Asaf Karagila
I don't have much choice...

How do you read a paper?

There are 2 comments on this post.

Some time ago I was talking to some people about how they read a paper. And I learned that I am somewhat significantly different from a lot of people. I spent some time thinking about it, and I arrived at some interesting conclusions.

So here is how I read a paper, and I'd like to ask you to think about how you read a paper, and why you read it this way.

If I wasn't looking for that specific paper (or if I haven't been directed to that paper somehow), I'm going to look at the title and authors. If at least one of them catches my eyes as interesting, I'll proceed.

Next comes the abstract, which I might scan if it seems too long or too dense. And this is where we diverge from the norm.

After the abstract comes the acknowledgments, if there are any. I'm not necessarily looking for my name there. Heck, my contribution to mathematics so far is fairly insignificant to be mentioned in the acknowledgments of most paper I'm going to open. But if you know who the authors are acknowledging, you know more about the psyche of the work. You know where it happened, and you know who were the people that were consulted while the research and writing were in progress. That tells you a lot, because you don't do mathematics in vacuum. In space no one can hear you prove.

Then I go back and read the introduction, and then I will decide whether or not I am going to read the rest, or just a part of the paper, or if I am going to scan the entire thing for interesting words in the theorems or definitions. When I do read, I won't usually get into the proofs. I will, eventually, if I feel that I need to, or if I've come looking for a particular proof or trick, but generally I will look at the theorem and try to understand why they might be true.

That's my process. I remember a lot of theorems, but very few proofs, which is not necessarily a bad thing, but not always a good thing (especially when you remember just 85% of the right formulation of the theorem!), but it all works out at the end. You can always go and read through the proof if you feel that you need to.

And it is interesting that a lot of people don't pay a lot of attention to acknowledgments. Recently I read a preprint by someone I know, and I remarked that there are no acknowledgments. They replied by saying that it's senseless to give acknowledgment to the referee now; but surely they had other acknowledgments to give (in fact, I know they have (not to me, if that matters)). And after that I noticed that it's not uncommon for a preprint not to have any acknowledgments, at least in the initial version.

I find that odd. And I'm sure that up until this point, I was the only one (modulo finitely many exceptions perhaps).

So how about you? How do you read a paper, and more importantly, why do you read it that way? (And please put acknowledgments in your preprints!)

There are 2 comments on this post.

By Evelyn
(Nov 20 2015, 08:43)

That's interesting. I hadn't really thought about it, but I do read the acknowledgments after the abstract. I don't know if it's as much to learn the psyche of the work as to put off the hard part for a few more moments. :) It does help you understand the paper, though, and get ideas of other people to read or talk with if you're interested in the topic.

By Pedro Sánchez Terraf
(Dec 01 2015, 19:52)

Browsing . Skimming. Skimming, multiple times, each one with a tiny bit more of precision than the previous one. After that, I can really read. I do not know why, that's the way my mind seems to work, at least most of the time.

Want to comment? Send me an email!