Authorship where Authorship is Due?

In a community known for the moniker of "publish or perish", it should come as no surprise that debate about authorship and credit are rampant. Information about authorship policies was volunteered as enticement in grad school interviews. Authorship order has been debated in the hallways of my department (or at least in one doorway), and interpretations of authorship have been discussed in everything from comics to scientific studies. Articles now come with a variety of disclaimers about what contributions each person made (whose idea was it, who collected/analyzed data, who wrote), or that authorship was determined alphabetically.

I had never felt the need to involve myself in these debates: my advisor lets students first-author their own research as a matter of course, and I'm not so committed to academic life that I particularly worry about "perishing". Still, I was intrigued by the results of the scientific study mentioned above, which have recently become even more relevant for my academic life.

According to that scientific study, the first author of three would be perceived to have been involved in 37% of the initial conception, 57% of the work performed, and 33% of the supervision. The last author - by convention in psychology, the "head honcho" - gets 20% of the work performed, but 49% of the initial conception and 54% of the supervision. The middle author gets the leftovers - considered to have contributed a meaningful but relatively small portion of the work.

At the time I came across that study, I was writing up a single paper as first author of three. I found the results to be acceptably reflective of the work each of us had put into this paper, certainly far more accurate than the comic (although of course you're hearing from a biased perspective). I was only tangentially involved in the inception ("hey, we can use those tasks in my thesis work, and then if nothing else we can have a replication to talk about..."), did all the data collection, and performed all the data analyses (with a great deal of electronic hand-holding from the second author). The head honcho, far from not having read the paper, holds most of the responsibility for its current organization, and several of the analyses.

Now, I keep mentally returning to that study because there may be a dispute of authorship on a related writeup, the attempt to get my master's thesis published. From this broader perspective the second author isn't really involved. The task he programmed and trained me on was just one of four, and even the results of that task are from a different angle than the one he helped with. He doesn't know what the results for the other tasks are, may even not remember what the other tasks are, if he ever knew. I have heard enough against author list inflation to think he belongs in the acknowledgments, not in the author list.

I'm not sure how intense my advisor's position on this author list is. His inclusion only comes up at all because the author list was copy-pasted from the smaller writeup. Upon consideration, adv might agree with me. Not wanting confrontation, however, my first volley in the debate is to turn in the first draft with just two authors listed. Perhaps adv will accept this without comment. If not, I will list the purported second author's contributions, and point out their similarity to the contributions of a person who is firmly in the acknowledgments.

The real question that preys on my mind is what to do if adv doesn't accept these arguments. I want this work published, and I don't need my relationship with my advisor to be any more rocky than it already is, so I'm not sure I can outright refuse to accept a superfluous author. If I let it happen, though, I know it'll bug me from here until eternity and forever tarnish my opinion of whatever paper we get out of this. Academic honesty meets academic authority...rock, meet hard place.