How to start a talk about a topic that cannot be explained without some background information?

Question

I am going to give a presentation of my research soon at an undergraduate conference. I have read that it is best to begin a presentation with the ‘thesis’, i.e. a brief explanation of the topic of my research.

However, I’m rather stuck as to how to explain it simply. It seems like it won’t make sense until I’ve explained the concepts behind it a little. To be more specific, the topic is a formalisation I’ve done in a proof verification system. But the students attending won’t know what a proof verification system is. Is it better to begin by explaining what mechanical proof verification is before explaining the topic?

Time parameters: 20 minutes for the talk, and 5 extra for questions.

Answer 1

As Wrzlprmft pointed out, you don't want to tell the audience things they don't understand. But sometimes that's the only audience you got and you have to make it work. Besides, this quote will keep bugging you if you give up:

If you can't explain it to a six year old, you don't understand it yourself.

― Albert Einstein

I would start with a heuristic. Compare it to something they do understand, something simple that everyone knows. For instance, when they explain software engineering, they often compare it to constructions of houses.

Some visual examples help. Yesterday I had to explain "statistical significance" to a friend of mine who has no idea about statistics. I flipped a coin, got heads and said that his coin had a 100% chance of rolling heads. He replied "Bull****!". I flipped it two more times and got heads and tails. I told him now it looked like it had a 67% chance of heads. He looked at me suspiciously. Then I said if I flipped it about 100 times, it would come to about 50% chance of heads, which is close to the truth. He got it after that presentation, even though I never even gave him a definition of "statistical significance". Inserting an easy-to-understand picture into the PowerPoint also does the trick.

Replace the technical lingo with common words whenever possible.

Also, trying to cover multiple difficult topics rarely works in any presentation. I usually pick the most important point and focus on it. If it is too broad, go a level higher. Zoom out and lose the details. Or pick a subtopic and dedicate your presentation to it. That's the trade-off you got to make.

But the most important thing is to do it. Practice, make mistakes, correct them, and improve. By the time someone wins a Nobel Prize, they are always able to explain their work to general public in a few minutes.

Answer 2

I have read that it is best to begin a presentation with the 'thesis': i.e. a brief explanation of the topic of my research.

I would not consider this piece of advice generally applicable. For example, in my field many people are applying the field’s techniques to some application from another field. As most of the audience is not familiar with this application (at least beyond what is generally known), there is no point in telling it anything about your detailed approach, before making it familiar with the application. And this is what good talks do: They first introduce the application and then they explain how they approached some specific problem.

More generally, one fundamental rule for a good talk (and in fact, any communication) is:

Do not talk about something that you do not expect your audience to understand.

Answer 3

I think the best advice would be to try to explain it to other people first, get feedback and then rework it. I'm not sure myself what that topic includes. I prefer having a map first thing and then letting the discussion fill in the left and right details but if I don't know what a map is I would assume I would need a lesson on maps. But that analogy may be completely useless for your topic.

Answer 4

For a 20 minute presentation you must be concise. For organizational purposes, you can introduce yourself and thesis/focus, then simply state an outline for your presentation: "first I will explain x, then talk about y and follow up with z." Then the audience will expect the background and you don't spend a lot of time on it. Give them the bare minimum background they need.

Answer 5

I have read that it is best to begin a presentation with the ‘thesis’, i.e. a brief explanation of the topic of my research.

Let's generalize this. Start with some hook or hint of what's to come. Something to motivate them to listen to the background explanation you have to provide.

This approach is somewhat analogous to the lead paragraph in a news story.

Additional suggestion: I like talks that provide a brief outline of the talk, at the beginning, and then brief reminders of the outline when moving from one section to another. This could be especially appreciated by your audience if your talk structure is a bit out of the ordinary.