3

I'm in collaboration with another researcher in the field of linguistics, she wants to try machine learning algorithms/analysis on her dataset. The dataset is not published before, and the contribution of this work is to use ML to better analyze linguistic theories and to support/challenge them with this new empirical evidence (ML).

So, my role here was to try appropriate ML tools based on the goal she described, and also I designed a new ML algorithm to serve the purpose. In addition, I perform the experiments, numerical evaluations, preparing the figures.

But since the main concept and the aim is fundamentally related to the linguistic field, she is in charge of writing down the main structure of the paper (Abstract, Introduction, the purpose, contribution). But, definitely, I'd write the ML parts and where I need to provide quantitative conclusions and also explanations regarding the new algorithm I designed and ML approaches that were used, and how to interpret the results.

Well, we are aiming for a conference in the area of computational linguistics, and it is not clear for me who should be the first author of this contribution. Also, I do not know if it is possible to claim equal authorship for such cases, and whether it'd be considered as the 1st authorship in my CV?

BTW, she is a post-doc, and I'm an almost finished Ph.D.! ;-)

Update: Since the value of this publication (if being accepted) is important to both of us regarding our future proposals and job-search, I personally prefer an equal authorship option, if possible! But I'm not sure if possible to show that in the CV!

  • 1
    My gut from reading just this description is she would be first author and you would be second. – Azor Ahai Jan 11 at 22:06
  • @AzorAhai Funny, I was thinking the exact opposite :) – Bryan Krause Jan 11 at 22:28
  • @BryanKrause It's definitely not cut and dry – Azor Ahai Jan 11 at 22:44
  • Just flip a coin already. – JeffE Jan 12 at 11:30
5

Here is a thought. (Perhaps it is naive. I haven't seen your paper, and I know nothing about the publication culture in computational linguistics.)

Could you perhaps write two papers, with you appearing as first author as one and your collaborator on the other? This would settle the issue symmetrically. Whether it is a good idea depends more on how "bi-disciplinary" your work is. By which I mean: is your work coming from two different camps, linguistics and machine learning? If not, splitting is a bad idea. If so, if you write one paper "splitting it down the middle" is there a ready-made audience who will appreciate the synthesis of both camps into a single paper? If so, splitting is a bad idea. But if not, you might find that your work reaches more people if you appeal separately to the linguistics and machine learning communities.

Again, just a thought.

2

Authorship order is very field-dependent.

From the perspective of biomedical sciences, I feel like your collaborator is acting mostly as a "senior author": she has identified a target and approach to research, and you are doing the implementation.

Since you are doing the bulk of the analysis and producing the results, in my field I would put you as a first author and your collaborator as the final author.

However, in my field, senior (last) authorship is given value and credit; their contribution is not recognized as more or less than the first author, just "different", and for people at different career stages the importance is different: productive professors, for example, may not have any first authored papers over long stretches of time besides an occasional review.

In your field this may not be the case, in which case it is much more difficult for you to determine authorship order. One factor to consider is how much your contribution to the paper really is. From reading your question, it seems like the bulk of the work is yours; however, since you are using your collaborators data that can quickly get complicated. If this is a data set that has already been used and this is just a new approach to old data, that is much different than if these are brand new data collected for this purpose which therefore have a lot of effort into the collection and your analyses may be a more minor addition to the ordeal.

  • I don't think senior authorship is very common in linguistics, at least not this kind of linguistics. – Azor Ahai Jan 11 at 22:45
2

Like others I can only offer a perspective from a different field (biological sciences). However in this field marking two authors as having equal contribution is common. This is usually done using an astrisk in the authorship list (*). So it would look so it might look like:

Student, PHD*, Researcher, PD*, Bigwig, A^.
* these authors contributed equally.
^ To whom correspondence should be addressed.

However, it is also common in the field for people to argue over who is the first "first" author, which suggests people don't entirely believe the equal really means equal. You would make it the same in your CV and hope the person reading the CV respected the * (which some do and some don't).

Of course, this is for Journal articles. For conference presentations (which are much less of a big deal in biology than in some other fields, and don't really count as a "publication", the first author would always be the person actaully standing up and giving the paper. But then its routine for the same paper to be given at multiple conferences and the order of authors might change between them.

1

You could investigate to see if there is a standard order in the linguistics field as there are in some. But otherwise you might just list authors alphabetically, which is common in CS, for example. You could include a brief note somewhere in the article on contributions, especially if they were different in kind, not just "amount".

I worked (retired now) in CS and never gave any thought to "first authorship", valuing collaboration over nearly all else.

  • 3
    Regarding your last sentence: right, me neither....but we work in fields in which we have the luxury of not having to. – Pete L. Clark Jan 11 at 22:09

Your Answer

By clicking "Post Your Answer", you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.