# speaking_part_three_guide - **Original Source:** `speaking_part_three_guide.txt` - **Category:** `speaking` - **Original Format:** `.txt` --- IELTS SPEAKING PART THREE. THE COMPLETE AUDIO GUIDE. This is one file. It replaces the compiled notes you were working from, which drew on seven or eight different sources and repeated the same four formulas under different names. Everything useful in that compilation is here, deduplicated, corrected where it was wrong, and followed by ten complete model answers so you can hear the length rather than only read about it. Written to be read aloud by text to speech software. Listen while commuting. The model answers are the part that matters most, so if you only have time for a fragment of this file, play the section containing them. WHAT PART THREE ACTUALLY IS. A discussion, not a monologue. The examiner asks, you answer, the examiner asks again. Four to six exchanges over roughly four or five minutes. Unlike Part Two, you are not trying to fill a fixed block of time, and you are not trying to be stopped. The questions escalate deliberately. The first set asks you to describe or compare, and targets around band six. The second set asks you to analyse, evaluate, or explain causes, and targets band seven. The third set asks you to hypothesize, predict, or discuss something abstract, and targets band eight or nine. This means something reassuring. If the questions turn strange or difficult near the end, that is not the examiner losing patience. That is the examiner checking whether your ceiling is higher than the previous answers proved. Difficult questions late in Part Three are a compliment. The examiner may also push back on what you said. Some people would argue the opposite. Does that always hold. That is a probe, not a disagreement, and there is a section on handling it further down. You may ask for one clarification. Sorry, could you rephrase that. Use it once, and only if you genuinely did not understand. CALIBRATION. YOUR TWO FAILURE MODES. Your recorded habits are clipping and drifting, and Part Three punishes both harder than any other section, because the answers are long enough for drift to develop and abstract enough for clipping to look like you have no ideas. The target. Four to five sentences. Roughly one hundred to one hundred and thirty words. Around forty five to fifty five seconds. Under three sentences, you have not shown analytical range and the examiner will move on early, which costs you turns. Over about seven sentences, three things go wrong at once. You start a second argument you cannot finish. The tense drifts, because you have left the frame the question set. And you use up the exam time in which you could have demonstrated range across more questions. A long answer is not a generous answer. The structural fix for drift is built into the shape below. The final move in every answer is a concession, and the concession is your exit. Once you have conceded, you stop. That is the whole discipline. CORRECTIONS TO THE COMPILED NOTES. Seven items. Each of these is either an internal contradiction in the source material or advice that will lower your score if you follow it. Correction one. The no I rule is overstated and, as written, dangerous. The notes say to frame arguments around demographics and institutions rather than yourself, and then supply dozens of phrases beginning with I would say, I tend to think, and in my humble opinion. Both cannot be right. The real rule is narrower. Do not use personal anecdote as evidence. Your cousin, your commute, and your weekend are not data. But the examiner is explicitly asking for your view, and refusing to give one sounds evasive rather than academic. Say I think, say I would argue, say my sense is. Then support it with something societal rather than something autobiographical. The correct target is a personal opinion supported by impersonal evidence. Correction two. Four competing formulas is worse than one. The notes give you A W E A, R E E, A E E C, Multiple Perspectives, and C T M E, several of them twice. Under pressure you cannot select between five templates, and the selection itself burns the seconds you needed for content. They are largely the same shape anyway. This guide uses one spine, described below, and treats the others as what they actually are, either abbreviations of it or content generators rather than structures. Correction three. The think like a billionaire hack should be dropped. It carries two risks. It pushes you toward naming real living individuals, which invites factual claims you cannot support, and the sample answers it produces in the notes are contemptuous about ordinary people, including a passage about workers having nothing better to do than scroll shopping sites to fake a fancy lifestyle. Content is not assessed, but contempt is audible, and it makes the register wrong. The useful part of that hack is only the shift from individual to institutional framing, and the three lens method below does that cleanly with no celebrity attached. Correction four. The wig answer is a bad model. The notes suggest that when the examiner asks something bizarre, you should answer bluntly, and give as their example a claim that people who wear wigs look ridiculous and only want attention. That is a rude, one clause answer that dead ends, which is exactly the failure the strange question was designed to expose. The correct response to a strange question is the honest ignorance protocol further down, which the same notes also contain, and which is far better. Correction five. Never self interrupt is too strong, and it contradicts the Part One track. The band descriptors reward self correction that resolves cleanly. What lowers your fluency score is looping, repairing the repair, or apologising for the slip. Repair once, briefly, and continue. Sorry, took, not take. Then keep going. Do not repair the same sentence twice and never say sorry twice. Correction six. Half of the supplied phrase bank is written English, not spoken English. Phrases like I am of the firm conviction that, it is abundantly clear to me that, a critical analysis of this trend reveals that, and to illustrate this point we need look no further than are essay openers. Spoken aloud in a conversation, they sound recited, and the same notes warn you at length that examiners are trained to detect exactly that. There is a corrected phrase bank below. It is much shorter on purpose. Ten natural phrases you can actually deploy are worth more than fifty you have to remember to insert. Correction seven. The length advice is inconsistent. The notes say four to six sentences in one place, and elsewhere import the Part Two instruction to keep talking until you are stopped. That second instruction does not apply here. In Part Three you finish your answer and hand the turn back. THE SPINE. ASSERTION, MECHANISM, EVIDENCE, CONCESSION. Four moves, in this order, every time, whatever the question. Move one, the assertion. Answer the question in one sentence, immediately, with a position in it. No preamble, no repeating the question back. Move two, the mechanism. Explain why or how. This is the move that separates band seven from band six, because it shows you can explain a process rather than only state a preference. Move three, the evidence. One generalized, real world illustration. A country, an industry, a demographic, an institution. Never a person you know. Move four, the concession. Name the limit of what you just said. This does two jobs. It demonstrates analytical range, and it ends the answer for you. The other formulas mapped onto this spine. R E E is this shape with the concession removed, which is why it produces shorter, weaker answers. A W E A is the same four moves under different letters. C T M E is not a structure at all, it is a quality checklist for the words you choose while executing the spine. Multiple Perspectives is not a structure either, it is a way of generating content when you cannot think of a mechanism. One spine. Two helpers. That is all you need to hold. THE SIX QUESTION ARCHETYPES. The spine does not change. What changes is what goes into moves two and three. Archetype one. Comparison, past against present, or one group against another. The mechanism explains what changed and why. Use used to, tended to, and whereas. Watch the tense, because this archetype invites you to start in the past and forget to return to the present. Archetype two. Cause. Why does this happen. The mechanism is the whole answer, so give the principal driver first and, if you have room, one compounding factor. Do not list three causes. Two, explained, beats four, named. Archetype three. Evaluation or solutions. What should be done. Split the answer across two levels, the institutional and the individual. On an institutional level, governments can do this. At the same time, individually, people can do that. The concession then observes that one level cannot work without the other. Archetype four. Prediction. What will happen. Hedge every claim. It is likely that. We will probably see. If governments fail to act, then. Never predict with certainty, because certainty about the future reads as naivety rather than confidence. Archetype five. Agree or disagree. Concede the opposing case first, briefly, then commit. There is some truth to that, but the stronger factor is. This ordering matters, because conceding first and committing second sounds decisive, whereas committing first and conceding second sounds like retreat. Archetype six. Abstract or hypothetical. Will this define the century. What would happen if. These arrive late and are the band eight and nine probes. Reframe them into something concrete you can discuss, and say that you are doing so. That depends on what we mean by. If we take it to mean, then. IDEA GENERATION. THREE MOVES, NOT SEVEN. When the topic is unfamiliar, you do not need an opinion about it. You need a lens. Lens one, the three perspectives. Look at the question as the individual, the economy, and society or the environment. Almost any topic yields something under all three. Private cars are convenient for the individual, valuable to a manufacturing economy, and damaging environmentally. That single move generates a mechanism, an example, and a concession in one pass. Lens two, institutions against individuals. Ask who has the power here and who bears the cost. Governments, employers, schools, and hospitals on one side, and households, workers, students, and patients on the other. Any question about a social problem can be answered by describing how those two levels interact, and it produces the exact vocabulary examiners are listening for. Lens three, the professional generalization. You know a great deal about clinical training, public health systems, and how institutions supervise people. That knowledge is usable, but only if you strip the first person from it. Do not say when I write evaluations for my residents. Say when clinical supervisors design evaluation criteria for trainees, the difficulty is usually that. Same expertise, correct register, and it will be more specific than anything a general candidate can produce. That third move is the most valuable thing in this file for you specifically. Your advantage in Part Three is that you have genuine institutional knowledge. The only barrier is the pronoun. THE CORRECTED PHRASE BANK. Short on purpose. Learn these and stop. To open with an assertion. I would say. My sense is that. The main thing is probably that. Broadly speaking. It really depends on, when the question is genuinely conditional. To give the mechanism. Mainly because. The main driver is probably. What tends to happen is that. Which in turn means that. To give the evidence. For instance. In a lot of countries. You see that in. A good example would be. To concede and exit. Having said that. Then again. That said. Though obviously that only applies to. To hedge a prediction. It is likely that. We will probably see. I suspect. If that continues, then. To buy two seconds. That is a difficult one. I have not thought about that before, but. Hmm, it depends what you mean by. Phrases to avoid, all of which appear in the compiled notes and all of which are written register. I am of the firm conviction that. It is abundantly clear to me that. There is compelling evidence to suggest that. A critical analysis of this trend reveals that. A clear manifestation of this is. To illustrate this point, we need look no further than. Without a shadow of a doubt. A textbook example of this phenomenon can be found in. Also avoid firstly, secondly, and in conclusion. Those are essay connectors. Say the main reason is, another thing is, and so on the whole. GRAMMAR TARGETS FOR PART THREE. Three things to make audible, chosen because they are where your marks are actually available. One, the cleft sentence. It is the media that shapes this, rather than the media shapes this. It is the cost of housing that drives the decision. One or two per Part Three, no more, and they carry a visible band eight marker each. This is the same structure you drill in Task Two, so it costs you nothing new. Two, agreement with abstract singular subjects. Part Three runs almost entirely on abstract nouns, which is why the s disappears here more than anywhere else. Access to healthcare determines. Migration drives. The cost of housing shapes. Automation replaces. Technology changes. Every one of those is singular with a plural noun sitting nearby. Three, tense control in comparison questions. When the question compares past with present, you will start in the past and stay there. The fix is structural. Put the past in the mechanism sentence and the present in the evidence sentence, and let the paragraph shape move you rather than trying to remember mid sentence. EMERGENCY PROTOCOLS. Protocol one, honest ignorance with an attempt. Use this for a strange or unfamiliar topic. To be honest, this is not something I have thought about much. But if I had to guess, I would say. Then answer for three sentences. The examiner is assessing whether you can navigate the unknown in English, and admitting ignorance fluently while still producing an argument is a strong performance, not a weak one. Protocol two, the conditional answer. Use this when the question demands a yes or no that neither answer fits. It really depends on. If the situation is this, then this follows. Whereas if it is that, the outcome is different. That is a full answer with an assertion, two mechanisms, and a built in concession. Protocol three, the geographical narrowing. Use this when the question is too broad to grip. I cannot speak for everywhere, but at least in Mexico, what has changed most is. Narrowing is not evasion. It is what a careful speaker does with an unanswerably broad question. Protocol four, the definitional reframe. Use this for abstract questions. That depends what we mean by success. If we take it to mean financial security, then. If we mean something closer to autonomy, the answer is different. This is the single most effective move against a band nine probe. HANDLING PUSHBACK. The examiner will sometimes challenge you. But some people would say the opposite. Do you think that is always true. Is that not just a matter of money. This is a probe, and it is a good sign. Do not abandon your position and do not defend it rigidly. Concede the specific point, hold the general one. The shape. That is fair, and in the case of X it probably does hold. What I would say though is that in most cases the stronger factor is still Y. Three phrases for this. That is fair, though I would still argue. There is definitely something in that, but. I would agree up to a point, and the point where it stops holding is probably. TEN MODEL ANSWERS. Each gives the question, the archetype, the answer, and what to notice. Every answer sits between one hundred and one hundred and thirty words. Say the assertion out loud before the voice does, then listen to how the rest of the answer is built. Model one. How has the way people work changed compared with a generation ago. Archetype, comparison. Here is the answer. I would say the biggest change is that work has stopped being tied to a place. A generation ago, almost every job required you to be physically present at a particular site for fixed hours, and the employer controlled the schedule. Now a large share of professional work needs little more than a laptop and a connection, so employers are effectively buying output rather than attendance. You see that in the amount of office space large companies have given up over the last few years, and in the fact that some of them now recruit across borders rather than within one city. Having said that, this only really applies to office work. For anyone in manufacturing, healthcare, or transport, the job is exactly as tied to a place as it always was. End of answer. One hundred and thirty two words. What to notice. The assertion arrives in the first eight words with no preamble. The mechanism sits in the past, the evidence returns to the present, and the paragraph shape carries the tense rather than your memory. The concession is not a hedge, it is a genuine limit, and it ends the answer cleanly. Agreement watch. A large share needs, singular. Employers are buying, plural. Model two. Why do people move from rural areas to cities. Archetype, cause. Here is the answer. The main driver is almost certainly employment. Rural economies tend to depend on a small number of industries, usually agriculture, and those industries have been mechanising for decades, so the number of jobs falls even when output rises. A young person who trains for anything else has no local employer to work for. Another factor that compounds it is that services follow the population rather than leading it, so once a place starts losing people, the school and the clinic close, which gives the people who stayed a reason to leave as well. In a lot of countries you can see that as a fairly predictable sequence rather than a sudden collapse. Then again, the flow is not always permanent, and some of it reverses when housing in the cities becomes unaffordable. End of answer. One hundred and thirty two words. What to notice. Two causes, both explained, rather than four named. The second is introduced as compounding rather than merely additional, which is what an analytical answer sounds like. Agreement watch. The number of jobs falls, singular, with the plural word jobs immediately before the verb. Services follow, plural. Model three. What can governments do to reduce stress in modern life. Archetype, evaluation and solutions. Here is the answer. I think it has to work on two levels at once. On an institutional level, the most effective measures are usually regulatory rather than therapeutic. Limiting working hours, protecting time off, and requiring employers to cover mental health support all change the conditions people live in, rather than asking them to cope better with conditions that have not changed. At the same time, on an individual level, there is a limit to what regulation reaches, because a great deal of stress comes from things like caring responsibilities that no employment law touches. For instance, several European countries now legislate a right to disconnect outside working hours. Though obviously a law like that only helps people in formal employment, which in a lot of countries is a minority of the workforce. End of answer. One hundred and thirty words. What to notice. The two level split is doing the structural work, and the concession attacks the coverage of the solution rather than its merit, which is a more sophisticated objection. Agreement watch. A great deal of stress comes, singular. Several countries legislate, plural. Model four. Do you think money is the main reason people choose a career. Archetype, agree or disagree. Here is the answer. There is some truth to that, particularly at the start of a career, when debt and housing costs weigh heavily and people take whatever pays. But I would say the stronger factor over a whole working life is status, in the broad sense of how a job is regarded and how much control it gives you. That is why people leave well paid roles for worse paid ones fairly regularly, and why some of the hardest professions to recruit for are not the lowest paid ones. In a lot of countries, public sector work pays clearly less than the private equivalent and still attracts people. Then again, that is a luxury argument. Where wages are near subsistence, money is not the main factor, it is the only one. End of answer. One hundred and twenty eight words. What to notice. The concession comes first and the commitment second, which sounds decisive. The final concession is the strongest sentence in the answer, because it identifies the condition under which the whole argument fails. Agreement watch. Public sector work pays, singular. The hardest professions are, plural. Model five. Will artificial intelligence replace doctors. Archetype, prediction. Here is the answer. Not replace, I suspect, but reshape fairly drastically. What these systems are good at is pattern recognition across enormous volumes of data, so the parts of medicine that are essentially classification, reading a scan or flagging an abnormal result, are likely to be automated first, and to some extent already are. What they are weaker at is everything depending on an unreliable account from a distressed person, which is much of clinical work. So the likely outcome is a redistribution of tasks rather than a replacement of the profession. If that continues, the skill that becomes scarce is not diagnosis but judgement about which diagnosis matters. Having said that, forecasts of this kind have a poor record, and most of them have overestimated the speed and underestimated the direction. End of answer. One hundred and twenty nine words. What to notice. Every prediction is hedged. I suspect, are likely to be, the likely outcome, if that continues. The two cleft style structures, what these systems are good at, and what they are much weaker at, are doing exactly what a cleft sentence does, placing the emphasis deliberately. The final concession is scepticism about prediction itself, which is a strong band nine move. Agreement watch. The skill becomes, singular. Forecasts have, plural. Model six. Is it easier to become famous now than in the past. Archetype, comparison. Here is the answer. Far easier to become known, I would say, and much harder to stay known. In the past, fame ran through a small number of gatekeepers, so a newspaper, a studio, or a record label decided who reached an audience at all, and that scarcity made the fame durable once you had it. Now anyone can reach an audience directly, so the barrier has essentially disappeared, but the attention is divided across so many people that very little of it lasts. You see that in how quickly the most watched account of one year becomes unrecognisable two years later. Then again, the older kind of fame has not disappeared, it has just narrowed, and the people who hold it are probably more powerful than their equivalents were before. End of answer. One hundred and twenty seven words. What to notice. The assertion contains a distinction, easier to become known and harder to stay known, which gives the rest of the answer its structure. Distinctions are the cheapest way to sound analytical. Agreement watch. The attention is divided, singular. The people who hold it are, plural. Model seven. Why do people collect things. Archetype, cause. This is the strange question, the anti memorisation probe. Here is the answer. That is a difficult one. I have not really thought about it before, but if I had to guess, I would say it is mostly about imposing order on something. A collection has rules, it can be completed or not completed, and it gives a person a small domain where the outcome depends entirely on their own effort, which is not true of most of the rest of life. There is probably a social element too, because collections tend to come with communities around them, and a lot of the pleasure seems to be in the comparing rather than the owning. Having said that, I imagine the motivation varies enormously between someone collecting for investment and someone collecting for nostalgia, and I would not want to treat those as the same behaviour. End of answer. One hundred and thirty two words. What to notice. Honest ignorance is admitted in eleven words and then abandoned. Everything after it is a real argument. This is exactly what the strange question is testing, and freezing or answering in one dismissive line is the failure it is designed to catch. Agreement watch. A collection has, singular. Collections tend, plural. A lot of the pleasure seems, singular. Model eight. Are smartphones good for children. Archetype, agree or disagree, answered conditionally. Here is the answer. It really depends on what they displace. If a phone replaces time that would otherwise have gone on sleep, unstructured play, or face to face contact with other children, then the effect is clearly negative, and the evidence on sleep in particular is fairly consistent. Whereas if it replaces passive television, the difference is much smaller than people assume. So the useful question is probably not whether children should have them, but what the alternative use of that time actually is. For instance, most school policies focus on banning devices during lessons, which is the easiest part of the day to regulate and probably not the part that matters. That said, there is one asymmetry worth acknowledging, which is that the risks are hardest to reverse in the youngest children. End of answer. One hundred and thirty words. What to notice. It really depends is not an evasion here, because a specific condition follows immediately. That is the difference between using the conditional strategy and hiding behind it. Agreement watch. The evidence is, singular. Most school policies focus, plural. The risks are, plural. Model nine. Should governments fund the arts, or should the arts fund themselves. Archetype, evaluation. Here is the answer. I would argue for public funding, though not for the reason usually given. The standard argument is that art has intrinsic value, which is true but not persuasive to anyone who does not already believe it. The stronger argument is a market failure one. Work that is expensive to produce and has no immediate audience will never be funded commercially, and a great deal of what later turns out to matter starts in exactly that category. In a lot of countries, the institutions that now generate substantial tourism revenue were subsidised for decades before they generated anything at all. Having said that, public funding tends to concentrate in capital cities and in established forms, so it can quite easily end up subsidising the audience that needed the subsidy least. End of answer. One hundred and twenty nine words. What to notice. The answer explicitly rejects the obvious argument before making a better one, which is a strong analytical move and takes only one sentence. Agreement watch. Work that is expensive will never be funded, singular. The institutions were subsidised, plural. Public funding tends, singular. Model ten. Will celebrity culture define this century. Archetype, abstract and hypothetical. This is a band nine probe, and the reframe is the answer. Here is the answer. That depends what we mean by define. If we mean that historians will look back and identify it as the organising feature of the period, then almost certainly not, because things like climate and demographic change operate on a scale that celebrity does not. But if we mean something narrower, that the mechanisms of celebrity, visibility, personal branding, and audience capture, have spread into areas that used to be separate from them, then I think that is already true. Politics is the obvious case, and increasingly so is academic and scientific communication. So my sense is that celebrity will not define the century, but the machinery behind it will turn out to have been more consequential than the celebrities themselves. End of answer. One hundred and twenty words. What to notice. The definitional reframe occupies the first sentence and reorganises the entire question. Two readings are then answered separately, which converts an unanswerable question into two answerable ones. This is the highest value move in the file for the final minutes of the test. Agreement watch. Celebrity does not, singular. The mechanisms have spread, plural. The machinery will turn out, singular. THE TWO FAILURE DEMONSTRATIONS. Clipping. Same question as model two. Because there are more jobs in the cities. Rural areas do not have many opportunities, so young people leave to find work. It is mainly an economic reason. Twenty nine words, about twelve seconds. Nothing in it is wrong, and that is the problem. There is an assertion and a restatement of the assertion, no mechanism, no example, and no limit. The examiner learns that you can answer a question and nothing else, and moves on, so you lose a turn as well as the marks. The repair is move two. Ask yourself why once more. The jobs are not there. Why not. Because those industries mechanised. That second why is the entire difference between band six and band seven, and it costs one sentence. Drifting. Same question again. The main driver is employment, because rural economies depend on agriculture and agriculture has been mechanising for a long time. Although it also depends on the country, because in some places it is about education instead, and in Mexico there is the additional factor of migration to the United States, which is a slightly different phenomenon but connected, and historically that has shaped which regions lost population first. I think it started in the nineteen forties with the bracero programme, and those routes tended to persist across generations, so families in certain states have been going to the same places for eighty years, which affects the internal migration too, and of course the cities themselves have changed a lot in that time, so what people are moving toward is not what it was... One hundred and thirty seven words and still going, with no concession in sight. The answer was complete after the first sentence and a half. Everything after although is a second topic, then a third, and the tense has moved into the past and stayed there. The repair is the exit. Once you have given a mechanism and an example, the next thing out of your mouth is having said that, followed by one limit, followed by silence. If you notice yourself opening a new subordinate clause with which, and, or because for the third time, land it immediately. Then again, that varies a lot by country. Stop there. CLOSING CHECKLIST. Four to five sentences. One hundred to one hundred and thirty words. Forty five to fifty five seconds. Then hand the turn back. Assertion, mechanism, evidence, concession. The concession is your exit, not an extra. Give your opinion. Withhold your biography. Personal view, impersonal evidence. When the topic is unfamiliar, reach for a lens, not for an opinion. Three perspectives, institutions against individuals, or your own professional knowledge with the pronoun removed. Hedge every prediction. Reframe every abstract question by defining the term first. One or two cleft sentences per Part Three. It is the cost of housing that drives it. Abstract singular subjects take the s. Access determines. Migration drives. Automation replaces. Technology changes. If you slip, repair once and move on. Never repair twice, never apologise twice. Hard questions late in the test mean you are doing well. Answer them like a compliment. End of guide.