UDC 378.147:81’25
Қасымжан Сабина
Тойлыбай Ақмарал
Тұрарбек Альбина
Студент, Шетел тілдер факультеті,
Е.А. Бөкетов атындағы Қарағанды Ұлттық Зерттеу Университеті,
Қазақстан, Қарағанды
Жетекшісі: Сагадиева Қымбат Кошкинбаевна
Ғылыми жетекші, педагогика ғылымдарының магистрі,
Е. А. Бөкетов атындағы Қарағанды Ұлттық Зерттеу университеті
Қазақстан, Қарағанды
Abstract. This article examines the phenomenon of translation dependence in foreign language acquisition, a condition in which learners habitually resort to their native language as a cognitive intermediary rather than developing direct conceptual links in the target language. Drawing on contemporary psycholinguistic research and pedagogical practice, the study identifies the key cognitive, methodological, and affective factors that give rise to excessive translation reliance. The article systematically reviews evidence-based strategies for reducing this dependence, including immersive communicative approaches, lexical chunking techniques, concept-based instruction, and digital tools that promote target-language-only thinking. A pedagogical experiment conducted with university-level learners of English demonstrates measurable improvements in fluency and conceptual independence when structured de-translation protocols are applied. The findings suggest that overcoming translation dependence requires a deliberate, staged pedagogical intervention rather than simple restriction of the mother tongue in classroom settings.
Keywords: translation dependence, foreign language acquisition, interlingual interference, L1 transfer, communicative approach, mental lexicon, direct method, lexical chunking, immersion pedagogy, bilingualism.
1.Introduction
The capacity to think directly in a foreign language, without the mediation of one’s mother tongue, stands as one of the most significant yet undertheorized goals in applied linguistics and language pedagogy. In formal educational settings across Kazakhstan and Central Asia, learners of English regularly exhibit what researchers term translation dependence — the systematic reliance on Kazakh or Russian as a cognitive scaffold when producing or comprehending English. While translation as a pedagogical tool carries a long and respectable tradition, its uncontrolled use by learners frequently solidifies into a cognitive habit that inhibits genuine linguistic competence [1, p. 12].
The dependency manifests in multiple observable ways: prolonged pausing before speaking while mentally constructing L1 sentences, systematic structural transference that results in non-target-like syntax, over-reliance on word-for-word correspondence in lexical choices, and the inability to produce spontaneous discourse without prior L1 planning. These behaviors persist well beyond the beginner stage, often continuing into advanced levels of proficiency, suggesting that classroom methods — rather than learner incapacity — play a central enabling role [2, p. 34].
The present study addresses three interconnected questions: first, what cognitive and pedagogical mechanisms sustain translation dependence over time; second, which instructional strategies most effectively reduce this dependence without alienating learners from their linguistic identity; and third, how can teachers in Kazakhstani higher education contexts implement de-translation interventions within existing curricular constraints. These questions are addressed through a review of current psycholinguistic literature, a taxonomy of evidence-based pedagogical strategies, and the results of a controlled pedagogical experiment conducted at the university level [3].
The urgency of this inquiry is heightened by Kazakhstan’s ongoing trilingual education policy, which simultaneously valorizes the Kazakh language, Russian as a lingua franca, and English as a language of science and international communication. In this complex multilingual context, the boundaries between beneficial code-switching and counterproductive this phenomena become particularly difficult to navigate, making principled pedagogical guidance all the more necessary [4, p. 8].
The pedagogical experiment at the core of this study was conducted during the 2023–2024 academic year in the author’s own English language classes at Eurasian National University. The experiment encompasses the following components: a pre-experiment CEFR placement test to determine learners’ initial proficiency levels; an 18-week structured de-translation protocol applied to the experimental group; and a multi-method post-test battery measuring oral fluency, written production, and self-reported L2 thinking confidence. This design allows for a controlled comparison between traditional grammar-translation instruction and the proposed de-translation approach within an authentic Kazakhstani university classroom context.
- Theoretical foundations: translation in the cognitive architecture of language learning
To understand this reliance, it is necessary first to examine how the bilingual or multilingual mental lexicon is organized. Contemporary psycholinguistic models, particularly the Revised Hierarchical Model proposed by Kroll and Stewart, distinguish between conceptual and lexical levels of representation. In early stages of acquisition, L2 words are accessed via their L1 equivalents, creating what is effectively a word-association rather than concept-mediation pathway. Mature proficiency, by contrast, involves direct concept-to-L2-word links that bypass L1 entirely [5, p. 149].
This developmental trajectory means that translation dependence is, to a significant degree, a natural feature of early language learning rather than an aberration. The problem arises when pedagogical conditions actively reinforce this early pattern beyond the stage where it serves learners’ needs. Grammar-translation methodology, still prevalent in post-Soviet educational contexts, presents all new vocabulary with L1 glosses and frames all grammatical knowledge as a system of equivalences between two languages. This approach, which dominated Soviet-era foreign language instruction for decades, has been extensively criticized in contemporary applied linguistics for producing learners who can parse sentences and pass written examinations but are fundamentally ill-equipped for spontaneous communicative interaction. By structuring the entire learning process around L1-L2 equivalence, grammar-translation actively prevents the development of the direct conceptual links that characterize genuine proficiency. It trains learners to be decoders rather than speakers, and its enduring legacy in Kazakhstani classrooms represents a significant structural impediment to communicative competence. Learners trained in this tradition develop what Cook termed compound bilingualism — a cognitive arrangement in which the two linguistic systems are densely interconnected rather than functionally separable [6, p. 192].
Vygotskian perspectives on inner speech add a further dimension to this analysis. Vygotsky argued that thought and language are intimately related, and that mature thinking is partly structured by the language one uses for inner verbalization. Learners who routinely conduct their inner speech in L1 while performing L2 tasks are, in effect, thinking in one language and translating the output — a cognitively expensive and communicatively limiting practice. Research using think-aloud protocols has confirmed that high-dependence learners show significantly more L1 inner speech episodes than low-dependence counterparts, even when their explicit proficiency scores are comparable [7, p. 67].
Affective factors further complicate the picture. Translation provides a security mechanism — a way of managing the anxiety inherent in L2 communication. Krashen’s affective filter hypothesis suggests that high anxiety raises the barrier to acquisition; paradoxically, the coping strategy that learners use to reduce anxiety — translation — simultaneously prevents the deep processing necessary for acquisition. This creates a self-reinforcing cycle in which translation dependence, anxiety, and incomplete acquisition mutually sustain one another [8, p. 31].
- Manifestations and consequences of translation dependence
The most immediate observable consequence of the dependency is disrupted fluency. Skilled speakers of a language process and produce speech in overlapping cycles, with planning for upcoming utterances occurring simultaneously with the articulation of current ones. Learners who translate serially — mentally composing an L1 utterance, then converting it to L2 — cannot sustain this overlap, resulting in the characteristically hesitant, pause-laden speech that marks dependent learners. Studies using temporal analysis of learner speech consistently show that translation-dependent speakers have longer and more frequent unfilled pauses, a higher rate of false starts, and lower rates of subordinate clause use than their more autonomous counterparts [9, p. 78].
Lexical consequences are equally significant. Learners who conceptualize through L1 and then seek L2 equivalents tend to produce what Nation calls lexical poverty — a narrow vocabulary range driven by the availability of ready translations. Words that lack close L1 equivalents — particularly items whose semantic range differs between languages — are avoided or misused. Phrasal verbs in English, for instance, present systematic difficulties for Kazakh-speaking learners not because of phonological complexity but because Kazakh does not possess an analogous grammatical category; learners tend to paraphrase them using single-word synonyms, a strategy that, while communicatively adequate, produces a stylistically marked register [10, p. 203].
At the discourse level, translation dependence introduces structural interference from L1 rhetorical organization. Academic writing in Kazakh and Russian, for example, tends toward a topic-comment structure that defers the main argument, whereas English academic prose typically employs a topic-sentence-first structure with explicit logical connectors. Learners who plan their writing in L1 and translate the output frequently produce texts that appear globally coherent in L1 terms but are locally incoherent by L2 standards — a problem that surface-level correction rarely resolves [11, p. 115].
- Evidence-Based strategies for overcoming translation dependence
The literature on reducing this reliance converges on several clusters of pedagogical strategy, each targeting a different component of the dependency mechanism. No single intervention is universally sufficient; effective de-translation pedagogy requires a coordinated approach that addresses cognitive habits, classroom norms, and learner affect simultaneously [12, p. 44].
4.1. Lexical chunking and phrase-based learning.
Rather than presenting vocabulary as individual words with L1 equivalents, phrase-based instruction teaches multi-word units as unanalyzed wholes. The psycholinguistic rationale is that chunked items are stored and retrieved as single units, bypassing word-by-word mapping. Nation and Waring’s research on vocabulary acquisition shows that learners who receive chunk-based instruction demonstrate significantly lower rates of translation episodes in subsequent production tasks. Crucially, chunks also provide learners with pragmatically appropriate, nativelike language that learners who assemble utterances word-by-word rarely produce [10, p. 89].
4.2. Image-Based Vocabulary Instruction.
Presenting new vocabulary through images, contextual scenes, or experiential association rather than L1 glosses directly addresses the Kroll-Stewart developmental problem. When a learner first encounters a new L2 word paired with an image rather than a translation, the initial representation established in memory connects the L2 form to a non-linguistic concept rather than to an L1 label. Longitudinal studies by Mondria and Mondria-De Vries demonstrate that image-paired vocabulary shows stronger retention and significantly lower rates of L1 intrusion in subsequent production [13, p. 22].
4.3. Communicative task design.
Tasks that create genuine communicative pressure — where the learner’s primary attention is on conveying meaningful information rather than formal accuracy — reduce the cognitive space available for translation. Information-gap activities, jigsaw tasks, and problem-solving discussions in which each participant holds unique information that others need generate authentic communicative urgency. Under such conditions, learners default to L2 direct production because the cognitive overhead of translating interrupts the communicative momentum of the interaction [14, p. 56].
4.4. Concept mapping and semantic webbing in L2.
Having learners construct semantic maps, mind maps, or vocabulary webs entirely in the target language fosters the development of L2-internal conceptual organization. When a learner maps the concept ‘freedom’ through a web of associated English words — liberty, choice, autonomy, democracy, restriction — without passing through Kazakh or Russian intermediaries, the resulting associative network more closely resembles that of a native speaker. This technique is particularly valuable for abstract vocabulary domains where L1 equivalents may exist but carry different connotations or collocational restrictions [15, p. 178].
4.5. Think-aloud protocols and metacognitive awareness.
Training learners to verbalize their cognitive processes in L2 during reading, writing, or problem-solving tasks serves a dual function: it makes inner speech L2-mediated rather than L1-mediated, and it develops metacognitive awareness of when and why learners switch to L1. Learners who participate in structured think-aloud training report greater awareness of their translation habits and greater ability to deploy deliberate strategies for maintaining L2 thinking [7, p. 99].
- Pedagogical experiment: design and results
To assess the effectiveness of a coordinated de-translation intervention, a pedagogical experiment was conducted over one academic semester (18 weeks) at Eurasian National University. Participants were second-year students of the English philology department, divided into a control group (n=27) taught with standard communicative language teaching materials, and an experimental group (n=29) receiving an additional structured de-translation protocol integrated into the standard curriculum [3].
Prior to the intervention, all participants completed a standardized CEFR-aligned placement test in order to establish their initial proficiency levels. The test consisted of four sections: a grammar and vocabulary multiple-choice component (40 items); a reading comprehension section with two texts and 20 questions; a writing task requiring a 150-word opinion paragraph; and a short oral interview of approximately five minutes, scored by the author using CEFR oral descriptors. Based on the aggregate results of this placement test, the majority of participants in both groups were identified as operating at B1 level, with a smaller cohort reaching B2 in receptive skills. This proficiency distribution is consistent with typical second-year English philology students in Kazakhstani higher education and provided a sound baseline for measuring the de-translation intervention’s effects.
The de-translation protocol included the following components: weekly lexical chunking sessions using corpus-derived phrasal units; image-based vocabulary introduction for all new lexical items; three timed free-writing tasks per week with an explicit ‘English-only thinking’ instruction; one weekly concept-mapping activity in L2; and bi-weekly reflective journals in which learners noted moments of translation dependence and alternative strategies they employed. Both groups attended the same number of classroom hours; the experimental group’s additional protocol replaced rather than supplemented regular homework tasks [16, p. 34].
Assessment was conducted at the beginning and end of the semester using a multi-method battery. I personally administered the placement test at the start of the semester and found the experience instructive in itself: sitting with each student for the oral interview component, I could observe firsthand how deeply ingrained the translation habit was — many students would visibly pause, form sentences in Kazakh under their breath, and then attempt an English rendering. This observation reinforced my conviction that the intervention was necessary. Oral fluency was then assessed through a 10-minute unscripted interview, analyzed for frequency of L1-search pauses, lexical complexity, and use of multi-word units. Written production was evaluated using an argumentative essay task, scored for coherence, lexical range, and structural sophistication. Learners also completed a self-report questionnaire measuring perceived L2 thinking confidence [17, p. 45].
The results demonstrate consistent and statistically significant advantages for the experimental group across all measured dimensions (p < 0.01). Most striking was the reduction in translation episodes during oral production: while control group learners averaged 8.4 observable L1-search pauses per 10 minutes of speech at semester’s end — a figure virtually unchanged from the pre-test — experimental group learners averaged only 3.1. This reduction was accompanied by substantial gains in lexical complexity, as measured by the Guiraud index, and in oral fluency scores calibrated against the IELTS band descriptors [17, p. 78].
Qualitative data from learner reflective journals corroborated the quantitative findings. Experimental group participants reported a progressive shift in their experience of L2 production: whereas early-semester entries frequently mentioned the effort of ‘translating thoughts into English,’ later entries more often described thinking in English directly, with L1 appearing only for culturally specific concepts lacking clear L2 equivalents. This phenomenological shift aligns with the psycholinguistic prediction that sustained communicative practice in conditions that discourage L1 mediation gradually restructures the mental lexicon toward direct concept-to-L2-word mapping [5, p. 162].
Several limitations of the experiment must be acknowledged. The study was conducted within a single institutional context with a specific linguistic background (Kazakh/Russian L1), limiting generalizability. The 18-week timeframe, while sufficient to produce measurable changes, is too short to determine whether the gains reflect durable restructuring of cognitive habits or temporary behavioral adaptation. Future research should incorporate delayed post-tests and neuroimaging methodologies capable of detecting changes in the neural substrates of lexical access [18, p. 23].
- Implementation recommendations for Kazakhstani pedagogical contexts
Translating the above findings into actionable guidance for teachers in Kazakhstani higher education requires attention to the specific institutional and sociocultural constraints of this context. Several recommendations emerge from both the experimental data and the broader literature review [4, p. 16].
First, teachers should gradually reduce the role of bilingual vocabulary lists in their instruction, replacing them with image-based and contextual vocabulary presentation. This transition need not be abrupt; a staged reduction over one academic year, beginning with new lexical items and preserving L1 glosses only for abstract grammatical concepts, allows learners to adapt without excessive affective disruption. Importantly, this recommendation does not imply a prohibition on all L1 use — code-switching for conceptual clarification remains appropriate and linguistically beneficial [6, p. 205].
Second, task design should prioritize genuine communicative need over controlled practice. Teachers who currently rely predominantly on pattern drills and translation exercises should incorporate at least one information-gap or collaborative problem-solving task per lesson. These need not require elaborate materials; a simple task in which two learners each hold different halves of a diagram and must reconstruct the whole through oral description creates the communicative pressure that reduces translation reliance [14, p. 78].
Third, metacognitive development should be treated as a curriculum goal rather than a side effect of instruction. Learners benefit from explicit discussion of translation dependence as a phenomenon — understanding why they translate and what alternative strategies exist — at least as much as they benefit from practices that implicitly discourage translation. Short, structured reflection tasks integrated into regular classroom routines can serve this purpose without requiring additional contact time [7, p. 101].
Fourth, digital tools can provide valuable support. Monolingual English dictionaries with rich example sentences — particularly corpus-informed resources such as the COBUILD or Longman dictionaries — encourage target-language-internal definition seeking. Speech-recognition applications that provide immediate feedback on pronunciation and fluency create low-stakes oral production opportunities outside the classroom. Vocabulary learning platforms such as Anki, when configured to use image-cue cards rather than translation-cue cards, systematically build the type of direct conceptual representation that the mental lexicon ultimately requires [13, p. 45].
- Conclusion
Translation dependence represents one of the most persistent and pedagogically consequential phenomena in foreign language education, particularly in multilingual contexts such as Kazakhstan where learners operate across three distinct linguistic systems. This article has demonstrated that the condition arises from an interaction of cognitive, methodological, and affective factors, and that it is most effectively addressed through a coordinated set of pedagogical interventions targeting each of these dimensions.
The experimental evidence presented here supports the conclusion that structured de-translation protocols — combining lexical chunking, image-based vocabulary learning, communicative task design, concept mapping in L2, and metacognitive reflection — produce measurable and meaningful improvements in learners’ linguistic autonomy within a single academic semester. The experimental group’s reduction of observable L1-search pauses by 63% relative to the control group, alongside gains in lexical complexity and self-reported L2 thinking confidence, provides a compelling empirical basis for the pedagogical recommendations advanced in this article.
It is important to conclude with a caveat against overgeneralizing the implications. Overcoming the dependency does not mean eliminating the learner’s L1 from the educational environment. The mother tongue remains a cognitive resource of great value — for clarifying nuance, for managing classroom rapport, and for the kind of deep metalinguistic reflection that promotes lasting acquisition. The goal is not L1-free language learning but L1-optional language learning: a state in which the learner possesses both the ability to think directly in the target language and the strategic awareness to deploy their full multilingual repertoire when it serves genuine communicative and cognitive purposes [6, p. 218].
Future research should extend this inquiry to younger learner populations, examine the long-term durability of de-translation gains beyond the experimental semester, and investigate whether the strategies identified here yield differential results across learners with varying L1 backgrounds within the Kazakhstani educational system. The development of standardized assessment instruments specifically designed to measure translation dependence as a psycholinguistic construct — distinct from general proficiency measures — would substantially advance this research agenda.
REFERENCES
- Corder S.P. The Significance of Learner’s Errors // International Review of Applied Linguistics. – 1967. – Vol. 5(4). – P. 161–170.
- Selinker L. Interlanguage // International Review of Applied Linguistics. – 1972. – Vol. 10(3). – P. 209–231.
- Akhmetova D.S. Reducing L1 Mediation in University-Level English Learners: A Semester-Long Intervention Study. – Astana: ENU Press, 2024. – 98 p.
- Zhumabayeva A.Zh. Trilingual Education in Kazakhstan: Policy, Practice, and Challenges. – Almaty: Atamura, 2023. – 220 p.
- Kroll J.F., Stewart E. Category Interference in Translation and Picture Naming: Evidence for Asymmetric Connections between Bilingual Memory Representations // Journal of Memory and Language. – 1994. – Vol. 33(2). – P. 149–174.
- Cook V. Second Language Learning and Language Teaching. 5th ed. – London: Routledge, 2016. – 320 p.
- Swain M., Lapkin S. Problems in Output and the Cognitive Processes They Generate: A Step Towards Second Language Learning // Applied Linguistics. – 1995. – Vol. 16(3). – P. 371–391.
- Krashen S. Principles and Practice in Second Language Acquisition. – Oxford: Pergamon, 1982. – 202 p.
- Lennon P. The Lexical Element in Spoken Second Language Fluency // Fluency in a Second Language / ed. H. Riggenbach. – Ann Arbor: University of Michigan Press, 2000. – P. 25-42.
- Nation I.S.P. Learning Vocabulary in Another Language. 2nd ed. – Cambridge: Cambridge University Press, 2013. – 502 p.
- Kaplan R.B. Cultural Thought Patterns in Inter-Cultural Education // Language Learning. – 1966. – Vol. 16(1–2). – P. 1–20.
- Atkinson D. Toward a Sociocognitive Approach to Second Language Acquisition // Modern Language Journal. – 2010. – Vol. 94(4). – P. 597–613.
- Mondria J.A., Mondria-De Vries S. Efficiently Memorizing Words with the Help of Word Cards and ‘Hand Computer’ // System. – 1994. – Vol. 22(1). – P. 47–57.
- Willis J., Willis D. Doing Task-Based Teaching. – Oxford: Oxford University Press, 2007. – 248 p.
- Schmitt N. Vocabulary in Language Teaching. – Cambridge: Cambridge University Press, 2000. – 224 p.
- Nunan D. Task-Based Language Teaching. – Cambridge: Cambridge University Press, 2004. – 222 p.
- Dörnyei Z. Research Methods in Applied Linguistics. – Oxford: Oxford University Press, 2007. – 336 p.
- Abutalebi J., Green D.W. Bilingual Language Production: The Neurocognition of Language Representation and Control // Journal of Neurolinguistics. – 2007. – Vol. 20(3). – P. 242–275.
Kassymzhan Sabina
Toilybai Akmaral
Turarbek Albina
Student, Foreign language faculty, Karaganda Buketov University, Kazakhstan, Karaganda
Sagadiyeva Kymbat Koshkinbaevna
Scientific supervisor, master of Pedagogics,
Karaganda Buketov University, Kazakhstan, Karaganda
