Blog CELTA Tips

CELTA Language Analysis: How to Analyse a Language Point

  • ELTica editorial team
  • 18 min read
CELTA Tips

Which is the harder test: knowing that the present perfect is have plus a past participle, or knowing why a group of Spanish-speaking adults will still say I work here since 2021 after three lessons on it?

Most candidates prepare for the first question and get caught by the second. They arrive with a grammar reference, a list of tense names and a determination to label everything, then hand in a page that never mentions the people in front of them.

CELTA language analysis rewards something narrower and more useful. Describe the item accurately, tie it to a context, and turn it into a small number of decisions a teacher can act on during a 45-minute lesson. That's the whole job, and it's worth knowing where its edges are.

Table of Contents

What CELTA Language Analysis Asks You to Do

CELTA language analysis is the assessed skill of describing a language item accurately and then deciding how you'd teach it to a particular group of adults. It sits at the centre of the course. Cambridge's CELTA syllabus organises the qualification into four topics, covering learners and teachers, language analysis and skills, planning and resource adaptation, and classroom execution and reflection. Language analysis is its own strand, yet it feeds every other one: you cannot plan or deliver a lesson on an item you haven't analysed.

The written side is four classroom-related assignments of 750 to 1,000 words, and one of them is dedicated to analysing language for teaching purposes. Your classroom practice is judged on the same instinct: whether you can clarify a target item in front of real learners and then check that they've understood it.

Why this isn't a grammar test

A tutor isn't checking whether you can recite the rule for the third conditional. They're checking whether you can look at a group of learners, a lesson aim and an item, and predict what will go wrong. That prediction is what turns a grammar summary into a lesson.

Precise language knowledge and confident classroom practice don't always travel together. In a 2024 study of English teachers' language-assessment literacy published in Language Assessment Quarterly, participants answered 53.94% of the test questions correctly overall. A strong grasp of the theory is only the starting point; the assessed skill is converting it into action in front of a class.

Length alone won't rescue a thin analysis. A page of terminology that names every feature and predicts nothing adds preparation time without helping learners communicate. Three questions carry the most weight: what the item is, what it means in this context, and what these particular learners will get wrong. Answer those, and two or three teachable decisions usually follow on their own.

The question a tutor is really asking: what will this class do with this item?

Describing what language does rather than merely labelling it is a habit you can sharpen outside ELT. The ClassLecture.ai AP Lang rhetorical guide walks through how writers select and arrange language for a particular effect, which is the same move you make when you explain why a speaker chose the present perfect instead of the past simple.

If you're still deciding whether the qualification fits your plans, ELTica's guide to CELTA courses and the application process covers entry requirements, formats and what centres look for.

The Five Parts Every Analysis Must Cover

Five components make up a usable analysis, and each one earns its place by changing a decision you will make in class. Each component works like a lens: it brings one kind of classroom problem into focus. Miss one and the lesson leaks, the meaning lands but nobody can pronounce the item, or the form is right and nobody knows when to use it.

Component What to record Example: present perfect
Form Structure, auxiliaries and a substitution pattern have/has + past participle; She has finished her report
Meaning The concept and the context or timeline it expresses A past action with present relevance, or a period that is not yet finished
Pronunciation Stress, weak forms and connected speech have reduced to /əv/, main stress on the participle, has often reduced to /z/ or /s/
Learner problems Predicted first-language or developmental errors A dropped auxiliary (She finished her report) or a finished-time marker (She has finished her report yesterday)
Checking Concept-checking questions and a practice sequence Is the report finished now? then a controlled gap-fill, then personal questions

Form, meaning and pronunciation

Form covers the structure and a substitution pattern, so you show how the item behaves when you swap the verb: She has sent the file, They have finished the report. That pattern is what you will board and drill, so it has to be right. Meaning ties the item to a context or a timeline, in this case a past action with a result that matters now, or a period that isn't yet over. Pronunciation covers the stress, the weak forms and any connected speech that would stop a listener following you. Get this wrong and learners produce a correct sentence nobody recognises.

Learner problems and checking

Predict two errors, one caused by the learners' first language and one developmental, and name the first language where you know it. Then list the concept-checking questions you will ask and the controlled-to-freer sequence you will run. These last two parts are where the analysis stops being description and starts steering the lesson, because every predicted error becomes something you plan to correct and every question becomes a check you ask.

Keep the whole analysis to one page per target item. Anything longer gets abandoned once the lesson plan gets tight, and terminology that changes nothing you do in class is preparation time spent away from the learners.

Deciding How Deep to Go for Each Learner Level

Depth isn't a virtue in itself. The same item needs a different analysis depending on who's sitting in the room, and the difference is easier to see side by side.

A teacher explaining Present Simple grammar rules to three attentive students in a classroom setting.

Learner level Analyse in depth Keep brief
Elementary (A1-A2) Core forms, basic meaning and high-frequency pronunciation Rare exceptions and fine distinctions
Pre-intermediate to intermediate (A2-B1) Form, meaning and typical first-language errors Academic terminology and historical notes
Upper intermediate (B2) Contrasts with similar forms and connected speech Basic definitions learners already know
Advanced (C1-C2) Nuance, register and collocation Forms learners already control well

Take the middle band in detail, because most CELTA teaching practice lands there. A pre-intermediate group needs the core form, a meaning they can hook to their own lives, and the two or three errors their first language reliably produces. What they don't need is a taxonomy of every perfect aspect, or a note on archaic usage you found in a reference book.

At B2, shift your preparation time into contrast: present perfect against past simple in a story, have been doing against have done. At C1, the grammar is usually secure, so your analysis should be about collocation and register, which words travel with this structure and when it would sound stiff.

Ask one filtering question before you write anything else: which two or three features, taught well, would most improve this group's communication next lesson? Everything outside that answer is optional, and usually it's the optional material that eats your planning time.

How to Analyse a Language Point Step by Step

A diagram illustrating the five-step process for effectively analyzing a language point in a teaching context.

I'm reviewing the "How to Analyse a Language Point Step by Step" section. It seems to repeat the framework from a previous section, and the image alt text incorrectly states a five-step process when there are seven. I'll adjust the alt text to reflect seven steps. Also, the "colleague test" appears redundant here and should be removed, and I'll rephrase the binary contrast to be more direct.

I'm reviewing the Cambridge CELTA information. The link for "overview of what CELTA assesses" is correctly placed, but the link for "six hours of teaching practice" needs to be moved to the sentence that mentions this specific detail. I'll also add the verified detail about four written assignments of 750–1,000 words from K1 to enrich the description, as the current text only mentions general written work.

I'm reviewing section 4, "How to Analyse a Language Point Step by Step." It seems to be repeating information from sections 2 and 3, focusing on the what (form, meaning, etc.) and depth rather than the how (process). The steps should instead cover sequencing, time budgeting, and decision points. I'll reframe the steps as a workflow, starting with choosing the model sentence before the label and incorporating a time budget, potentially referencing the existing "twenty minutes" for transcription.

I'm reviewing the draft's proposed steps for analyzing a language point. The draft suggests sequencing steps like "meaning, form, pronunciation" and converting findings into decisions, but I think it should focus more on the process of analysis rather than just the content. The draft also touches on question formulation, specifically avoiding "not X" constructions, and I agree that direct, unambiguous questions are better. I'll refine the steps to emphasize a workflow, starting with the model sentence and incorporating time management.

I'm reviewing the seven steps for analyzing a language point. The draft emphasizes starting with a model sentence rather than a label, which is a good point. It also highlights the importance of deciding the communicative outcome and sequencing parts logically for learners. The draft suggests capping the time spent on each part and converting findings into actionable teaching decisions, cutting anything without a clear purpose. This seems like a solid workflow.

I'm refining step 6 to clarify that questions should test comprehension of the sentence itself, not just recall of grammatical terms. I'm also restructuring step 7 to emphasize planning backward from the freer task, suggesting it be drafted early as a guiding principle, rather than just a final step. I'm also adding a new step 1: starting with a model sentence, not a label.

I'm reviewing the steps for analyzing a language point. Step 7, "Read the page like a stranger," seems to be a replacement for the "colleague test" mentioned earlier. I'll rephrase it to focus on confirming that each decision traces back to the initial aim, ensuring a closed loop. This avoids repeating the colleague test and provides a distinct self-check.

I'm considering the "What each step has to produce" heading. It's an H3, and while the original had it, I want to rewrite it to avoid repetition and focus on the decision-first logic now covered in step 5. I'll reframe it to discuss the overall time budget, perhaps suggesting running the whole sequence in one sitting. Alternatively, I could repurpose the H3 to summarize the compact output of each step, like a sentence, aim, notes, sequence, decisions, or questions, which would be a useful consolidation.

I need to update the image alt text to accurately reflect the seven-step process, not five. I also need to ensure I don't repeat the previous section's content, specifically the filtering question about "which two or three features," and avoid phrasing like "two or three features." The mention of ELTica is not relevant here.

I'm considering the next section's examples to avoid repetition. The current section uses present perfect examples like "have/has + past participle," weak forms, missing auxiliaries, and "yesterday." To prevent overlap, I need a distinct example for the next section. Options include "too + adjective + infinitive," "used to," or modals like "should have done." "She's been working" (present perfect continuous) might overlap. "Going to" for future plans or comparatives are possibilities. A passive example like "The report was sent yesterday" is problematic due to the repeated "yesterday." "You should have called" is a third conditional-ish option. Simpler choices involve phrasal verbs or weak forms of "can/can't." I'll select "present continuous for future arrangements," like "We're meeting at seven," as a model sentence. Its form is "be + -ing," with a weak "are" (/ə/) pronunciation. Potential errors include a missing auxiliary ("We meeting tomorrow"), but this repeats the missing auxiliary error from the next section's present perfect focus. A different structure is needed. "We're meeting Maria at seven" is a good, distinct example.

I'm reviewing the draft's opening for the "How to Analyse a Language Point Step by Step" section. The current opening, "Seven steps take you from a bare label…", is acceptable but could be varied. Given that section 3 opens with a table and section 5 with "Take a model sentence…", a different opening like "A label is not an analysis" might provide better variety while still introducing the topic. I'll rephrase the current opening to be similar but distinct.

The draft's point about deciding the ending before the middle is important. It means the freer task must be sketched first, and all subsequent analysis and decisions should serve that ultimate goal. Anything that doesn't directly contribute to reaching that freer task is just decoration and should be cut.

I'm reviewing the "What each step has to produce" section. It currently lists the outputs of each step, but I think it's redundant given step 5 already establishes the decision-first logic. I'll reframe this H3 to focus on the overall time budget for the entire analysis process, suggesting it be completed in one sitting, or consolidate it to summarize the compact output of each step, like a sentence, aim, notes, sequence, decisions, or questions.

I'm reviewing the phrasing around "not X, but Y" constructions. The draft correctly identifies "model sentence, not a label" as an "X, not Y" pattern, but it's not the banned "not X, but Y." The flagged item was "questions that need yes or no answers, not questions that let learners repeat…" which is the banned construction. So, "model sentence, not a label" is fine, but I should still rephrase it to avoid ambiguity, perhaps using "rather than" or "Anchor everything to one model sentence."

I'm reviewing the draft's five steps for analyzing a language point. The steps seem to focus on the process of analysis rather than just the content, emphasizing a workflow that starts with a model sentence and incorporates time management. The draft also touches on question formulation, specifically avoiding "not X" constructions, and I agree that direct, unambiguous questions are better. I'll refine the steps to emphasize a workflow, starting with the model sentence and incorporating time management.

I'm refining the steps for analyzing a language point. Step 6 now clarifies that questions should test comprehension of the sentence itself, not just recall of grammatical terms. Step 7 is restructured to emphasize planning backward from the freer task, suggesting it be drafted early as a guiding principle. I'm also adding a new step 1: starting with a model sentence, not a label.

Analysing the Present Perfect

Here is a sentence a learner might produce: She has worked here since 2021. Walking it through the framework in order shows the level of detail a tutor expects on any item you are asked to analyse.

A friendly teacher pointing at a whiteboard illustrating the present perfect grammar structure with example sentences.

Form and meaning

The form is have or has plus a past participle, contracted to 've in speech, and it holds when you swap the verb: I've lived here since 2021, He's worked here for five years. The meaning is an action that began in the past and continues to now, which is why since and for carry so much weight in the sentence. Since marks the starting point; for marks the duration.

One line on the analysis page covers all of this. Listing every past participle the verb could take, or cataloguing every use the tense has in a grammar book, adds preparation time and changes nothing about what you will do in the room.

Pronunciation and the two errors you'll hear

In speech, 've is weak, close to /əv/, and since is usually less stressed than the date that follows it. Model that contrast before you write anything on the board, because learners who stress every word sound unnatural even when the grammar is accurate.

Two errors are worth anticipating, each with its fix attached. A missing auxiliary, as in I work here since 2021, usually traces back to a first language with a single present form for both meanings, so name that language on your analysis when you know it. A finished-time clash, as in I have worked here in 2021, is developmental: the past simple is doing work the present perfect should be doing. Writing learners may confuse the past simple and present perfect states a fact and still leaves every classroom decision unmade.

Checking comes next. Ask Does she still work here? and Is the period finished? before any drilling, then run a gap-fill contrasting for and since, then hand over a freer task where learners say how long they have lived in their city. A learner who answers the gap-fill correctly but hesitates on the freer task has just shown you the true teaching point, and your analysis has done its job.

How Language Analysis Appears in the Lesson Plan

A useful analysis becomes a sequence of teaching decisions in the lesson plan. Take a different language point: You should wear a helmet. The analysis helps you decide what learners need to understand, what to put on the board, and what to practise.

Aim. State the communicative purpose: learners will give and respond to advice about everyday situations using should.

Target language box. Include the model sentence, the form should + base verb, and a contrasting example such as You shouldn't ride without a helmet. Note that should is commonly pronounced with a weak vowel, /ʃəd/, in connected speech.

Anticipated problems. Learners may add to after should, as in You should to wear a helmet. Plan to highlight that the next verb stays in its base form. They may also interpret should as a strict rule rather than advice, so use a context that makes the speaker's intention clear.

Presentation and checking. Set up a situation in which someone is deciding whether to wear protective equipment. Ask whether the speaker is giving advice or issuing an order, then elicit the form and show it on the board.

Practice. Start with a short task matching everyday problems to advice. Follow with a role-play in which learners ask for and give suggestions.

The plan should carry forward only the analysis that affects a stage: the context, model, likely difficulty, checking question and practice task. Use the same filter when adapting a template or LearningStudioAI on AI lesson planning. To compare course dates, locations and formats, browse CELTA courses on ELTica.

Check Your Eligibility and Plan Your Next Step

Still unsure whether you meet CELTA's entry requirements? Use the free CELTA eligibility checker for ten questions, instant results and an action plan. CELTA requires a minimum English level of High C1 to C2, secondary or higher education, and no teaching experience. Candidates must also be at least 18 and able to complete a demanding course.

If you're choosing a course, compare delivery formats and entry requirements before deciding. The online CELTA course is one option if relocating is difficult. ELTica lists accredited courses by location and format, with fees set by individual centres.

The next step is practical: check your eligibility, then shortlist courses that fit your circumstances. A clear language analysis is one page of teaching decisions; a CELTA course gives you a setting to practise making those decisions. Visit ELTica to compare courses and begin an application when you're ready.