sentences with the target words.On both tasks, the participants performed better when the headwords were defined analytically.The authors, though, did not account for their findings.The result of this and similar studies could have been more informative if the entries from the dictionaries under study had been scrutinised with reference to the way information is distributed at the level of definitions, that is, spread through the definition or otherwise loaded in a particular position of the definition.Literature has shown that a considerable amount of information, transmitted by content words, is not evenly distributed at the level of English sentences.Sentences hold lower amount of information in the initial and higher amount in the final positions.
Preference for dictionary use in translation was studied by Atkins and Verantola (1997).The look-ups were mainly recorded to be performed in bilingual dictionaries.However, directionality played a part so that monolingual dictionaries were used more frequently in translation into the participants' mother tongue while bilingual dictionaries were the references in both directions.
The results of the studies on the usefulness of dictionaries of different types and their grammatical information or user-friendliness of coding systems have impacted our knowledge on second language learners at college or high school levels.Despite the dynamic development of research in dictionary use, to the knowledge of the researchers, no research has been devoted to the question of how the quality of contextual equivalents extracted from monolingual dictionaries may depend on the format of the defining strategy.
While most learners' dictionaries are increasingly and selectively employing FSD (Rundell 2008), the strategy is still unknown to translation trainers and translation trainees.Not being aware of different approaches to defining lexical entries, translation trainers in Iran introduce the Oxford Advanced Learner's Dictionary of Current English to be consulted as a monolingual dictionary for meaning extraction when bilingual dictionaries fail to assist the trainees.Therefore, even if we accept the argument that FSD has a higher pedagogic value when vocabulary leaning is concerned, we might still need to see and test their values for translation trainees by examining their performance on the quality of contextual equivalents.If the strategy is a better guide in meaning extraction, it should result in better equivalents by comparison with SDS.The essential point that has not been discussed so far is the usefulness of different definition formats for translation trainees and translators, and for different groups of words.
In view of the above context, the aim of the present study was to explore whether exposure to different definition formats, namely substitutable and full sentence definitions, would influence the quality of contextual equivalence in translation tasks.In particular, in this study, the success of meaning extraction was judged by the degree to which the participants appropriately provided the contextual meaning of the target lexical items, mental verbs, represented by a single lexical item as well as a stretch of words.

THE PRESENT STUDY
The objective of the present study was to examine if there is any significant difference between the quality of contextual equivalents for unknown mental verbs, guided by SDS and FSDS.
It was hypothesised that the COBUILD dictionary can facilitate the translators' job in extracting the context related meaning of unfamiliar mental verbs.METHODOLOGY PARTICIPANTS Since randomisation of individuals was not feasible, two intact classes of 105 Persian native speakers who studied English translation at Sheikhbahaee University in Isfahan, Iran made the convenient sample of the study.They took an Oxford Placement Test (OPT).Out of 105 candidates, 72 were intermediate to advanced in English (according to the ranking of the OPT).About 80% (n=57) of the participants were female and 20% (n=15) male, the average age being 22 years.

INSTRUMENTS
The instruments of the study comprised of two parallel twenty-item contextualised translation tests.In one of the tests, Oxford Test (OT, hereafter), the target verbs were replaced by nonce or pseudo words which were defined by employing SDS.A nonce word is invented for a particular occasion and is only used once.In the other test, COBUILD Test (CT, hereafter), the nonce verbs, representing the same mental verbs, were defined by FSDS.Each item held one target mental verb embedded in a short context.In order to be sure that the words were unknown to the participants and they found the equivalents only with the help of the definitions rather than their background knowledge, a pair of nonce words were initially coined for each mental verb; one for the items of the CT and the other for the items of the OT.The pseudo verbs were morphologically inflected in their context in accordance with English regular inflectional rules.To delimit the findings to the defining strategy, the tasks were merely followed by the definitions taken from OALDCE 7 th edition, and COBUILD 5 th edition, since some studies (e.g.Dziemianko 2006) have shown that there were cases where dictionary definitions proved to be less helpful than the following examples.The participants, therefore, had to rely only on the definitions to extract the meaning of the target words.Moreover, there was no grammatical indication accompanying the definitions since it was found in the related literature (e.g.Tono 1988) that they are largely ignored by dictionary users.
Below are examples of items in the tests, one from the CT and one from the OT.The target item was 'Perceive'.
OT item: The more honest you become, the more accurate your model of reality will be.As you voltar reality with increasing accuracy, your decisions will improve, and in turn so will your actions and thereby your results.
CT item: A sound spiritual philosophy must be firmly rooted in truth.This requires that we strive to atrim reality as accurately as possible.How exactly can we achieve accuracy when trying to atrim the true nature of reality?-Atrim: If you atrim something, you see, notice, or realise it, especially when it is not obvious.
A PILOT STUDY Prior to the administration of the translation tasks, two 25-item tests were developed and administered to a group of 30 translation students.It was intended to check the clarity of test instruction, to remove the items found to be problematic, to calculate the reliability of the instruments, and to make the tests parallel.Five items were removed from both tests since they were either answered or missed by more than 95% of the participants.The researchers also found that the instruction of the test should be provided in Persian so that the participants understand what they are expected to do during the translation tasks.
To make the tests as parallel as possible, the number of the items and their format were the same.The researchers also made sure that each pair of items in both tests conveys relatively the same type of content.After piloting the tests, the means and standard deviations for the tests were calculated and found to be very close (OT, M=0.63, SD= 0.12, CT, M= 0.52.SD=0.14).
After piloting the tests and removing five items reliability of the instruments was calculated.The estimated Cronbach's alphas were 0.71, for the OT, and 0.67 for the CT.
Since the analysis might have involved subjective judgments, a colleague, who was also lecturing courses in translation studies, scored half of the papers (15 sets).The correlation coefficients for the OT and CT were r=0.78, p<0.05 and r=0.72, p<0.05 respectively.

THE TARGET LEXICAL ITEMS
The reasons for selecting verbs were threefold.First, the researchers intended to delimit the domain of the lexical items.Second, as Jennings puts it "[a] verb is a power of all speech, […] [i]t brings to birth."(Jennings as cited in Aarts-Meyer 1995, p. 1).And third, due to the centrality of verbs in the orientation and meaning of sentences, it is less likely that they are subject to arbitrary or obligatory omission in translation.However, mere absence of the intended verb equivalence in the verb position may not be an indication of mistranslation, since shifting in part of speech was fairly probable.Mental verbs were specifically the focus of the present paper since they are of notably high frequency both in spoken and written texts (Biber, Johnsson, Leech, Conrad, & Finegan 1999).The tests (OT and CT) were administered during regular class times, without time limitation, in two classes with an interval of two weeks.The participants could consult any dictionary to translate the bold sentence containing the target verb represented as a nonce word in each translation task.They were required to translate the sentences in Persian considering the context.However, the scoring was based on the accuracy of propositional meaning and co-occurrence restrictions imposed by the context on the target lexical item (Baker 2011).To score the tests, one point was assigned to appropriate equivalents represented in whatever part of speeches or even extended in the sentence.Otherwise, the score of 0 was assigned.Missing responses were also given a score of zero.

DATA ANALYSIS
The Data obtained from the OT and CT were used for running paired-samples t-tests to test the research hypothesis introduced above.Initially the participants' scores on the OT were compared by running a paired-samples t-test with those on the CT, to see which definition strategy better guided the participants in extracting the meaning of the unknown lexical items.
Subsequently, a set of paired-samples t-tests were conducted for individual test items to investigate, more deeply, on which items in each test the participants performed significantly better.

RESULTS
In this section, the participants' performance in the translation tasks will be presented together with some examples of their performance on the items.
The results of descriptive statistics showed that the participants were less successful in extracting the contextual meaning and the relevant information from the items in which the nonce words were defined by FSD than the items in which the nonce words were defined by SD.On the items of the OT and CT, the means were 0.48 (SD= 0.17), and 0.43 (SD= 0.19) respectively.
Significance of the difference was tested with a paired-samples t-test carried out in SPSS.Table 1 shows the results of the statistical analysis.As the table illustrates, even though the mean difference is only 0.05, it proved to be statistically significant: t (71) = -2.09,p˂0.05.Therefore, the participants had a significantly better performance on the OT than the CT.
To obtain a clear picture of meaning extraction, results obtained for each item were compared across the tests, i.e. the participants' performance on each target mental verb on the CT was compared with their performance on the item related to the same verb in the OT, running a set of 20 paired-samples t-tests.The results are given in Table 2.As can be seen, on six items of the CT, the participants performed significantly better than the corresponding items on the OT.The items were related to the verbs: recognise, experience, prioritise, deserve, integrate, and devise.
There were 6 items in the OT, related to the verbs: satisfy, identify with, assign, gratify, perceive and manipulate, on which the performance of the participants were significantly better than the corresponding items on the CT.For the remaining 8 items (calculate, justify, deduce, turn in to, verify, characterise, discriminate and contribute), the difference between the participants' performance was not significant on the OT and CT.

DISCUSSION
Contrary to the prediction in the research hypothesis, on the whole, meaning extraction proved to be more successful when the verbs were defined by substitutability principle.The participants failed to recognise and correctly interpret the meaning in relation to the context of each translation task.The worse performance of the participants on the items followed by FS definition suggests that providing the words' typical environment and behaviour, which has been claimed to be critical to any account of their semantics, makes the definitions more demanding to process, and it may exert opposite effects on contextual meaning extraction.Being exposed to substitutable definitions, the Persian learners may, on the face of it, become habituated to benefiting from the familiar format for definition.
The more valid reason for providing acceptable equivalents drawing on analytic definition might lie in the fact that, making the definitions longer in the COBUILD dictionary to convey the illocutionary force of the lexical items (Rundell 2008) failed to help the participants in their job.The following examples would illustrate the point.
The COBUILD dictionary has defined the verb "Perceive" as: If you perceive something, you see, notice, or realise it, especially when it is not obvious.Inspecting the contextually wrong equivalents, such as Moshakhas Kaedan (to identify) and Bavar Kardan (to believe), together with the context in which they had to provide the equivalence (see Appendix) revealed that the final part of the definition made them confused in their task of meaning realisation.They used the contexts in the translation tasks to conclude on an equivalence which does not violate the collocational restrictions associated with that equivalence.The short and easy-to-process Oxford definition (Become aware or conscious of (something), though, could guide them more effectively in this regard.In other words, while the claim is that the semantic and contextual integrity of each entry word in a COBUILD dictionary provide the users with a quite different kind of information, which is characterised by abstractness, completeness, objectivity and intelligibility (Sinclair 1987), the participants are more confused while trying to understand the propositional meaning of the entries.
Failure in deriving the equivalent might also be attributed to the fact that the translation trainees are accustomed to finding a synonym which is unconsciously assumed to be easier and more frequent than the entry.As an example, the definition of the word 'manipulate' is presented together with the erroneous equivalents.
Manipulate (COBUILD definition): If you say that someone manipulates an event or situation, you disapprove of them because they use or control it for their own benefit.
Manipulate (Oxford definition): To control or influence somebody, often in a dishonest way so that they do not realise it.
The wrong equivalents provided by the participants for this item of the CT, such as Nadide gereftan (To ignore), Tahte taasir gharar gereftan (to be affected by), Dorugh Goftan (To lie), and the correct equivalents produced by the same participants for the same item of the OT, such as Be nafe khod tagheer dadan (skillful change in something in favor of somebody), Farib dadan (to deceive) reveal that, no matter what type of strategy is used for definition, they look for a synonym that can be substituted for the word in the sentence (see Appendix).
This bears out the Gromann and Schnitzer's (2015) findings in addressing the resources, selection processes and consultation strategies of learners of five languages.Studying business and economics, the users of monolingual dictionaries in their study reported that, looking for synonyms for text translation and production tasks were the main reason to use the dictionary.It also might be in line with Fabiszewski, Jaworski and Grochocka's (2010) conjecture that based on the obtained evidence, some familiar parts of non-substitutable definitions were considered as synonyms of some headwords.
Examination of the items on which the participants' performance did not differ significantly on the OT and CT suggested another possibility-that both definitions are conducive to the kidrule principle which consists of "the extraction of a readily known substring from an item's definition and treating that substring as equivalent in meaning to the item defined" (Miller & Gildea 1987, p. 88).Meaning extraction from both analytic and full sentence definitions seems to almost equally feature the use of kidrule strategy.
The apparent inconsistent behaviour of the participants in contextual meaning extraction might also be attributed to the way information content is distributed within the definitions.Not only is sentence comprehension influenced by syntactic and semantic factors, but it also is affected by intra-sentence distribution of information content.The close examination of the definition pairs, one taken from the Oxford and the other from the COBUILD dictionaries, as well as that of the wrong equivalents provided by the participants, demonstrated that information is extracted from different parts of the definitions.From the full sentence definitions, the participants selected the substring from the mid or final positions where, in comparison to the initial position, more information content is carried (Yu, Cong, Liang, Liu 2016).Yet, inspecting the wrong equivalences for the OT items, on which the participants performed significantly less successfully, revealed that from the analytic definitions, where extra information is provided to differentiate the meaning of the entry from that of the superordinate, the initial position of definition which contains the greatest load of substitutable information is assumed to be mainly selected.The phrasal format of the definition renders it short enough to be processed as a complete synonymous unit.This can further be related to the initial occurrence of high-frequency words in the FSDs in the COBUILD dictionaries, and the occurrence of low-frequency words bearing more load of information in other positions in this format of definition.High frequency words are the words which occur frequently in written as well as spoken materials, for example, 'and', 'the', 'there', 'wh-forms', 'as' and 'it'.This group of words are mostly function words, and have little meaning on their own.However, they contribute a great deal to the meaning of a sentence.Low-frequency words, on the other hand, with higher word-length than function words, constitute majority of lexicon and transmit more information.The analytic and full sentence definitions for the verb discriminate will illustrate the point.
Discriminate (COBUILD definition): If you can discriminate between two things, you can recognise that they are different.
As shown in the examples, the core of meaning in the Oxford definition lies in the whole short phrase which can be taken as a synonym simply adjustable to the context of translation task (see Appendix).The whole phrase consists of two content low-frequency words that bear the load of information.However, the initial position of the FSD in the COBUILD dictionary is occupied by high-frequency words, if, you, and can.Therefore, it is the final part of the COBUILD definition that holds the information load.

CONCLUSION
Having observed the results of the study, it can be concluded that, overall, the translation trainees find substitutability principle easier to draw upon, through which the meaning of simple paraphrases and synonyms can be derived.The contextual equivalents provided by the translation trainees supplied some evidence that, regardless of definition format, the participants' route to meaning extraction is looking up for substitutable synonyms that can be adjusted to the translation context.Seeking synonyms, which is barely distinct from kidrule strategy in process, was presumed to account for some erroneous equivalents as well as correct ones.
The study also raised the possibility that applying kidrule strategy is subject to the position of information content where the content lexical items are loaded.While the Persian translation trainees of English texts expect to find the information load in the final position in full sentences, they may look for the load of information in the short substitutable phrasal definitions which can conveniently work as synonyms.The information added at the final position of the definitions, which play the distinctive role to differentiate the headword from the subordinate introduced at the initial position of the phrase, can account for the errors in finding equivalents taken from the analytic substitutable definitions.
Research on dictionary use is a relatively new area, and it still faces many unanswered questions.Some points which fall beyond the scope of this study might be an area of further research.First, no special attention was paid to the participants' L1 competence which is a basic component to consider in all translation assessment tasks.Taking this as a variable can be interesting in further analysis of the errors in equivalence extraction and in categorising the errors based on their L1 knowledge.
Second, future research might yield more reliable results if the quality of equivalents in naturalistic situations for looking-ups is to be investigated.Third, to generalise from the findings about the way definitions of different formats might be used by translation trainees, future studies need to be conducted with participants of various proficiency levels in English.However, as the focus of the current study was on the usefulness of definition strategy in equivalent extraction, the findings could be illustrative for translation instructors, in that they raise the learners' awareness of the way dictionaries use different strategies in setting forth the meaning of lexical items.

PROCEDUREA
standard Oxford Placement Test (OPT) was administered a week before the administration of the research instruments.The placement test was carried out on the total population of 105 B.A. students.Out of this number, 72 were found to be intermediate or above intermediate translation students, making the participants of the study homogeneous.

TABLE 1 .
Paired samples t-test on the participants' mean scores in the CT & OT

TABLE 2 .
Paired samples t-tests by item