Hedgpeth v. Pulido – Oral Argument – October 15, 2008

Media for Hedgpeth v. Pulido

Audio Transcription for Opinion Announcement – December 02, 2008 in Hedgpeth v. Pulido

del

John G. Roberts, Jr.:

We’ll hear argument next in Case 07-544, Hedgpeth versus Pulido.

Mr. Friedlander.

Jeremy Friedlander:

Mr. Chief Justice, and may it please the Court: Respondent has agreed with us that the ninth circuit was wrong when it said that the error here and the one in Stromberg were structural defects.

The remaining question is whether Respondent can salvage the Stromberg rule by making it part of a proper harmless error test as he attempted to do.

His harmless error test is wrong for several reasons, but the Stromberg part of it was wrong for basically one reason.

He makes the Stromberg rule into a rule of harmless error, and it is not even a correct rule of error.

This Court has defined “instructional error” to require not merely a mistake in instruction, but a reasonable likelihood that the jury misapplied the law.

So when Respondent says, invoking Stromberg, that to instruct the jury on valid and invalid theories is harmless error if the jury adopted a valid theory, what Respondent is saying is nothing more than that no error occurred in the first place.

Because the jury found everything it needed to find in order to convict and, therefore, did not misapply the law.

Antonin Scalia:

Do that again, will you?

Jeremy Friedlander:

We need to have a reasonable likelihood that the jury misapplied the law.

That’s how we get to error.

Respondent invokes Stromberg to say if the jury is instructed a right way and a wrong way and they went the right way, we have harmless error.

No.

What we really have is no error at all.

Because there is the — the jury did not misapply the law.

Moreover, Stromberg is not even a correct rule of error.

Because under Stromberg you have error if the jury could have misapplied the law.

John Paul Stevens:

Can I interrupt you?

You are saying no error occurred, and on page 11 of your brief you say,

“An unconstitutional instructional error occurred in this case because there was a reasonable likelihood the jury found. “

and so forth and so on.

Jeremy Friedlander:

Yes.

I agree that error occurred in this case.

What I’m trying to argue here is that Respondent is entirely off base in importing the Stromberg rule into a harmless error determination.

Stromberg is, at most, a rule of error.

The jury went the wrong way.

That is a rule of error, but it is not even a correct rule of error because under Stromberg you say, could the jury have gone the wrong way; whereas, under this Court’s modern precedent you have to have a reasonable likelihood that the jury–

John Paul Stevens:

But you said that’s what happened here.

That’s what your brief says.

Jeremy Friedlander:

–Yes.

I agree that there was an error here but not because of Stromberg.

Because this was a reasonable likelihood that the jury misapplied the law.

Stephen G. Breyer:

But you should have said “no” to the question, I think.

Because I thought the question — whatever it was, you are saying that there isn’t a reasonable likelihood that its error influenced the outcome.

You are not saying that?

Jeremy Friedlander:

There is a reasonable likelihood that the jury misapplied the law.

That’s the error question.

Stephen G. Breyer:

Yes.

Jeremy Friedlander:

Then the harmless error question is: Given that they misapplied the law, do we know under the requisite degree of certainty — Chapman on direct review, Brecht here — that they would still have found him guilty of felony murder under a proper instruction?

John Paul Stevens:

Do you agree that the California Supreme Court gave an incorrect explanation to the answer to that question?

They said there was no harmless error because the — the jury made a specific finding that rebutted that, and that finding depended on the difference between the word “and” and “or”.

Jeremy Friedlander:

I’m not sure I followed the last part of the question, but I agree with you if what you are saying is that the California Supreme Court, in effect, applied a Stromberg-like test.

They said “necessarily resolved”.

John Paul Stevens:

No.

I’m — I’m questioning the basis for the Supreme Court’s conclusion that the error was harmless.

Jeremy Friedlander:

Their basis for the conclusion was the jury’s “special circumstance” finding.

John Paul Stevens:

Right.

Jeremy Friedlander:

Now, they did not consider, when they made that “special” — when they interpreted that “special circumstance”, the “and/or” mistake that was embedded in the instructions.

John Paul Stevens:

Which meant that the jury really hadn’t answered the question they thought it had answered.

Jeremy Friedlander:

No, it didn’t mean that.

We don’t agree with that.

We don’t believe that that is a correct interpretation of the instruction.

But let me skip past that if you’ll allow me.

Even if you — even if you accent any meaning from that “special circumstance” finding — and we don’t agree with that but even if you did — you had at a minimum here the nature–

John Paul Stevens:

If you do that, you are removing the basis for the California Supreme Court’s decision.

Jeremy Friedlander:

–Well–

John Paul Stevens:

Is that not right?

Jeremy Friedlander:

–The basis–

John Paul Stevens:

At least the basis–

Jeremy Friedlander:

–The basis for the California Supreme Court–

John Paul Stevens:

–The basis–

Jeremy Friedlander:

–I think — I think that–

John Paul Stevens:

–for the decision was a — was a specific finding–

Jeremy Friedlander:

–The reasoning behind the decision, yes, I think that’s right.

John Paul Stevens:

–All right.

Jeremy Friedlander:

Okay.

But it doesn’t matter because we are going to apply Brecht anyway.

Samuel A. Alito, Jr.:

Could you clarify the — the category of cases that would be affected if your argument is accepted?

There would be — an error would be — would not be harmless under Brecht if it had a substantial or — or injurious effect, right?

Jeremy Friedlander:

True.

Samuel A. Alito, Jr.:

Now, what is the standard under California law for submitting a theory of liability to the jury?

Presumably, you have to have some evidence in support of it.

Jeremy Friedlander:

I’m not sure I understand the question.

I think that goes to the question of whether there is error in the first place in submitting a factually unsupported theory.

Samuel A. Alito, Jr.:

Well, you are — you are arguing that there are instances in which, when a theory of liability is submitted to the jury, the error — and it’s erroneous — the error can be harmless.

Jeremy Friedlander:

Yes.

Samuel A. Alito, Jr.:

Correct?

Now, what are the — what is the standard under California law for submitting the theory of liability to the jury in the first place?

You have to have — presumably, there has to be some evidence in support of that, right?

How much?

Jeremy Friedlander:

I don’t think — I don’t think I know the answer to that question.

It would certainly be error under California State law if there is a reasonable — if they have the same reasonable likelihood standard that the Federal courts have, if there is a reasonable likelihood that the jury misapplied the law.

Now, in a — but you seem to be focusing on an evidentiary sufficiency, and I think that’s governed by Griffith where if you — if you have an insufficient — if you have insufficient evidence to support the theory, then it’s inappropriate–

Samuel A. Alito, Jr.:

What I’m asking is: When is there going to be a case in which a theory of liability is submitted to the jury and it’s — and it’s erroneous — multiple theories are submitted; one erroneous.

And yet it would turn out that the submission of the erroneous theory did not have a substantial or injurious effect.

Jeremy Friedlander:

–This one — this case — this case was submitted to the jury on two alternative theories: One of them correct, pre-killing aiding and abetting; one of them incorrect, post-killing aiding and abetting.

And it didn’t have a substantial injurious effect because we know that at an absolute minimum here the jury found that the defendant acted — aided and abetted robbery with reckless indifference to human life and was a major participant in the robbery.

And those findings were not compatible with post-killing aiding and abetting.

Ruth Bader Ginsburg:

I thought there was — that there was this — there were three possibilities, and one was that the defendant didn’t do anything before, didn’t do anything during, but was simply an accessory after the fact.

Ruth Bader Ginsburg:

If the jury believed that, then there was no way it could convict him of the crime that he was convicted of.

Jeremy Friedlander:

Well, the jury rejected accessory after the fact but–

Ruth Bader Ginsburg:

How do we know that?

Was there a special verdict?

Jeremy Friedlander:

–They were instructed on accessory after the fact as the lesser offense and they found in essence robbery and felony murder.

David H. Souter:

They found the greater offense, so there would have been no need for them to go to the lesser included offense.

Your trouble is — and I think the problem that we are having is that in finding the greater offense, they may have done so under a theory of accomplice liability that was consistent with the accomplice simply coming in at a late stage in the proceedings.

Jeremy Friedlander:

Yes.

That’s right.

David H. Souter:

I mean, that’s the difficulty we are having.

Jeremy Friedlander:

Yes.

There is a reasonable likelihood that in deciding felony murder they failed to decide that he engaged in the robbery before the killing.

David H. Souter:

Right.

Jeremy Friedlander:

However, they went and found the special circumstance in which they found reckless, aiding and abetting with reckless indifference to human life and as a major participant.

David H. Souter:

I thought the, the difficulty with using that as a means of answering the first error was that the way the special circumstance instruction was phrased, they could have found the special circumstance without finding anything more than that he came in at a late stage.

Jeremy Friedlander:

No.

That’s not correct.

I don’t believe that’s correct.

It is true that part of the instruction, that and/or difficulty, focused on a different question, a different point, said that if you — if they — I’m blanking out but–

David H. Souter:

The second — go ahead.

Jeremy Friedlander:

–It didn’t address the reckless indifference to human life and the major participant, but that was embedded in the instructions, and from the start the courts that have looked at this and have reached this question have all agreed that there was a jury finding of reckless indifference to human life and major participant, and they have disagreed as to what it might have meant.

Okay?

Now, the U.S. district court–

Ruth Bader Ginsburg:

How do we — in this — how do we know what the jury found?

They just came in with a guilty verdict; is that right?

Jeremy Friedlander:

–No.

They came in with a special circumstance finding.

The special circumstance finding that this was–

John Paul Stevens:

Isn’t the special circumstance finding the one that had the word “or” in it, and therefore allowed the jury to give a postkilling interpretation of the event.

Jeremy Friedlander:

–No.

Jeremy Friedlander:

The special–

John Paul Stevens:

I thought that was what the California Supreme Court relied on?

Anthony M. Kennedy:

I agree with Justice Stevens.

I thought that was the whole problem, the “or” and the “and” is the problem for you.

Jeremy Friedlander:

–The — I’ll read from JA 14.

This is the special circumstance instruction that was — was mistaken:

“The murder was committed while the defendant was engaged in the commission or attempted commission of a robbery, or the murder was committed in order to carry out or advance the commission of the crime of robbery, et cetera. “

Now, that — that finding alone we agree if you — if you — if you just follow that finding, that doesn’t implicate the defendant.

But you have to read at JA 13, the instruction, JA 13:

“The defendant with reckless indifference to human life and as a major participant aided and abetted in the commission of a robbery which resulted in the death of Flores. “

That’s at JA 13.

So there were multiple findings here.

And — and the district court which — this is at the bottom of the first paragraph of JA 13.

David H. Souter:

Yes, but isn’t — maybe I’m not getting it.

Help me out here.

One possibility under — under the instruction at JA 13 is

“or assisted in the crime of robbery. “

And the other instructions because they are erroneous leave open the possibility of finding that he assisted in the crime of robbery only if he came in late.

And that’s why it does not solve the original error.

Jeremy Friedlander:

No.

The — the problem with the, I think you’re mixing up two things here.

David H. Souter:

I may be doing that.

Jeremy Friedlander:

The felony murder instructions did not speak to this question of contemporaneity.

They left it open.

That was error.

We acknowledge that that was error.

Okay?

So they arguably find felony murder, aiding and abetting felony murder based on post-killing aiding and abetting.

Now, the jury has to get to the special circumstance finding, right?

David H. Souter:

Yes.

Jeremy Friedlander:

And now it has to decide, among other things, as I — as I read to you,

“did he aid and abet robbery with reckless indifference to human life and as a major participant in the commission of a robbery which resulted in the death of Flores? “

That’s at JA 13.

So what we–

David H. Souter:

Where does it say that he must have assisted prior to–

Jeremy Friedlander:

–It doesn’t.

It doesn’t say that.

David H. Souter:

–Then that doesn’t cure the error.

Jeremy Friedlander:

Yes, it does, and the reason that it does is because the harmless error test is not whether the jury actually made the finding that it needed to make; rather, the harmless error test is do we know from the evidence and from the findings that the jury would have made the findings that it was required to make had it been properly instructed?

John G. Roberts, Jr.:

It seems to me that you’re arguing one of two things: First, that there was no error at all under Stromberg; or second, that if there was error, it was harmless error.

And I thought the case was about whether or not the Ninth Circuit erred in concluding that harmless error analysis didn’t apply to a Stromberg error.

Jeremy Friedlander:

Yes.

I’m not arguing the first point.

I — I concede there was error.

I just wanted to make the point that Stromberg itself is an incorrect rule of error.

But I can see that that’s led us down a path that I didn’t need to go down.

Now, we are certainly arguing harmless error.

As for the Ninth Circuit’s opinion, we said — the Ninth Circuit said this is structural error.

We don’t even start a harmless error analysis.

Ruth Bader Ginsburg:

The district court didn’t say that.

Jeremy Friedlander:

No, it did not.

And — and we said that’s wrong.

This is trial error, not structural error.

Ruth Bader Ginsburg:

So if we were to say, Ninth Circuit, you were wrong, but the district court was right–

Jeremy Friedlander:

Well, the district court was right to the extent that it found it to be trial error.

It was wrong to the extent that it found it to be prejudicial trial error.

But Respondent has agreed with us that this is not structural error, that we need to do a harmless error analysis, and what Respondent has done is import into this harmless error analysis we are saying the Stromberg rule; and it doesn’t belong there.

The Stromberg rule is not a proper rule of harmless error.

It’s at most a rule of error.

I’d like–

John Paul Stevens:

But do you not agree that the harmless error — the — the rationale of the California Supreme Court was incorrect?

Jeremy Friedlander:

–Was the rationale incorrect?

If–

John Paul Stevens:

In the harmless error analysis.

Actually, I’m agreeing with you that there should have been harmless error.

Did it perform the proper–

Jeremy Friedlander:

–It’s incorrect in the way that you’re thinking it is incorrect, but it was also incorrect in another way.

It was incorrect in a way that it applied too strict a standard.

It said has this question been necessarily been resolved; and it said yes, it has been necessarily resolved.

Well, that’s not the question.

The harmless error question–

John Paul Stevens:

–The — answer was incorrect, too, wasn’t it?

And that answer was incorrect?

Jeremy Friedlander:

–That answer was incorrect; and as well as what you’re thinking, it was incorrect, in that they premised it on a factual basis that was not in fact the case.

But I better reserve time for rebuttal.

John G. Roberts, Jr.:

Thank you, counsel.

Mr. Shah.

Mr. Chief Justice, and may it please the Court: I think the Chief Justice framed the main issue in this case properly.

Consistent with the last two decades of this Court’s instructional error jurisprudence, the Court should hold that the type of error at issue is not structural and is subject to harmless error review.

The Ninth Circuit’s contrary interpretation of Stromberg creates the untenable result that adding an invalid theory — adding a valid theory to an invalid one somehow makes the error worse.

Even Respondent now rejects that interpretation.

It follows that this Court should reverse the Ninth Circuit and remand for application of harmless error review under Brecht.

The main issue–

Ruth Bader Ginsburg:

That’s what the district judge did, right?

–Yes, Your Honor.

The district court did apply harmless error review under Brecht.

The court of appeals, of course, did not.

Ruth Bader Ginsburg:

It could just review the district court’s decision using the standards that the district court used?

This Court could do that, but the longstanding and customary practice of this Court is that when the court of appeals does not address the harmless error question, that it remands to that court.

It did that in Rose V. Clark, in Pope V. Illinois and California V. Roy, and more recently in Recuenco.

And in several of those cases the intermediate court or the district court, in the habeas context, did address it and the court of appeals didn’t in all of those cases that the Court did remand for application of harmless error.

Ruth Bader Ginsburg:

I wasn’t questioning that.

The issue, is the United States position say structural error was wrong, harmless error was right, and then just have the Ninth Circuit review the district court’s decision that held that there was harmless error the way it would ordinarily review a district court decision?

Yes, Your Honor, but I would add that this Court should also clarify the scope of that harmless error inquiry.

I think there are two important principles.

There is no logical reason to apply a narrower form of harmless error review in this case than the Court did in Neder.

So I think that this Court should clarify that the principles of harmlessness articulated in Neder should apply on remand to the Ninth Circuit.

I think those of those principles bear mention.

First, the question is not what the jury actually or necessarily found, but rather, it encompasses what a rationale jury would have found absent the error.

And I think the second important point for this Court to clarify is that harmlessness review is not limited to cases, is not limited to cases where the relevant elements are undisputed, but rather even when disputed, a reviewing court should consider the entire record and determine whether, in light of the jury finding, there is sufficient evidence to support a contrary verdict.

And then it should remand, Your Honor, to the court of appeals to apply those principles.

Antonin Scalia:

Why do you say in light of the jury finding when you have disavowed any necessity that the jury had found this?

Well, Your Honor, we do–

Antonin Scalia:

I mean

“in light of the jury finding. “

why don’t you just drop that.

–Well, Your Honor, we think it’s important because there are certain jury findings that we know the jury made in this case.

The jury returned a special circumstance verdict.

And we know based upon that verdict that at a minimum, the jury found that Respondent was a major participant in the crime — just taking the facts of this case, the Respondent was a major participant in the robbery, that he exhibited a reckless disregard for human life.

And we also know that the jury rejected a duress defense.

The jury was instructed on that defense, and yet the jury came back guilty.

So I don’t think that the — that the inquiry should take place in a vacuum, but, of course, the reviewing court should have the benefit of the limited jury findings that we do know were made, and then apply the sufficiency inquiry in light of those jury findings.

There is no reason to disregard the jury findings that were made.

John G. Roberts, Jr.:

Suppose it’s not, in the abstract anyway, necessarily true that this case follows a fortiori, because here, at least you have a valid theory.

That may make it a harder case, because the jury had based its entire verdict on an invalid theory.

It may be more difficult than a situation where you’ve got a theory of the case that is invalid because a particular element is missing.

And you can look and see whether or not that was harmless or not.

I would disagree, Mr. Chief Justice.

I think there are two possibilities in a case with alternative theories.

One possibility is that the jury relies upon the defective instruction.

For example, there is one element omitted just like the instruction in Neder.

If that’s what the jury had done, then this case is no different than Neder, no better, no worse, right.

The other possibility, however, is that the jury relies on the completely valid theory.

If the jury relied on the completely valid theory, then there is even less of a problem than Neder.

There is no problem at all.

So I don’t see how the error in this alternative valid theory and defective theory circumstance could be any–

John Paul Stevens:

The difference is when you only got one theory, you know what he relied on.

When you got alternatives, how do you know which one he relied on?

–Well, you don’t — you don’t, Your Honor.

And that’s what the reviewing court has to decide.

But in either scenario, if it — if you assume that the jury relied on the defective theory, then we are in the same box as in Neder, and the reviewing court would decide whether there was — you know, they would apply the harmless error review to the missing element and decide whether there was sufficient evidence to render that error harmless; or the court could ask, well, is that uncontraverted or overwhelming evidence that the defendant would have been found guilty under the valid theory.

And that would also be a permissible inquiry.

John G. Roberts, Jr.:

Well, I guess, though, the point is when you have invalid and valid theories, it would be a little harder than Neder, because they don’t have to make any findings under the valid theory.

You know, because you don’t just have to fill in a missing piece of the puzzle, as in Neder.

You have to — they might have been working on an entirely different puzzle.

Well, Your Honor, in almost all conceivable cases there is going to be at least some overlap between — between the two theories.

And what the — the reason why that’s relevant is because you know that when the jury has returned a general verdict of guilt, they have necessarily found beyond a reasonable doubt that each of those elements has been satisfied.

John Paul Stevens:

No.

But in this case the question is whether the uncle did it or the nephew did it.

And if there is a lot of doubt about — about — if there is substantial reason to believe maybe it was the uncle, they might have taken the easy case out.

I think this is vastly different from a single theory case.

Well, Your honor, here’s why–

John Paul Stevens:

Because it’s an — there is plenty of evidence that he might have been guilty after the fact.

There is no doubt about that.

But there is doubt about the former.

–Your Honor, here’s why I think the cases aren’t that different, because, you know, at a minimum from the general verdict of guilt if you take a case like this, you know that the jury found that there was an unlawful killing.

You know that the jury found that there was a robbery.

You know that the jury found that Respondent was involved in the robbery.

The only element that we don’t know–

John Paul Stevens:

Perhaps after — after the death occurred.

John Paul Stevens:

That’s all we really know.

–It is possible that the jury would have found that, and that’s the question that the reviewing court would perform in harmless error review, just like if there — basically that is the exact same question the Court would form, would do under Neder if the timing element were omitted all together.

The reviewing court would look at the evidence and decide is there sufficient evidence to support an after-the-fact participation.

John G. Roberts, Jr.:

So what we are talking about, in other words, is the application of harmless error, which may or may not be harder in the multiple theory case, and not the question of whether harmless error analysis should be applied.

Yes, Your Honor.

That’s going to be completely fact dependent whether it’s harder or not.

But I think analytically it’s exactly the same as in Neder.

Samuel A. Alito, Jr.:

Aren’t there going to be cases in which there’s enough evidence to support the submission of a theory of liability to the jury, but so little that the court is going to be able to say that it’s harmless under — that the submission of the invalid theory was harmless under Brecht?

Yes.

I mean, you can image in a case — take Neder.

By this Neder instruction, the one omitting the materiality element, was submitted to the jury certainly enough evidence to support it, and then there was another valid theory–

Samuel A. Alito, Jr.:

There is difference.

An element of the offense has to be submitted to the jury.

And it may be that that it’s — the element is undisputed, but a theory of liability isn’t going to be submitted to the jury unless there is some evidence to support it.

–Well, Your Honor, if — if there is some evidence to support it, if we were in a direct review context, that’s the inquiry that Neder says.

So on a clean slate it might be that if there is sufficient evidence to support a contrary verdict, that that would end the inquiry.

But in many of these cases we are going to know that the jury has made at least some findings, has found some of the elements.

And that might inform the harmless error inquiry, just like in this as we know that the jury has found things like a respondent was a major participant, exhibited a reckless disregard for human life, discounted a significant portion of this — of his statement by rejecting the duress defense.

If there are no further questions, Your Honor.

John G. Roberts, Jr.:

Thank you, Mr. Shah.

Mr. O’Connell.

J. Bradley O’Connell:

Mr. Chief Justice, and may it please the Court: Unlike the Sarausad case argued earlier this morning, this is a case of conceded constitutional error under the reasonable likelihood standard, the same question that consumed so much of the Sarausad argument.

We acknowledge that this is a trial error and is subject to the Brecht prejudice standard.

And as I believe Justice Souter commented during the Sarausad argument, in a Brecht situation, the defendant wins the benefit of the doubt.

That is a lesson of this Court’s decision.

Stephen G. Breyer:

I agree with that.

And my question is going to be — at some point, both of you seem to agree with that so — pretty much, it’s a harmless error standard of review.

I think a wrote a case on that, which says that–

J. Bradley O’Connell:

Yes.

You wrote on O’Neal, Your Honor.

Stephen G. Breyer:

–Right.

So why don’t we just send it back and say apply the standard?

J. Bradley O’Connell:

Your Honor, although the Ninth Circuit admittedly used a structural defect nomenclature with which we don’t agree, we believe that the Court can’t — that upon considering both the Ninth Circuit decision and the district court decision, this Court can affirm because the inquiry which the Ninth Circuit did conduct was actually a case-specific inquiry, not really a structural defect inquiry, and when an examination of the evidence and the verdicts and the instructions and anything else indicative of the jury’s thinking still leaves the court uncertain whether the jurors relied on the valid or the invalid theory, that’s a substantial and injurious influence.

That is grave doubt.

That is equipoise, in the language of Your Honor’s opinion.

Stephen G. Breyer:

You’re saying that I should have grave doubt myself and any judge would have to.

He made an argument here I think that wasn’t such a bad one.

He said look at page 14 and you’ll see the mistake, you know, in the joint appendix.

J. Bradley O’Connell:

Yes, Your Honor.

Stephen G. Breyer:

There it is.

And it’s absolutely wrong.

The basic instruction under the law says you have to be engaged in the crime and you have to have a certain state of intent.

You have to have committed this murder in order to help the crime or at least in order to escape.

J. Bradley O’Connell:

Yes, Your Honor.

Stephen G. Breyer:

And that’s wrong.

Put them together, but then look at the instruction on page 13, which he actually gave, and the one on 13 really does seem to say that, or more so.

It says, you can only apply — jury, you can only apply this special instruction if two things are true: First, in lines 9 through 12 of that instruction, it seems to repeat that you have to be engaged in the crime.

And then just before that, it says, and you have to have one of two states of mind.

You have to either intend to kill the person or at least be recklessly indifferent.

And since that’s what we said on page 13, the fact we made a mistake on page 14 doesn’t matter that much.

That’s one of his arguments that I’m sure they have a lot of evidentiary ones and a bunch of other ones.

J. Bradley O’Connell:

Your Honor, I think the answer to that question is in the underlying trial reporter’s transcript at pages 1015 to 1016, the district attorney’s cross-examination of Mr. Pulido.

The district attorney developed the theme quite effectively that, even under Mr. Pulido’s own account, he exhibited reckless indifference because he saw the mortally wounded victim lying there, and the D.A. said, did you try to resuscitate him?

Have you ever had a CPR class?

I did once.

Well, did you do CPR on him?

No.

Did you call 9-11?

You didn’t care whether he lived or died, did you?

So this entire line of cross-examination, to which the district attorney came back again, though somewhat less dramatically, at 1330 of the transcript and again at 1337 to 1338, commenting that he — that Mr. Pulido did not go back in at some later point to check on the victim’s condition, that entire line of cross-examination encouraged the theme that even under Mr. Pulido’s own account of the facts, he showed reckless indifference.

J. Bradley O’Connell:

And–

John G. Roberts, Jr.:

So the error wasn’t harmless.

And that’s an argument you can make before the Ninth Circuit on remand.

J. Bradley O’Connell:

–That certainly is an argument that if the Court — if the Court is convinced that the Ninth Circuit’s analysis in substance — because I realize the nomenclature — the nomenclature was inconsistent with harmless error, and I’m not going to defend that.

If the Court is convinced that the substance of the Ninth Circuit’s analysis was so far removed from what harmless error analysis requires in this context, we agree that the usual course would be to send it back.

John G. Roberts, Jr.:

Well, why don’t we just take them at their word?

As you say, they are applying the wrong nomenclature.

It would be easy enough for them when they get it back to say, oh, we meant, you know, harmless error under Brecht.

But that seems to me to be at least an open question.

J. Bradley O’Connell:

Your Honor, I think I could best answer that by offering my own critique of the Ninth Circuit opinion and — which may explain why it may be unnecessary.

In my opinion, what the analysis that the Ninth Circuit conducted was equivalent to a sufficient analysis, equivalent to the point of equipoise discussed in O’Neal.

That is to say, it reviewed the evidence and said — and the verdicts and the instructions — and said we’re uncertain, we can’t tell which way the jury went.

My critique of the Ninth Circuit’s — at least its per curiam opinion, as opposed to one of the concurring opinions, is that in fact this is not just a general verdict case.

We know a great deal about the jurors’ thinking in this case.

We know that by a vote of either 8-4 or 4-8, the jurors rejected the prosecution’s primary theory, which was that Pulido personally shot the victim and that, in the prosecutor’s words, Aragon had nothing to do with this, Aragon wasn’t there, Aragon isn’t a murderer.

We also know — and this will be another echo of the Sarausad case — we also know that throughout their five days of deliberations, the jurors submitted question after question directed to what are now — what are concededly defective erroneous instructions, and moreover, those questions focused right in on the defect in the instructions and in particular the timing of Mr. Pulido’s assistance.

And I refer the Court in particular to the jurors’ handwritten diagrams of alternative conceptions of felony-murder aiding and abetting liability, which are at 36 to 38 of the joint appendix, and also the jurors’ question at page 41 of the joint appendix whether — this goes right — this goes, as the district court said, to the crux of the issue: Does the defendant have to have knowledge before or during the crime of the — of the unlawful purpose of the perpetrator?

And as to both of those questions the judge said: Go back and reread those same instructions, which of course were erroneous.

So my critique of the Ninth Circuit opinion, which is again slightly different, I think, than the State’s, would be that it neglected to mention that compelling evidence that in fact the jurors actually found a set of facts corresponding to the invalid late joiner theory, because in my view what — the analysis the Ninth Circuit conducted gets us to a point of equipoise, and that’s enough for Brecht.

The defect in the Ninth Circuit’s analysis is they neglected to also factor in those juror queries, and in my view the juror queries move us from a state of equipoise to a very high probability that the jurors relied on the — and adopted a factual scenario corresponding to an invalid theory.

John G. Roberts, Jr.:

Under Brecht, you have the burden of showing a substantial injurious effect.

J. Bradley O’Connell:

Well, actually, Your Honor, as understand O’Neal the Court specifically renounced describing it as a burden, and in fact said that we are not going to use the “burden” terminology, and rather than talk about allocating burdens, we are going to address it from the perspective of the judge and say, if the judge is in equipoise, that is sufficient.

We also know that — drawing upon the relationship between the Brecht standard and others, we know that a Strickland standard is deemed to — that if you have Strickland prejudice, you don’t have to go through a separate Brecht analysis.

Because of Strickland, reasonable probability also satisfies Brecht’s substantial and injurious influence.

Well, this Court has said again and again, in construing Strickland, it does not require a more likely than not showing; it simply requires a showing sufficient to undermine confidence in the outcome.

And–

John G. Roberts, Jr.:

So you think you win under Brecht?

J. Bradley O’Connell:

–I think we will under Brecht, Your Honor.

John G. Roberts, Jr.:

You think the Ninth Circuit, although it didn’t use that label, actually applied the Brecht standard?

J. Bradley O’Connell:

In substance, it did even though, as I said, I will — I can’t — I can’t run from the language the Ninth Circuit actually used.

John G. Roberts, Jr.:

Sure.

Well, then it seems to me you agree with your friend that the Brecht standard — and the government — that the Brecht standard applies.

J. Bradley O’Connell:

Yes, we do, Your Honor.

John G. Roberts, Jr.:

And all the fight is over is whether the Ninth Circuit applied that standard or not?

J. Bradley O’Connell:

Yes, Your Honor, whether it applied it in substance.

John G. Roberts, Jr.:

Why don’t we just send it back and ask them?

Say: We can’t tell.

The parties are having disagreement about whether you did.

What did you do?

J. Bradley O’Connell:

Your Honor, as I said a little earlier, if I fail to persuade the Court that the substance of what the Ninth Circuit did was — effectively satisfied Brecht as interpreted in O’Neal, then I agree that’s an appropriate course, and we indicated that in the final section of our brief.

There is not–

Stephen G. Breyer:

I’m sure that would be — I mean, what worries me is proliferating arguments among judges.

Why not just say, you know, apply Brecht.

I’d like it with the O’Neal clause because it seems to me the O’Neal clause applies.

J. Bradley O’Connell:

–I agree, Your Honor.

Stephen G. Breyer:

Just go do it.

And then they can say “we already did it”, or they can say

“we didn’t already do it, but we’ll do it now. “

Let them say what they want to say.

J. Bradley O’Connell:

Your Honor — and, again, I don’t — I don’t strenuously object to that because, as I’ve mentioned if the court, if the court remains, remains of the view that what the Ninth Circuit did was simply incompatible with Brecht, then I agree it should be sent back and I — and I understand why that might be the Court’s inclination.

Let me offer another angle for why what the Ninth Circuit did is entirely compatible with Brecht.

In civil cases, this Court routinely applies the same sort of general verdict analysis which in criminal cases has become to be known as the Stromberg line.

In fact, it did it as recently this summer in the Exxon Valdez case.

Now civil cases don’t have Chapman, they don’t have structural defect; civil cases are subject to a harmless error standard which ultimately goes back to the identical statute which was construed in Kotteakos.

And going back to the opinion of the Court in O’Neal, in addressing the relationship between Kotteakos/Brecht harmless error review and civil harmless error review, the Court indicated he think rather strongly that those standards are either equivalent or to the extent that they differ, the Brecht/Kotteakos standard is less forgiving of error because of the liberty interest at stake.

Now that means two things.

From a retrospective point of view of simply construing this Court’s different lines of cases, I think it demonstrates that this general verdict analysis is entirely consistent with a Kotteakos type harmless error analysis, but it also has some perspective implications, because if this Court were to wholly disavow that type of analysis and say that has nothing to do with harmless error review, that’s going to impact civil cases, too; and obviously in all this Court’s opinions, the Court is considering not only the immediate case, but what is going to happen if we — if we rule one way or another.

So this case — this case supports a finding of prejudice under that O’Neal equipoise and as for the Ninth Circuit did, that’s not what it called it, it even more strongly supports prejudice when one factors in the rest of the record, notably the juror — the juror queries and another item which I think may be responsive to some of Justice Alito’s earlier questions, which is the strength of the evidence on the various theories.

There were three theories and obviously there was evidence in support of the prosecution’s main theory that Pulido shot the victim himself.

There was evidence — obviously there was substantial evidence in the form of Pulido’s own testimony in support of the invalid late-joiner theory.

J. Bradley O’Connell:

Out of these the one theory that had least support in the record before the jury was the valid aiding and abetting theory, because no witness testified to a scenario in which Pulido assisted in some fashion or another prior to the shooting; nor did the prosecutor attempt to develop that type of scenario in argument because the prosecutor was very clear: Prosecutor said you’re going to get aiding and abetting instructions just in case.

They only apply — if — they only come into play if you were to believe the defendant’s story which was a real whopper; but if you believe the defendant’s story the felony murder rule applies.

But then the prosecutor immediately went back,

“but I don’t think it applies at all, Aragon wasn’t there. “

“Aragon didn’t do it. “

“Pulido did it. “

and prosecutor went back to his primary theory.

So the theory which the — present in the instructions was least before the jury, was the valid aiding and abetting theory, yet defining this conceded error harmless depends on the notion either that the jury actually adopted a pre-shooting aiding or abetting or that they necessarily would have adopted one if properly instructed.

There is one final point I’d like to make, which is there is a qualitative aspect to Brecht; and when one has an error under which the defense testimony even if believed renders the defendant guilty erroneously, that derails the entire defense.

That it seems to me is the quintessential substantial and injurious influence; because it practically renders the trial nullity by erroneously telling the jury even if you find this set of facts, the defendant is guilty.

So for these reasons and additionally for the very thorough prejudice analysis found in the district court opinion, we urge the Court to affirm, despite the Ninth Circuit’s inapt nomenclature, but in the event the Court is dissatisfied with that explanation, we have acknowledged that the appropriate course in that instance would be to send it back under Brecht.

Unless the Court has any further questions, I’m prepared to submit.

John G. Roberts, Jr.:

Thank you, Mr. O’Connell.

J. Bradley O’Connell:

Thank you, Your Honor.

John G. Roberts, Jr.:

Mr. Friedlander, you have three minutes remaining.

Jeremy Friedlander:

We agree that we need to do Brecht.

We disagree on what Brecht is.

Respondent is telling you that Brecht is the error test, whether the jury relied on the valid or invalid theory.

That’s what he says Brecht is.

In Calderon v Coleman, this Court said there is a difference between the error test and the harmless error test.

All these questions that the jury asked went to the question of error.

Did the jury misapply the law by failing to find a contemporaneity aspect of felony murder?

They did not go to the harmless error question of what the jury would have done had it been properly instructed.

We urge the Court in remanding the case to make clear that the error analysis must be kept separate from the harmless error analysis.

This is a mistake Judge Thomas made in his concurring opinion.

We — he assumes — he sees a mistake in an instruction; he assumes that the mistake rises to the level of constitutional error; and then he does a prejudice analysis by deciding whether the mistake caused the jury to misapply the law.

In other words, he — he applies an error analysis to decide prejudice.

And what he does is he includes all this evidence of the questions the jury asked when those questions went only to the misapplication of law that was the error, and did nothing to show the basis of the jury’s special circumstance finding, which is the heart of the harmless error determination here.

If there are no further questions.

John G. Roberts, Jr.:

Thank you, Mr. Friedlander.

John G. Roberts, Jr.:

The case is submitted.