Why a script makes it worse
The instinct is to remove all risk by writing every word. It backfires in three ways. Written Japanese delivered aloud has the wrong register, with long sentences and full forms that are hard to follow by ear. Reading flattens intonation, so the audience loses the prosodic cues that mark structure. And a script has a single point of failure: lose your place and there is nothing underneath it.
Improvised speech at half speed is genuinely easier to follow than a fluent reading, because it has natural emphasis and real pauses. Audiences are far more tolerant of slow, simple language than of a wall of correct sentences delivered without shape.
The compromise that works is to script the structure and the transitions, and improvise the content inside each section.
A structure that suits a Japanese audience
Explicit structure matters more here than in many presentation cultures. Saying how many points you have, naming each transition, and summarising at the end are not padding; they are what lets an audience follow a speaker whose delivery they have to work slightly harder to parse.
The conventional shape is simple and worth following rather than improving on: greeting and self-introduction, the purpose, the number of points, the points themselves, a summary, and an invitation to ask questions.
| Japanese | Romaji | Meaning |
|---|---|---|
| 本日はお時間をいただき、ありがとうございます。 | Honjitsu wa ojikan o itadaki, arigatō gozaimasu. | Thank you for your time today. |
| 本日は三点についてお話しします。 | Honjitsu wa santen ni tsuite ohanashi shimasu. | Today I will speak about three points. |
| まず、一点目です。 | Mazu, itten-me desu. | First, the first point. |
| 次に、二点目に移ります。 | Tsugi ni, niten-me ni utsurimasu. | Next, I will move to the second point. |
| 最後に、まとめさせていただきます。 | Saigo ni, matomesasete itadakimasu. | Finally, let me summarise. |
| ご質問があれば、お願いいたします。 | Goshitsumon ga areba, onegai itashimasu. | If you have questions, please go ahead. |
Announcing the number of points at the start is the highest-value sentence in the whole presentation. It tells the audience how to allocate attention and it commits you to a structure you cannot drift out of.
Signposting so nobody gets lost
Between sections, say where you are. Learners tend to move between points silently, which is fine in a language the audience processes effortlessly and confusing in one delivered with an accent and an unfamiliar rhythm.
Signposting is also a gift to yourself. Each transition phrase is a rehearsed, automatic sentence, which gives you two or three seconds to remember what the next section contains.
| Moment | Say | Effect |
|---|---|---|
| Starting a point | まず / 次に | Marks the boundary |
| Giving evidence | 例えば | Signals an example |
| Contrasting | 一方で | Prepares a shift |
| Concluding a point | つまり | Marks the summary |
| Returning | 先ほどの話ですが | Reconnects a thread |
| Ending | 以上です | Signals you have finished |
The last one matters more than it looks. Without it, audiences are unsure whether you have finished, and the silence after a presentation becomes awkward.
Rehearsing to the structure
Rehearsal should make the structure automatic and leave the wording free. If you rehearse the words, you will notice every deviation on the day and it will unsettle you; if you rehearse the shape, every delivery is a valid one.
Rehearse aloud, standing, at the pace you will actually use. Silent rehearsal takes about half the real time and gives a badly wrong impression of how long the presentation runs.
- Rehearse from a five-line outline, never from the full text.
- Say it aloud at least four times, varying the wording deliberately.
- Time the third run; slides always take longer aloud than in your head.
- Rehearse the opening and the closing to the word; those two can be fixed.
- Practise saying your final sentence and then stopping, which is surprisingly hard.
Surviving the questions
The question session is where preparation usually stops and where the presentation is actually judged. It is also the only genuinely unpredictable part, which makes it the part most worth practising.
Two things carry you through: a set of phrases for buying time and asking for clarification, and having decided in advance that not knowing an answer is acceptable. Both are more useful than trying to anticipate every question.
| Japanese | Romaji | Meaning |
|---|---|---|
| ご質問ありがとうございます。 | Goshitsumon arigatō gozaimasu. | Thank you for your question. |
| すみません、もう一度お願いできますか。 | Sumimasen, mō ichido onegai dekimasu ka. | Sorry, could you repeat that? |
| 確認させてください。……ということでしょうか。 | Kakunin sasete kudasai. ...to iu koto deshō ka. | Let me confirm: do you mean...? |
| その点は今、資料がないので、後ほどお答えします。 | Sono ten wa ima, shiryō ga nai node, nochihodo okotae shimasu. | I do not have the data here; I will answer later. |
| 勉強不足で申し訳ありません。 | Benkyō busoku de mōshiwake arimasen. | I am sorry, I am not well informed on that. |
| ほかにご質問はありますか。 | Hoka ni goshitsumon wa arimasu ka. | Are there any other questions? |
Practising the unpredictable half
You can rehearse a presentation alone; you cannot rehearse a question session alone, because the whole difficulty is that the questions were not yours. That half usually goes unpractised, which is why confident presenters fall apart at the end.
The fix is to have something ask you questions about your material out loud. Explaining your topic to an AI Sensei in Unihongo's immersive 3D classroom and then taking its questions reproduces the actual condition: spoken, unanticipated, and requiring an answer now. A chat window cannot, because unlimited composing time removes the entire difficulty.
The session report afterwards tells you which answers collapsed, and those are the two or three you should prepare properly before the real thing.
Run the question practice twice. The second time, the buying-time phrases arrive automatically, which is the whole point of them.