Twombly: Trimming Some of the Possible Worlds

newplanet.jpgOver at Prawfsblawg, Scott Dodson has a post up commenting on the latest judicial decision to attempt to make sense of the Supreme Court’s civil procedure blockbuster last term, Bell Atlantic v. Twombly. Twombly held, seemingly contrary to what courts had been saying for decades, that a complaint must not only provide a short and plain statement of the claim, but must also plead “enough facts to … nudge[ ] their claims across the line from conceivable to plausible.” Although this set off alarm bells among many commenters suggesting that the Supreme Court had just struck notice pleading dead, the Third Circuit, in Phillips v. County of Allegheny, held that Twombly says only that Rule 8(a)(2) means what it says: a plaintiff must provide not only a short and plain statement of the claim, but one “showing that the pleader is entitled to relief.” No plausible facts, no showing. In that case, Twombly would be a non-event.

Scott proposes a different interpretation:

A better reconciliation of “plausibility” and Rule 8 would have returned to the Third Circuit’s initial insight: context matters. In the unique circumstances of antitrust, inferential allegations of conspiracy must contain facts showing the conspiracy was plausible in order to show entitlement to relief. By contrast, in a garden-variety negligence case, for example, plausibility is not required; rather, even a bare averment of “negligently drove” may need no factual elaboration on the question of negligence in order to show entitlement to relief.

These interpretations are both reasonable. But here’s mine, which I suggested to my Civ Pro class when we covered this case a few weeks ago: Twombly rules out low-probability inferences in complaints.


Philosophers use a neat terminology to discuss counterfactuals and probabilities: they refer to each possible course of events as occurring in a “possible world.” The probability of the event actually occurring (or having occurred) is represented by the “distance” of that possible world from our own. For example, a nearby possible world has me eating a peanut butter and jelly sandwich tomorrow. (It might even be the actual world.) In a much more distant possible world, I win a million dollars tomorrow. In an even more distant possible world, Earth is invaded by aliens tomorrow.

Conley v. Gibson‘s “no set of facts” language — that “a complaint should not be dismissed for failure to state a claim unless it appears beyond doubt that the plaintiff can prove no set of facts in support of his claim which would entitle him to relief — seems to say that, as long as there is even one possible world, consistent with the allegations in the complaint, in which the plaintiff wins, a 12(b)(6) motion must be denied. Twombly, by contrast, seems to hold that a motion to dismiss can be granted where the plaintiff only wins in very distant possible worlds — i.e., under circumstances that have a low probability of being actual.

I think that’s the only way in which the Twombly complaint could be dismissed, but Form 11 (née 9) and Swierkiewicz remain on the books. In Twombly, the plaintiffs alleged what were, to them, suspicious circumstances: parallel conduct by all of the Baby Bells in denying competing local phone service and internet providers wholesale access to their networks, and in not competing with each other. Therefore, given the parallel conduct, the plaintiffs alleged

upon information and belief that [the Baby Bells] have entered into a contract, combination, or conspiracy to prevent competitive entry in their respective … markets and have agreed not to compete with one another ….

How is that too large a leap from the facts pled, but Form 11 — “on [date], at [place], the defendant negligently drove a motor vehicle against the plaintiff” — isn’t? Or the complaint in Swierkiewicz, which alleged that the plaintiff was a 53-year-old native of Hungary, that he was phased out of his job by some French nationals, and then asserted that he had been fired due to age discrimination and national origin discrimination? I think the Supreme Court is drawing an implicit line between inferences that are relatively probable (although not necessarily 50% probable), and those whose probabilities are below some undeclared threshold. Or, to put it back in possible worlds terminology, there’s a distance limit to how far an inferential leap can travel in a complaint.

The Twombly majority evidently believed the “suspicious circumstances” — the parallel conduct — alleged by the plaintiffs in that case only gave rise to an extremely slim chance of an actual conspiracy; somewhere on the order of my winning a million dollars tomorrow. (Imagine this inference: I had a donut today; therefore, on information and belief, I will win a million dollars tomorrow. Sure, the second fact is consistent with the first. But the donut-eating worlds are hardly anywhere close, as a group, to the million-dollar-winning worlds.) Indeed, Justice Souter in footnote 4 takes the trouble to point to discussions of hypothetical parallel conduct that does make a conspiracy relatively more probable: for example, “complex and historically unprecedented changes in pricing structure made at the very same time by multiple competitors, and made for no other discernible reason.”

Contrast the Supreme Court’s assessment of the Twombly complaint to Form 11. Car accidents happen a lot, and often someone is negligent in them. The mere fact of a car accident, which is what Form 11 pleads, is relatively close to possible worlds in which the driver of the car was negligent. Similarly, the Court has to be saying the same thing about Swierkiewicz: when a middle-aged foreigner is fired by people from a different foreign country attempting to “energize” the company, it’s sometimes due to national origin discrimination and age discrimination. Close enough.

Don’t get me wrong: I’m not saying the Court assessed the probabilities correctly in Twombly. Who says that the world in which the Baby Bells have entered into an antitrust conspiracy is incredibly distant from the world in which 7 Baby Bells all exclude resellers from their networks, and all fail to compete in each others’ markets? Nevertheless, I think the Supreme Court thinks courts can make such intuitive judgements reliably, and the purpose of Twombly is to give district courts the discretion to dismiss when the only “set of facts” under which the plaintiff wins is possible, but highly improbable.

You may also like...

10 Responses

  1. John says:

    Very interesting post. One interesting practice note – as I have noted in my practice and in the practice of others, Twombley gives defendants new incentive to attach documents to a motion to dismiss as “evidence” that a claim is not plausible.

    Let me explain: under the Fed. R. Civ. P., evidence cannot be submitted with a motion to dismiss, lest the defendant risk converting his motion into one for summary judgment. In practice, there are all sorts of ways around this bar; the most common is that documents which are considered “integral” to the complaint may be attached and considered by the court. Common examples are contracts, bills of sale, SEC filings, etc.

    Pre-Twombley, there was not much incentive to attach lots of exhibits to a motion to dismiss; even with some evidence supporting the motion, if there was any possibility that the plaintiff’s claim was possible, dismissal would be denied. Post-Twombley, litigants have incentive to attach evidence to the complaint in an attempt to impress upon the judge the implausibility of the complaint. I did it myself just days after Twombley was decided, and I really think it may have helped color the judge’s opinion.

    Now, there is a sweet spot – a defendant can only attach enough exhibits to be helfpul, but not so many as to convert the motion into summary judgment. Pre-Twombley, there wasn’t much incentive to push the envelope. Post-Twombley, I think there is great incentive. An interesting quirk, I think.

  2. Bruce Boyden says:

    Interesting point, John, thanks.

  3. Howard Wasserman says:

    That is an interesting take. I would add that it also requires the court to disregard broad, information-and-belief allegations. Suspicious circumstances + an agreement would have stated a claim; suspicious circumstances alone do not (for the reasons you suggest). So it depends a bit on eliminating that manner of pleading facts, at least in some cases.

  4. Mike O'Shea says:

    This is quite thoughtful, Bruce — and helpful. I just started teaching Twombly this week for the first time.

    Before Twombly, the last time SCOTUS swerved from the relaxed, Swierkiewicz view of notice pleading was Dura Pharmaceuticals v. Broudo, 544 U.S. 336 (2005), a securities fraud case. Dura held unanimously that the plaintiff class failed to comply with Rule 8’s pleading requirements when the only evidence of “loss causation” pleaded by plaintiffs was an allegedly inflated purchase price for defendant’s shares.

    It’s true that the PSLRA was looming in the shadows in Dura, but the text of the statute did not impose a heightened pleading requirement for loss causation, and the Court expressly assumed that regular old Rule 8 pleading standards applied.

    So here’s my supporting gloss on Bruce’s post: It seems significant that in both of the recent cases where the Court has seemed to demand the pleading of extra facts to elucidate a causal relationship — Dura and Twombly — the intervening causal mechanism was the market economy.

    Perhaps when the defendant is a large economic actor, the universe of variables in play, the number of possible influences on the company’s business behavior (Twombly) or stock price (Gura), are so multifarious that the number of surrounding possible worlds increases very fast — hyperbolically? — once one moves any distance from the “world” pleaded in the complaint. Probabilities decrease very fast in this space compared to the possibility-space that surrounds a slip-and-fall or a car crash. Justice Breyer spoke in Dura of “the tangle of factors affecting price”; that echoes a lot of Justice Souter’s rhetoric in Twombly.

    When proffering these sorts of theoretically satisfying distinctions, it’s also fair to ask ourselves what a Crit might say about the distinction between the two types of cases. (Ahem) — “Possible worlds? [rolls eyes] Look, there’s no need to haul in David Lewis here. It’s simple. Form 11’s car crash is two small fry suing each other; the courts figure, who cares about weeding these out? Dura and Twombly are large corporate defendants with elite-firm lawyers. What’s hard to understand about it?”

    I support equal time.

  5. Mike O'Shea says:

    “Dura and Twombly are large corporate defendants”

    — that is, Dura and Bell Atlantic (now Verizon) are.

  6. Guest says:

    I agree with this post, but from a practitioner’s standpoint, I would sell it to the judge somewhat differently. Instead of pointing out the (useful) analogy of possible worlds, I would analogize the Twombly standard to the “probably cause” and “reasonable suspicion” standards in the criminal context. These are standards the judge is familiar with, and I would argue that Twombly put the showing a civil litigant must make on the same spectrum as the showing a prosecutor or officer must make. I would say that “plausibility” is a showing that is lower than “reasonable suspicion.” But if the plaintiff cannot even meet that, then his or her complaint should rightly be dismissed.

    And by the way, I never understand why people say that the facts pleaded in Form 11 are inconsistent with a “plausibility” standard. I mean, the guy hit someone. Doesn’t that make it at least plausible that he was negligent?

  7. John says:

    I would analogize the Twombly standard to the “probably cause” and “reasonable suspicion” standards in the criminal context

    I would be wary of doing that, only because then you risk misinterpretation of your analogy. As a practical matter, I always try to stay away from analogies, because they have the tendency to backfire (or just confuse the proceedings). You say that “‘plausibility’ is a showing that is lower than ‘reasonable suspicion'” – why? I think a good argument could be made that pluasibility and reasonable suspicion are essentially the same thing. Or maybe they are different. Who knows? That’s the point.

    I think the “plausibility” standard of Twombley (which, let’s face it, isn’t the model of clarity), will probably best be defined by slow osmosis, at least in non-antitrust cases. Unless we get more clarity from the Court, individual judges will become known for their willingness to embrace more fact-rich pleading. Some judges will allow pleadings to be as they were pre-Twombley, and some will not. Litigants will have to get to know the individual judge. Ironically, this is much the way it works in criminal cases – some judges have high thresholds for probable cause, some have lower thresholds. I would just stay away from the confusing analogies.

  8. Howard Wasserman says:

    Mike’s crit argument dovetails with what Judith Resnick has argued about the evolution of civil litigation under the Rules: The paradigm civil action in 1938 (and what the drafters had in mind) was a basic one-to-one diversity tort or breach-of-contract claim; the system of the Federal Rules has changed as the substantive law has changed underneath it to where the paradigm is more complicated federal antitrust or securities or environmental claims, with large corporate defendants almost always on the defense side.

    But how does the crit explain Swierkiewicz and Leatherman. Is it the market influence that is present in antitrust and securities, but less so in civil rights (especially claims v. Govt)?

  9. Alex says:

    Mike,

    I don’t think Dura is relevant to this discussion. That case wasn’t a pleading case in the sense of Twombly or Swierkiewicz. Rather the Court was considering whether, as a matter of law, a securities plaintiff establishes loss causation by alleging (and later establishing) that the price of the security on the date of the purchase was inflated because of the defendant’s misrepresentation(s). The Court held, as a matter of law, that that type of allegation did not establish loss causation. The Court did not assess whether the Dura plaintiff’s allegations were inadequate under Rule 8(a)(2), i.e., that their securities fraud claim was implausible (to Twombly’s language) because of some pleading deficiency (in terms of a lack of facts). Indeed, the Supreme Court assumed that the Dura plaintiff’s allegations were true – that they did pay inflated amounts for Dura’s stock. But in its view, as a matter of substantive securities law, that type of allegation would not suffice for establishing a 10b5 claim. And so, in that sense, the Supreme Court did exactly what one would expect a court would do on a Rule 12(b)(6) motion – it determined whether, as a matter of law, the Dura plaintiffs could state a securities fraud claim based on the allegation contained in their complaint.

    That the Supreme Court required a securties plaintiff to plead loss causation is in my view no different than requiring a plaintiff, who brings a garden-variety negligence claim, to include allegations that she was damaged by the defendant’s conduct. Dura therefore stands for the narrow propositon that a securities plaintiff who alleges she paid an inflated amont for a security, without more, does not state a securities fraud claims because she wasn’t injured, i.e. suffer any damages attributable to the fraud.

  10. Randy Picker says:

    If you are interested in more reading on Twombly, I have a forthcoming paper that talks about it, along with the other three antitrust cases from the 2006 Term. It is available at http://papers.ssrn.com/sol3/papers.cfm?abstract_id=1091498 (typo and all in the title).