Editing Openai/69668f73-ea68-800d-ac56-44a64b19dae4 (section)

=== Assistant: This is a really good way to frame it — and you’re right that the only live uncertainty in your model is whether Sotoma… ===
This is a really good way to frame it — and you’re right that the only live uncertainty in your model is whether Sotomayor/Kagan/Jackson behave as a bloc or splinter. Everything else you’ve bracketed pretty sensibly.

Let me answer your actual question directly, and then stress-test your frame a bit.

==== ### ====

Yes — much more often than not — but not mechanically, and not on procedural/institutional questions.

They are:
* A tight ideological cluster on civil rights, criminal procedure, administrative law, voting, and access-to-justice issues.
* Less of a bloc on technical, procedural, commercial, and institutional design questions — especially when: - The case is fact-specific - The stakes are structural rather than normative - The question is “what rule should courts adopt” rather than “who should win”

Hikma v. Amarin is much closer to the second category.

==== They are highly cohesive when: ====

|     |     |
| --- | --- |
|     |     |
|     |     |
|     |     |
|     |     |
|     |     |
|     |     |

In those areas, they are functionally a voting unit.

==== They splinter much more when the case is: ====
* Commercial / economic
* Highly technical (IP, securities, tax, bankruptcy)
* Procedural and rule-design oriented
* Focused on incentives rather than rights

In those, you often see:
* Kagan peel off to join conservatives
* Or Sotomayor dissent alone
* Or Jackson take a distinct position

And critically: Twombly/Iqbal / pleading / inference standards fall squarely in this splinter-prone zone.

==== Here’s the subtlety that matters most for your model: ====

All three are uncomfortable with pleading rules that bar plaintiffs from access to facts that are asymmetrically held by defendants.

That gives them a shared instinct against tightening pleading standards too much.

But:
* Kagan is more concerned about floodgates and administrability.
* Sotomayor is more concerned about fairness and asymmetry.
* Jackson is likely very sensitive to power asymmetry and informational imbalance.

So their instincts point the same way — but their thresholds differ.

That means:
* They may align in result (deny cert),
* But not necessarily in rationale or intensity.

And that matters for whether they vote together at the cert stage.

==== Another key point: Justices are less ideological and less bloc-driven at the cert stage than at the merits stage. ====

At cert, they are thinking:
* Is this a good vehicle?
* Is there a clean split?
* Is the record right?
* Will this produce a good rule?
* Is this the right time?

Those questions scramble ideology.

So even if they’d all ultimately be skeptical of Hikma’s position on the merits, one of them (especially Kagan) could still say:

: 

And vote deny cert, even if she’d agree with Hikma in theory.

==== ### ====

: 

That’s plausible.

But the more realistic picture is:
* They are not a mechanical bloc in this type of case.
* They are likely to diverge on vehicle quality and institutional value.
* Kagan is especially likely to defect from a “bloc yes” if she thinks this case is a poor platform.

So the default expectation should be:

: 

Which pushes your base case toward cert denied.

==== You’re assuming: ====
* Kavanaugh = yes (reasonable)
* Other conservatives = no (plausible)
* Liberals decide the outcome

But it’s also plausible that:
* Roberts or Barrett could be persuadable on institutional grounds,
* Or that Kavanaugh’s “yes” is not firm if the SG brief didn’t add anything new (as you’ve pointed out).

So the coalition math is not quite as clean as 1 vs 5 vs 3.

==== Do the three liberals tend to move as a bloc? ====
→ Yes, in rights and ideology cases. No, in technical, procedural, or vehicle-dependent cases.

Is Hikma v. Amarin in the bloc-prone category?
→ No. It’s in the splinter-prone category.

So your read is right to focus on them — but the conclusion should be:

: 

Which means your overall handicap should still lean toward denial, unless you see an independent path to a fourth vote outside that bloc.

That’s the cleanest, least story-driven way to model it.