Some time ago I muttered something along the lines of “there are lies, damn lies and authorship lists”. It is really one of the more contentious sides of research and an endless source of strain to some people (and to everyone on occasion) as with more than one author there has to be some kind of order of authors on the paper (Jones and Smith, or Jones, Smith and Smythe), and at least one person is probably going to be less than 100% happy, and it is almost certainly not going to really represent quite who did what on the manuscript.
Moreover, who even gets on the paper and for what reason can vary from field to field (lab techs seems to regularly be included as authors on molecular papers, whereas fossil preparators are rarely included on palaeo ones) and certainly within fields too, and there are some arcane rules hidden away in those lists. Indeed I’m sure there are others that I’m unaware of even within biology/ palaeontology and certainly beyond.
Things also vary according to the structure of a paper too. If you already have half a dozen or more authors on a small project and someone comes in with an important point or can add an extra analysis or whatever, then it’s unlikely to upset the applecart to add them in as an author, and it probably represents a reasonable approximation of the contributions with so many involved. If however you’re refining a 50+ page monster you constructed yourself then the same amount of help will likely only get you into the acknowledgements: it’s not really fair to split the credit as Smith & Jones as opposed to Smith alone for that (proportionally) little. Some people probably will of course, when others might not go that far in the first example, but such is life.
It’s also the reason why younger researchers (I think) get more concerned about this, when you only have a handful of papers to your name, missing out on one, or dropping down the authorship list a few places can make a big difference to your overall record, whereas seasoned researchers will have this effect balanced out over a number of papers and are generally more generous with credit (and can afford to be). Certainly it can be frustrating when some people seem to endlessly get added to work by their colleagues or supervisors or there are 20 authors on a 3 page paper, it looks like someone(s) isn’t pulling their weight and are getting undue credit. To be honest I think this is a pretty open, if dirty, secret but it survives because really there’s no demand by journals for authors to justify exactly what they did, and in any case people can (and I suspect at least some would) simply lie.
Obviously the order generally reflects the level of work committed to a project. The first author did the most, the last author the least, but it’s rarely that simple. In palaeo at least and certainly most fields of biology I know, it’s common for the last author in a string to have been in some kind of supervisory role. Senior researchers at the head of labs are often in this position though they might have made a considerable contribution to the paper. In know that in China however, this is generally not the case and authorship lists there tend to more reflect levels of contribution, though confusingly this can also integrate levels of seniority (so Professor X might be listed above researcher Y even though the latter did a bit more).
Certainly different people can be on papers for very different reasons and it’s certainly not uncommon for people to be in papers for which they have written not a word. That of course doesn’t mean they didn’t contribute – they could have collected important data, run analyses, found or prepared the original material, been involved in discussions or generated the ideas the work is founded on, generated funding to support the work or a host of other technical and academic contributions without necessarily being involved in the actual writing. In some cases the final paper contains none of the persons’ contributions because it was cut out, but they remain on their in reference to their efforts and time (and quite probably money) devoted to the project.
All of this adds up to a confusing picture and while some papers can be read and each author’s contributions laid bare, others are rather more complex. If nothing else, there is a huge difference between a genuine collaboration between ‘Smith and Jones’ where each did 50% of the work (and you can even see joint first authored papers these days) and a ‘Smith and Jones’ where the latter was the former’s supervisor who helped the project into being, and deserves his credit, but did perhaps 10% or less of the paper. Hence the inevitable issues when you get to 5 or even 10 authors and the conflicts that can arise and of course, the confusion that can arise when reading a paper.