Avi Steiner emailed me with a set of questions that are too good not to turn into a blog post:
Being a math/science major at a small liberal arts college, I unfortunately never get the “full” experience of a math/science talk. Since I do plan on eventually attending grad school, I thought it might be beneficial to get an idea as to what the aforementioned “full” experience is. Therefore, I present to you and your readers the following questions:
1. At what point in a group’s/individual’s research will they choose to give a talk?
2. What sort of questions are asked?
3. Are there any recurring audience-member “personalities” (e.g. Haughty Famous Guy Who Thinks You’re Going At The Problem Wrong)
4. Does anybody actually ever sleep through talks?
Depending on where you are, you may still be able to get the full seminar experience at a small college– the talks won’t be quite as technical, but all of the associated quirks can be present. The key to the seminar experience is the presence of academic scientists, and there are plenty of us at small colleges as well.
Taking these in order, there’s no hard and fast rule for when in a graduate career you might find yourself giving a seminar talk (I assume the question is about the 45+ minute sort of talk given to a department somewhere, not a contributed conference talk). Some people never give one until they’re done with their Ph.D. and looking for jobs; others will graduate with extensive experience. It all depends on what you want as a student, and how your research goes.
This also often depends on where you are and who you know. I gave two seminar talks before I got my Ph.D.: one was a talk at my alma mater, which I arranged myself because I was planning to be in town anyway for a rugby alumni game; the other was at UConn, where the group who originally came up with the time-resolved collisions idea and suggested it to me was located. In both cases, I was talking to people who already knew me, which made the experience a little more comfortable.
When I did my time as the colloquium organizer here at Union, it occurred to me that undergrad students don’t necessarily need to have a Big Name Professor come to give a talk, so I’ve invited a bunch of grad students and post-docs here for colloquia over the years. This has provided a few people with the opportunity to give long talks earlier in their career than they otherwise might’ve.
As to the second question, the questions that get asked depend a great deal on the talk itself, as well as the details of the audience. If something you’re working on happens to run into one of the pet issues of a local faculty member (there’s a pretty good chance of this if somebody there knows you well enough to invite you), you may get a bunch of highly technical and specific questions. If the audience contains a lot of beginning students and undergraduates, you may get some really basic “please define this jargon term” questions. If your work isn’t that close to the local faculty areas of interest, you can get some surprisingly random questions.
The one fairly safe bet is that if there’s a specific technical point you’re really worried about, and spend a lot of time preparing an answer for, you won’t get asked about that thing (unless you deliberately put a really obvious hook for that question into your talk). The universe is perverse that way. Odds are very good, though, that you will get asked a difficult question about some aspect of the work that it never would’ve occurred to you to think about, let alone prepare an answer for.
As to the third question, there are some general Types that show up in most departments. Nearly every physics department has at least one Interrupter, who will break into your talk several times to ask questions as they come up (Bill Phillips is the ultimate example of this– during my graduate career, I developed a mortal terror of the phrase “Before you leave that slide…” which heralded a really difficult question). Many have an Obsessive, who can manage to ask a question about their pet interest even when it has nothing whatsoever to do with the talk (“Your discussion of Cassini images of Saturn was very nice, but what can you tell me about the Fermi surface of the semiconductor used to make the CCD?”). Lots of departments have a Polymath, the guy who asks a question at every seminar no matter what the field (I tend to be this guy, though I try not to be too annoying about it).
Lots of departments have a Crazy Person as well– this is most common at really large institutions, but Crazy People turn up even at small colleges. These are the most nerve-wracking from the speaker’s perspective, because if you’re not warned ahead of time, it can be really hard to tell a random Crazy Person from a distinguished member of the faculty. Which means you can find yourself struggling to come up with a sensible interpretation of a question that’s just insane– I saw a speaker stopped dead once after a talk on the quark structure of the proton by the question “As we all know, protons decay. How does this affect your results?” The correct answer to this is, of course, “Protons do not decay, as far as we have been able to measure. Their nonexistent decay thus does not affect my results,” but the speaker didn’t realize that the question was just crazy, and stammered for a bit trying to find a way to interpret it that would lead to a sensible, non-blow-off answer.
As for sleeping through talks, it happens all the time. One of my thesis students a few years ago never went anywhere without a digital camera, and started his senior thesis talk with a gallery of images of people sleeping at colloquia. Including one shot of me dozing off during some talk or another (providing lunch with the talk, as we do, is really deadly in this regard– I eat before the talk, then lapse into a food coma).
What’s really impressive is the occasional person who can doze off during a talk, then wake up in the question period, and ask relevant questions. There was a guy at NIST when I was there who would close his eyes during the title slide of any talk, and as far as anybody could tell, drop into a deep sleep– faint snores, REM twitches, the whole nine yards. If the talk was in an area that interested him, though, he would wake up during the question period, and ask questions that I swear indicated a complete knowledge of everything said during the talk, including questions asked by people at the seminar.
That just floored me. I would love to be able to do that, but alas, when I get past the whiplash-nod stage, I don’t recall anything.
I’m sure I’m forgetting some important aspects of the Seminar Experience. Feel free to correct my omissions in the comments. And thanks for the question, Avi.