What not to do in talks

Fearing Tavern‘s comment on the previous post links to his exhaustive list of “things not to do in a conference talk”:

use cliche’ expressions
include unecessary equations
read paragraphs from slides
animate unecessarily (vis-a-vis powerpoint)
get caught lying
read formula’s out loud
ignore your time requirement
cry (or sound as though you will)
assume nontrivial background knowledge
mispronounce people’s names
use a separate laptop from previous speaker (often causes technical difficulties)
forget to conclude
change notation
use .AVI movies (or anything else specifically Windows)
pander to famous audience members
be afraid to ask for clarification on audience member questions

I think I’ve done seven of these, though I won’t reveal which.  What about you?

Tagged ,

13 thoughts on “What not to do in talks

  1. Hmm, I’ve done seven, too. That includes one that I actively try to avoid but occasionally mess up (“use cliched expressions”), as well as some that I think are only bad in certain circumstances (using your own laptop is ok if there’s 30 minutes between talks and you start setting yours up as soon the previous talk finishes.)

  2. Greg Martin says:

    One of the given “things not to do in a conference talk” is “lie”. I disagree with this!

    The purpose of a conference talk is to eduate the audience; specifically, since it is impossible to fully educate the audience on every detail of a paper (say), we should try to educate them on the big picture and the important landmarks of the supporting technicalities. And if lying about something actually improves their understanding, then why not use the opportunity?

    Often we can easily mention that, or even how, we’re lying when we lie, without counteracting lying’s benefits. Often, but not always. I’m not recommending actively searching for opportunities to lie to our colleagues … but I am recommending keeping in mind that the primary goal of giving a conference talk (or a calculus lecture, etc.) is to help other people understand something. All other choices, I submit, should be made in the context of that goal.

    This philosophy automatically covers six of the “things not do do”, tohugh I won’t reveal which.

  3. When I was an undergraduate, my advisor told me that when giving a math talk one should “emphasize the obvious and obscure the essential”, which I think is very true and something I’ve tried to take to heart over the years. So I agree with Greg that a little “lying” is a good thing when done right…

  4. Christine says:

    I disagree with the lying from the other commentators. I think we should instead be looking for a way to present a big picture that doesn’t involve any lying. As a student, lies are particularly bothersome because unlearning something is difficult and years from now I may remember that one particular statement as truth when in fact it’s false.

  5. Graham says:

    I made this in a fit of frustration after a particularly bad AMS meeting, and inspired by a blog post I can’t now recall that was talking about bad talks: Commutative Algebra Bad Talk Bingo. Reload for a new random card. Not everything on there is a sin, per se — they’re not even all negative — but they satisfy my griping quotient.

    (“Goes over time” appears several times in the database, because I hate it so much.)

  6. I have done at least four of them myself. One thing I have not done though is:

    * insert apostrophes in non-possessive plurals

    That’s pretty egregious, no?

    And I agree with Christine. You should not lie in a talk without warning: that is needlessly confusing to the speakers (few, but significant) who are actually trying to follow your argument logically. As long as you give any kind of warning — e.g. stating that a result is
    true under “additional [unspecified] technical hypotheses” — then this could be a great expository choice. (Of course, you should be willing to provide details or appropriate references if you get asked about the technical stuff you are trying to slide under the rug.)

  7. JSE says:

    It’s a direct quote, Pete: hassle FT about it, not me!

    I’m with Pete on lying: I think it’s OK if you’re honest about doing it. Going overtime, though, that’s never OK.

  8. Paul says:

    Leaping to FT’s defense: the apostrophe is clearly an attempt at “Cliché.”

    If you lie, but warn the audience you’re lying, and they protest anyway, does it count as being caught? Because I’m all about lying.

  9. Jason Starr says:

    I have been organizing a seminar aimed at grad students about writing, preparing job app materials, etc. What I have realized is that it is easy to make lists of “things to avoid”, but that is not very helpful for students who haven’t done this sort of thing before (I am sure it is helps the author of the list to blow off steam). Why not make a list of “things to strive for”? Here are a few obvious ones: tell a joke, answer a good question from the audience, give a simple example before giving a difficult one, explain some history, explain why some famous problem is hard, break the chalk so it doesn’t squeak, repeat a soft-spoken audience member’s question / remark so everybody hears it, . . .

  10. Christine says:

    @JSE, is being honest about lying really still considered lying? Especially when the false statement and the admission of the lie happen in the same breath? If so, then you have a stricter moral code than I do.

  11. Andy P. says:

    When I first started giving talks, I was terrified about missing a single technicality. Eventually I calmed down and got the courage to tell small lies (though always with a warning to the audience), and my talks became a lot easier to follow.

    For example, once you understand orbifolds/stacks, you can (almost) treat them like ordinary spaces. In a talk not aimed at absolute experts, I think it is fine to mumble an apology for doing so at the beginning of the talk and then go ahead and pretend that there are no issues.

  12. Willie Wong says:

    @Christine: I think it has less to do about moral code than definition. The relevant entries from Princeton WordNet gives “tell an untruth; pretend with intent to deceive” and “a statement that deviates from or perverts the truth”. In most cases (in analysis anyways, I am not sure about other fields), the lie/recant often goes something like, “… and this inequality holds. Well, to be technically precise, we need to insert an error term due to [so-and-so]. But let us move on and pretend that error term is not there…” So if we take lying to be “making a statement that deviates from the truth”, then the statements in the presentation is obviously false, and hence a lie. But if we take lying to have “the intent to deceive”, then the above situation is not a lie, since it clearly indicates the picture as incomplete.

    Personally I’ve found that in those <30 minute talks, it is completely impossible to finish a report about PDEs that does not involve any lies or omissions. And in a 1 hour talk, it is completely impossible to maintain the attention of the audience if you give every technical detail. So from my experience in my field, I'd say that lying is okay if (a) you are honest about it and (b) you are prepared to give the gory details if someone asks afterward.

  13. [...] Ellenberg stirred up a discussion of bad practices that happens in talks. Then he decided (after a commenter) that positive reinforcement of good [...]

Leave a Reply

Fill in your details below or click an icon to log in:

WordPress.com Logo

You are commenting using your WordPress.com account. Log Out / Change )

Twitter picture

You are commenting using your Twitter account. Log Out / Change )

Facebook photo

You are commenting using your Facebook account. Log Out / Change )

Google+ photo

You are commenting using your Google+ account. Log Out / Change )

Connecting to %s


Get every new post delivered to your Inbox.

Join 505 other followers

%d bloggers like this: