Interlanguage is the developing language system a learner builds while acquiring a second language. It is separate from both their first language and the target language, and it follows rules of its own. Larry Selinker named it in 1972, reframing learner errors as evidence of a grammar in progress rather than plain failure.
The short version
Before 1972, a learner’s incorrect English was usually explained in one of two ways. Either it was interference from the mother tongue, the assumption behind contrastive analysis, or it was simply a deficit, a gap where correct English should be. Both explanations treat the learner’s output as a damaged version of the target language.
Selinker proposed something different. The learner, he argued, constructs a separate linguistic system, complete in itself, sitting between the first language and the target language. It is not English with holes in it and it is not the first language in disguise. It is a third system with its own internal logic, and it changes shape as learning proceeds. Selinker called it interlanguage. Around the same time, William Nemser described the same object as an approximative system, and S. P. Corder had already argued that learner errors were systematic rather than random.
The claim matters because it changes what an error is. If the learner has a system, an error is a window into that system rather than noise to be removed. Corder’s phrase for this was that errors are significant: they tell the teacher what the learner currently believes about English.
Learner errors cluster. A class that has just met the past simple will produce goed, buyed and teached in the same week, from learners who have never heard any of those forms from anyone. The forms are invented, but they are not arbitrary. They are the regular past tense rule applied where English happens to be irregular.
Accuracy often dips before it climbs. A beginner may produce went correctly as an unanalysed chunk, then start producing goed once the past simple rule arrives, then return to went months later with the rule and the exception both in place. This is the U-shaped curve familiar from acquisition research, and it is the single most useful thing to know when a class appears to have got worse after a good lesson.
The system takes material from several sources at once. Selinker named five: transfer from the first language, transfer of training, which is influence from how the language was taught, strategies of learning, strategies of communication, and overgeneralisation of target language rules. A single error can come from any of them, which is why a confident guess about cause is often wrong.
Fossilisation is Selinker’s term for an interlanguage that has stabilised short of the target and stopped developing, despite continued exposure and instruction. The long-term resident who has spoken fluent, successful, permanently article-free English for twenty years is the standard example. Whether fossilisation is genuinely permanent or only looks permanent is contested, and that argument is covered below.
A B1 class of sixteen adults, mixed first languages, four weeks into a general English course. The teacher sets a speaking task about last weekend and collects what the learners actually produce. Three learners generate three different errors, and the useful move is to treat each one as a reading of a different system rather than as three instances of bad grammar.
| What the learner said | What the system is doing | What the teacher should do next |
|---|---|---|
| “I goed to my sister house.” | Overgeneralisation. The regular past rule has been acquired and is being applied everywhere, including where English is irregular. | Do not reteach the past simple. Treat irregular forms as vocabulary and give retrieval practice on the twenty highest frequency ones. |
| “I went to the my sister’s house.” | Transfer. The learner’s first language allows a determiner and a possessive together, so the interlanguage permits it too. | A single contrastive noticing task on determiner slots, aimed at that learner rather than the whole class. |
| “Last weekend I have visited my sister.” | Transfer of training. The present perfect was drilled heavily the week before and is now the default past form in production. | Contrast the two tenses on a time line with finished time markers. The problem is the boundary, not either tense. |
All three learners produced a past tense error. Correcting all three the same way, by supplying the right form, would have taught nobody anything about their own system. Reading the interlanguage first turns one generic correction into three different and much shorter interventions.
| Pair | The difference that matters |
|---|---|
| Interlanguage vs L1 interference | Interference is one input into interlanguage, not a synonym for it. Transfer explains some errors. Overgeneralisation, training and communication strategies explain the rest. |
| Interlanguage vs an error | An error is a single observable output. Interlanguage is the underlying system that produced it. You correct errors; you teach into the system. |
| An error vs a slip | Corder’s distinction. A slip is a performance lapse the learner can self correct when pointed at. An error reflects the current system and cannot be self corrected, because the learner believes it is right. |
| Interlanguage vs fossilisation | Interlanguage is the system. Fossilisation is one possible fate of it, where development halts short of the target despite continued exposure. |
| Fossilisation vs stabilisation | Stabilisation is a plateau that may still shift. Fossilisation claims permanence. Michael Long argued that almost every classroom case labelled fossilisation is really stabilisation, because permanence needs years of evidence to demonstrate. |
| Interlanguage vs a pidgin | A pidgin is a shared community language with its own speech community. An interlanguage belongs to one individual learner and changes as that learner changes. |
Selinker’s 1972 paper in International Review of Applied Linguistics in Language Teaching is the source of the term, but the ground had been prepared. Corder’s 1967 paper argued that learner errors were systematic and worth analysing rather than eliminating on sight. Nemser, writing in 1971, described learner speech as an approximative system with its own structure. Selinker’s contribution was to name the system, to propose the five processes that shape it, and to attach the claim about fossilisation.
The concept is now standard in second language acquisition and sits behind most modern thinking on error correction. It is not uncontested, and the page is more useful if it says where.
Variability. Elaine Tarone and others showed that a learner’s interlanguage is not one stable system but shifts with task, audience and attention. The same learner produces third person s reliably in a careful written exercise and drops it in free speech. If the system changes with the task, the idea of a single grammar to describe becomes harder to hold.
Fossilisation is difficult to prove. Demonstrating that a form will never change requires evidence over years, and most studies do not have it. Michael Long’s argument is that stabilisation is what is usually observed and fossilisation is what is usually claimed. ZhaoHong Han’s work maps how narrow the genuinely fossilised cases turn out to be once the criteria are applied strictly. For a teacher, the practical consequence is worth stating plainly: describing a learner as fossilised is a judgement that the evidence in an ordinary classroom almost never supports, and it tends to become self fulfilling.
It describes more than it predicts. Interlanguage tells you that a system exists and that errors are systematic. It does not tell you which error a given learner will make next, or which intervention will move them on. That is a real limit, and it is why interlanguage informs diagnosis rather than replacing a syllabus.
Sources are given in full below. Where no stable public link was confirmed at the time of writing, the source is cited in print form rather than linked.
| Source | What it establishes |
|---|---|
| Selinker, L. (1972). Interlanguage. International Review of Applied Linguistics in Language Teaching, 10(3), pages 209 to 231. | The original statement of the term, the five processes, and fossilisation. |
| Corder, S. P. (1967). The Significance of Learners’ Errors. IRAL, 5(4), pages 161 to 170. | Errors as systematic evidence of a developing competence, and the error and slip distinction. |
| Nemser, W. (1971). Approximative Systems of Foreign Language Learners. IRAL, 9(2), pages 115 to 123. | The parallel formulation of learner language as a structured system in its own right. |
| Tarone, E. (1983). On the Variability of Interlanguage Systems. Applied Linguistics, 4(2). | Interlanguage varies with task and attention, which complicates the single system account. |
| Long, M. H. (2003). Stabilization and Fossilization in Interlanguage Development. In Doughty and Long, The Handbook of Second Language Acquisition. | The case that most claimed fossilisation is stabilisation and the evidential bar is rarely met. |
| Han, Z. (2004). Fossilization in Adult Second Language Acquisition. Multilingual Matters. | A book length treatment of what fossilisation requires before the label can be applied. |
| Council of Europe, Common European Framework of Reference for Languages | The proficiency scale teachers use to describe where a learner’s system currently sits. Checked 11 September 2026. |
| British Council TeachingEnglish, Pathways: Knowing the subject | Professional development material on language systems and learner language. Checked 11 September 2026. |
Three practical consequences follow, and none of them requires a new syllabus.
It is the version of English a learner has built so far. Not broken English and not their first language, but a working system of its own with rules the learner is applying consistently. It changes as they learn, and their errors show which rules are currently in place.
Larry Selinker named it in a 1972 paper in International Review of Applied Linguistics in Language Teaching. He built on S. P. Corder, who argued in 1967 that learner errors were systematic, and on William Nemser, who described learner language in 1971 as an approximative system.
No. Interference, or transfer, is one of the forces that shapes interlanguage, but it is not the whole of it. Selinker listed five processes, including overgeneralisation of English rules and the effect of how the language was taught. Many learner errors have nothing to do with the first language at all.
Fossilisation is Selinker’s term for an interlanguage that has stopped developing short of the target, even with continued exposure and teaching. It is contested. Michael Long argued that most cases teachers call fossilisation are really stabilisation, a plateau that can still shift, because proving permanence takes years of evidence.
No. An error that reflects a rule the learner is currently building is worth addressing. A one off slip the learner could fix themselves is not. During fluency stages, selective and delayed correction aimed at the underlying pattern is generally more useful than correcting each instance as it happens.
Mainly by changing where the evidence comes from. Instead of planning the next grammar point from the coursebook order, collect learner output first, read the patterns in it, and let that decide what the next lesson targets. The syllabus stays; the sequence inside it becomes responsive.
Interlanguage sits alongside two closely related entries in the TEFL teaching methods and classroom techniques glossary: error correction and recasting, which covers what to do with the errors interlanguage produces, and L1 interference, which covers the transfer process specifically. The affective filter explains why heavy correction of a developing system can slow it down.
For the qualification and career terminology that sits outside this glossary, see The TEFL Institute’s guide to TEFL qualifications, and its account of the grammar translation method, the tradition interlanguage research was reacting against. For Irish school and Transition Year procedure, see The TEFL Institute of Ireland on Transition Year TEFL programmes.
How to cite this definition
tefl.ai (2026) “Interlanguage”, TEFL Teaching Methods and Classroom Techniques Glossary. Available at: https://tefl.ai/interlanguage/ (Accessed: 11 September 2026).
Last updated 11 September 2026. Reviewed against the sources listed in the reference table.
The past simple is the English verb form used for an action or state completed at a finished time in the past. Most verbs form it by adding -ed, while a closed set of common verbs change their form instead. It is the ordinary narrative form of English and the first past form most learners meet.
The present perfect is the English verb form made with have or has plus a past participle, used to link a past action to the present moment. It covers experience up to now, a situation that began in the past and still continues, and a past event whose result matters now. It never takes a finished time reference.
