This spring has been the season of talks for me. This is great, because I love giving talks! Wait, I should say that I love giving good talks; the mediocre ones not so much. There are plenty of really good reasons to give talks (see this excellent post of Adriana’s, with its many excellent links), and a lot of good advice out there about giving math talks (here are three of my favorites, from Terry Tao, my first Algebra professor Eric Moorhouse, and Gizem Karaali).
I think that math talks are important and good talks are really powerful—talks that are enlightening instead of intimidating, that aim to communicate instead of impress, that bring out the fun and wonder of math. With this in mind, I usually spend a lot of time worrying about and planning my talks before I give them. Usually this works, and I have given many talks that I was really happy with! However, even the best intentions can miss the mark, and mine have definitely missed on occasion. Then comes the agony of post-talk deconstruction. Of course this can become a black hole, a talk-related disaster of its own. Having already over-indulged in post-talk soul searching, though, I thought maybe I could save others some suffering by sharing a few of my talking disappointments with you. Alert: these disappointments were not, for me, fully preventable. But maybe some warning, and putting a name to these possibilities, can be at least a little helpful.
Sometimes it is just really hard follow the good advice. For example, the good advice is that you should give big-picture and context. Don’t be too technical. Some of my most unsatisfying post-PhD research talks were about brand new work that I didn’t have much context for, and honestly didn’t understand as well as I had understood my earlier work. I remember one talk especially where I know I gave more technical details than big-picture ideas. I remember it probably because there were multiple people sleeping in my seminar. I was embarrassed about this (though, in my defense, I’ve seen people sleeping in really good seminars), but at that point in my understanding it was just the best I could do. I understood the computations I had done but didn’t deeply get why I had done them. Preparing and giving the talk was also part of the learning process, and helped me understand the work better later on. So it was not my best talk, but I can console myself with the fact that at least I learned something from it even if some others did not. I think that it is still worth giving a less-than-perfect talk—you just have to respect your audience by really trying to give them the best talk possible. If you really care, your not-perfect talk will still be loads better than a talk from an expert who doesn’t care very much.
Another instance of not being able to follow good advice came up in the transition from graduate school to working life. People rightly say that you should practice your talk and get feedback on it. But in a new job, whom can you ask to watch your brand new and maybe very shaky talk? I practiced my first talks after grad school by myself in an empty room, but it took me a while to get used to this. I really missed my friends and their honest feedback (who knew I would miss hearing “Too much text! Boring! Give me an example!”?). Especially at first, my new talks were just not as well-organized as I might have wished. Looking back, I am really grateful for that community of people, who must have watched me give some awful practice talks.
Other disappointing talks happened when I thought hard about audience, as the advice suggests, but turned out to be wrong about who would be in the audience. For example, when I started giving undergraduate-focused talks, I didn’t realize that the audience for these talks usually includes a lot of professors. Though the organizer might claim you can choose any topic that the students might like, choosing certain mainstream topics will bore the professors in the audience nearly to death. This is not your fault as speaker, but it is very disappointing to *be* that speaker. I suffered this fate when I explained public key cryptography to one group. I could feel their silent screams as my talk went on. Now, I try to choose topics that include at least something that I think most professors in any audience won’t know very well. I still give the crypto talk, but if there are professors in the audience I start talking about lattice-based cryptography way earlier, so they have something a little different to think about.
One other issue that the good advice did not save me from: sometimes we just expect too much response. For example, after graduate school, I started to feel like my talks were not making much of an impression. Then, I realized I was used to people being extra nice to me because I was a graduate student. Of course people have higher standards and are a little more sparing with the praise after you get a PhD. I didn’t really realize how wonderfully kind and encouraging people had been until I was out of school, and noticed I really missed it! On the other hand, since I figured out what was going on, I adjusted my expectations and can even appreciate that people don’t feel the need to go easy on me anymore.
In another example of expecting too much, I realized that depending on the culture of the school, undergraduates may not be super responsive even in talks they really like. They don’t often start conversations with a speaker afterward. I had a lot of experience getting undergraduates to respond in the classroom, but didn’t realize it might be so different in talks until I started giving a lot of undergraduate focused talks. I love giving these talks and have had some great experiences connecting with undergrads at talks—I just now know not to feel bad if nobody introduces themselves!
Okay, so there are a few difficulties I’ve run into—I hope that they can serve to make someone’s talks better, or make some disappointment shorter-lived. More warnings or good ideas? Please share in the comments!
Thanks for the post! The pint about professors being in the audience for undergrad talks is a good one, thanks for pointing that out. Also, adjusting expectations for feedback from being a grad student to after is a good tip. I remember my advisor complaining that this happened to him when he became a tenured prof.
This is a nice article and I congratulate you for your honesty in describing your experiences: the good, the bad and the ugly. The advice I give my grad students is to structure their talks in 3 parts: past (what is the problem; why important; motivation); present (what did I do; why important); future (what is left to do). I ask the students to spend more time on the first part and not to get too technical in the 2nd part (a simple proof here and there is ok, but not more). Giving a good math talk is difficult and practice is important. One more thing, you should not feel guilty about giving talks advertised and targeted at undergraduates because they do not contain materials for faculty.