(Post title inspiration.)
I'll have to wait until I read the full paper before I make a serious judgment, but on first glance Larry Ribstein's argument that the law should not prohibit Milberg Weiss's paying plaintiffs, because the prosecutors who made a plea agreement with Bershad also are paying off someone to participate in a court proceeding, seems really strange. Yes, of course prosecutors risk getting bad testimony by giving a benefit to a witness for his testimony that the witness would not receive if he didn't testify. But it's a cost-benefit analysis: until an intolerably large number of people go to prison on the word of false jailhouse rats or lying plea bargained co-conspirators, the number of people who go to prison on the word of accurate witnesses will be seen as too much of a good to give up.
Paid plaintiffs, on the other hand, provide no public good that I can discern. Reportedly, most named plaintiffs in securities class actions don't have to do much beyond signing their names to papers, and unless payments are extremely widespread, we're not short on people who will volunteer to be named plaintiffs without bribes to do so. All of the good provided by paid plaintiffs goes to the people paying them, i.e. the attorneys who are more likely to be assigned as lead counsel for the class by reason of having been first to file.
So why are the two, paid plaintiffs and plea bargained witnesses, worth comparing? Expert witnesses frequently give somewhat dubious testimony that is partly motivated by the desire to get paid by the party in whose favor they spin the facts. Do such witnesses also make the law behind barring pay to plaintiffs suspect? Some people just aren't likely to participate in a judicial proceeding unless they are compensated. Neither the jailhouse rat nor the expert witness has a self-interested reason to get involved in a civil or criminal trial of other people. Indeed, the rat has self-protective reasons not to get involved. Sure, both rat and expert could altrustically, in service of the truth -- or in order to screw someone else -- volunteer to give testimony with no compensation of any sort, but that makes a large demand on selfish human nature.
Plaintiffs, in contrast, have an inherent interest in participating. After all, they are supposed to be vindicating their own rights or interests. If a plaintiff does not feel she has been in any way harmed, or if harmed, that the harm is not another person's fault, then the lawyer who hunts her down and convinces her that she has been harmed is an ambulance chaser. The lawyer who pays her to pretend to feel a harm -- to be a plaintiff when she otherwise would not be -- potentially manufactures litigation that otherwise would not exist. (Though possibly every case Milberg Weiss has had required a race to the courthouse because of other, unpaid plaintiffs who had espied a harm on their own, which would mean the litigation would have occurred regardless of whether a plaintiff had been paid or not.)
As for referral fees to law firms being equally (il)legimate as fees paid to plaintiffs, I don't know much about how referral fees are regarded in the legal profession. The American Medical Association has declared them unethical both when the doctor refers a patient to another physician/ facility, as well when the doctor refers a patient for a clinical trial (a situation where it is more often called a "finder's fee"). But to my knowledge, the fees are not illegal.
Again, one can distinguish this on self-interest grounds: why should a firm give a referral unless they are rewarded for it? As with physicians, I think this rationale is shakier than with the jailhouse rat or expert witness because doctors and lawyers are supposed to be professionals with fiduciary duties to their patients and clients. The doctor should refer a patient to a hospital or clinical trial because it benefits the patient, not because it benefits him, and the same should be true of the lawyer referring to another firm. But these are demands of disinterest that we make on professions; they are not the same as the self-interest -- the concrete and particularized injury -- a plaintiff inherently should have as part of standing.
 Ideoblog commenter Richard Painter says in support of Ribstein's point, "Expert witnesses in civil li[t]igation get paid a fixed amount -- usually an hourly rate -- regardless of how effective their testimony is for the party that retains them (or at least they should be)." That's true, but expert witnesses are chosen in advance based on how effective their testimony will be for the party that retains them. If I am suing Merrell Dow, I will retain only an expert who says that Bendectin causes birth defects, not the ones who say it isn't a risk factor. Painter's mistake is in looking backward on the plaintiff-expert relationship.