Cosme v. State of Florida

DISTRICT COURT OF APPEAL OF FLORIDA

SECOND DISTRICT

JOSE LEONARDO COSME, JR.,

Appellant,

v.

STATE OF FLORIDA,

Appellee.

No. 2D2025-0001

May 22, 2026

Appeal from the Circuit Court for Pasco County; Joshua Riba, Judge.

Matthew R. McLain of McLain Law, P.A., Longwood, for Appellant.

James Uthmeier, Attorney General, Tallahassee, and Krystle Celine

Cacci, Assistant Attorney General, Tampa, for Appellee.

KHOUZAM, Judge.

Jose Leonardo Cosme, Jr., appeals his judgment and sentence on

two counts of lewd or lascivious molestation on a victim less than twelve

years of age. Because the State improperly bolstered the victim's

testimony, we are constrained to reverse for a new trial.

Cosme was charged with the lewd or lascivious molestation of his

goddaughter, A.D. At trial, A.D. testified that she believed Cosme was

her father. She would visit and stay with him often. She was about

eight years old when Cosme began touching her inappropriately. The

abuse, which she described in detail during her testimony, would occur

almost every time she visited him. This continued until, at ten years old,

she disclosed the abuse to a school guidance counselor. She then

recounted the abuse to a school resource officer and a child protection

team interviewer. Without any physical evidence of abuse, the victim's

allegations were the crux of the State's case against Cosme, and his only

theory of defense was that the victim's story was fabricated.

The State presented recordings of the victim's statements through

testimony from the child protection team interviewer and school resource

officer who interviewed her. But the State went further by also eliciting

testimony from these witnesses to vouch for the victim's credibility. For

example, the child protection team interviewer testified that he observed

no signs of deception from the victim; that the victim had no motive to

fabricate; that her disclosure, unlike in other cases, did not appear

influenced; and that he "found no inconsistencies" in her account. Then,

the school resource officer testified that the victim was successfully truth

qualified; that she appropriately answered his questions; that she "knows

the difference between telling the truth and a lie"; and that he, from his

perspective as a law enforcement officer, "did not see any signs of

deception with her." After that, the State called another officer, who had

never had any personal interaction with the victim, to testify based on

his professional review of the reports and the victim's interviews that she

was "reliable" and "very consistent" and that "each time she gave the

account, it was described the same way." He contrasted her behavior

with that of children who lie, opining that she exhibited none of the

typical discrepancies.

2

Cosme argues that, in the context of his trial, this testimony

constitutes improper bolstering and that the trial court reversibly erred

by allowing it. We agree. "It is elemental in our system of jurisprudence

that the jury is the sole arbiter of the credibility of witnesses. Thus, it is

an invasion of the jury's exclusive province for one witness to offer his

personal view on the credibility of a fellow witness." Davis v. State, 360

So. 3d 809, 814 (Fla. 2d DCA 2023) (quoting Sierra v. State, 230 So. 3d

48, 51-52 (Fla. 2d DCA 2017)). "This principle applies equally in cases

where the testimony offered directly comments on the truthfulness of a

minor victim." Smith v. State, 292 So. 3d 46, 48-49 (Fla. 5th DCA 2020)

(collecting cases). Indeed, where a case turns on the credibility of the

victim—a common occurrence in molestation cases like this one—

improper bolstering can be particularly harmful. See Johnson v. State,

177 So. 3d 1005, 1008 (Fla. 1st DCA 2015) ("Improper bolstering of a

witness is especially troubling in a case that rests solely on competing

witness accounts to establish an element of the crime."); see also

Cavaliere v. State, 147 So. 3d 628, 630 (Fla. 2d DCA 2014) ("Because the

primary evidence against Cavaliere was the victim's testimony, we cannot

say that the State has proved beyond a reasonable doubt that the error

was harmless.").

"[W]hen a police officer, who is generally regarded by the jury as

disinterested and objective and therefore highly credible, is the

corroborating witness, the danger of improperly influencing the jury

becomes particularly grave." Davis, 360 So. 3d at 814 (quoting Martinez

v. State, 761 So. 2d 1074, 1080 (Fla. 2000)). "Thus, '[w]here a police

officer's testimony is used to bolster the credibility of a victim, the error

cannot be deemed harmless.' " Id. (alteration in original) (quoting Sierra,

230 So. 3d at 52); see also Cavaliere, 147 So. 3d at 629 ("[E]rrors

3

committed from the improper admission of a police officer's testimony

that can be used to bolster the credibility of a victim's trial testimony

cannot be deemed harmless." (quoting Lee v. State, 873 So. 2d 582, 584

(Fla. 3d DCA 2004))). For example, in Cavaliere, 147 So. 3d at 630, this

court held that harmful error occurred when a detective improperly

vouched for a child victim by telling the jury what truth qualification is,

that the child victim had been truth qualified, and that "by looking at her

and kind of getting a feel for her I could tell that, you know, she was—

she was acting appropriately. This wasn't a joke to her."

Civilians are also prohibited from vouching for the victim's

credibility. "Although the prohibitions on civilians vouching for a victim's

credibility have not been treated as harshly by the courts, such

testimony can still be harmful error." Cavaliere, 147 So. 3d at 629 (citing

Rhue v. State, 693 So. 2d 567, 568 (Fla. 2d DCA 1996)). For example, in

Paul v. State, 790 So. 2d 508, 510 (Fla. 5th DCA 2001), the court found

that an investigator for the child protection team "improperly vouched for

the credibility of the victim by stating that the victim was honest, that

she heard nothing which would indicate that the victim had fabricated

her statements, and that the victim would have no motive for lying."

Similarly, in Fuller v. State, 540 So. 2d 182, 183-84 (Fla. 5th DCA 1989),

the court found that the medical director of the child protection team

improperly vouched for the child victim when he testified, "I believe [the

child] told me the truth" and "I don't have any reason to believe that he

lied to me." (Alteration in original.) There, the court concluded that "[i]t

was error to ask [the director] to vouch for the credibility of the child, and

that error cannot be considered harmless." Id. at 184.

We review the trial court's ruling on the admissibility of evidence

for an abuse of discretion, but the trial court's discretion is limited by the

4

rules of evidence. See Smith, 292 So. 3d at 48. The State has the

burden to show that any error was harmless beyond a reasonable

doubt—"or, alternatively stated, that there is no reasonable possibility

that the error contributed to the conviction." State v. DiGuilio, 491 So. 2d

1129, 1138 (Fla. 1986). "Application of the [harmless error] test requires

not only a close examination of the permissible evidence on which the

jury could have legitimately relied, but an even closer examination of the

impermissible evidence which might have possibly influenced the jury

verdict." Id.

Here, the defense appropriately objected when the child protection

team interviewer, Matthew Bonner, testified as follows:

Q: And in speaking with [the victim], were you able to discern

any sort of motive of her to make this up or anything to that

effect?

A: Like deception?

Q: Yes.

A: No.

The court overruled the objection. This was a clear abuse of discretion.

After the court erroneously allowed this testimony, the State

proceeded to severely compound the error. The State further questioned

the interviewer:

Q: Okay. And was [the victim] resolute with her disclosure?

A: I found no inconsistences.

Then, the State called multiple law enforcement officers who improperly

bolstered the victim's testimony. Deputy Shaun Duval, the school

resource officer who interviewed the victim, testified:

Q: And were you able to successfully truth-qualify [the

victim]?

5

A: Yes. She answered the question appropriately, and the

way she interacted and spoke, I feel that she knows the

difference between telling the truth and a lie.

....

Q: And throughout your interview with her, did you, as a law

enforcement officer, see any signs of deception?

A: No. I did not see any signs of deception with her.

Officer Clifford Williams, who had never even interacted with the victim,

testified:

Q: And in your capacity as the Major Crimes detective at the

time, did you compare Deputy Duval's report to the report of

the subsequent forensic interview of [the victim]?

A: Yes. I believe it was a Mr. Bonner, he completed the

forensic, then I had Deputy Duval, and even statements from

the guidance counselor, everything had—I mean, she was

reliable.

....

Q: And in looking at the two reports of her two interviews,

did you find that [the victim] was consistent in what it is that

happened to her?

A: Yeah. Very consistent. In the majority of the time if

children are lying, they will change things or omit things or

there's some type of discrepancy. And each time she gave the

account, it was described the same way.

Especially when considered all together, it is undeniable that this

improper testimony placed a "cloak of credibility" upon the victim's

testimony, thereby usurping the jury's role. See Cavaliere, 147 So. 3d at

629-30 (quoting Lamb v. State, 357 So. 2d 437, 438 (Fla. 2d DCA 1978)).

The State then further exacerbated the error in closing argument

by expressly relying on the improper testimony in order to establish the

victim's credibility. For example, the State commented:

6

First we heard from [the victim] herself. Fourteen yesterday.

Ten at time of disclosure. Eight to ten at the time Mr. Cosme

was molesting her. She got up there and she was open,

honest, forthcoming with you. She answered questions from

both myself and the defense attorney. She said when she

didn't know. She said when she didn't remember.

....

She's never once said, no, he didn't do that, never once. That

would have been the easy way to get out of this, not having to

be here, to say, no, I don't want to do it, it didn't happen.

And that never happened because it did happen. And that is

why a child has been consistent for something that happened

as long as five years ago, because it happened to her. She

remembers unfortunately. But she remembers.

....

As Matthew Bonner told you . . . she was resolute. She was

direct. She answered. If there was something—he asked her

a ton of questions. . . . And she was open and honest and

she answered those questions. . . . She was honest. She was

always quick. Shook . . . her head, no, no, no.

These comments in closing again placed the prestige of the government

behind the victim's testimony and invaded the jury's factfinding role. See

Johnson, 177 So. 3d at 1008 ("Improper bolstering occurs when the State

place[s] the prestige of the government behind the witness or indicates

that information not presented to the jury supports the witness's

testimony." (quoting Spann v. State, 985 So. 2d 1059, 1067 (Fla. 2008))).

The totality of this improper testimony as well as commentary from

the State was especially harmful because there was no physical evidence

in this case and therefore the victim's credibility was the pivotal issue for

the jury to evaluate at trial. Under these circumstances, we cannot say

that the erroneous admission of testimony improperly bolstering her

credibility was harmless, and we are compelled to reverse for a new trial.

7

We reject the State's argument that this issue has not been

preserved because the defense objected only to Interviewer Bonner's

testimony. As explained above, the cumulative effect of the testimony

and commentary at issue rendered the preserved error harmful in the

particular context of this case. But it appears that the cumulative effect

of these errors would likely constitute fundamental error as well. We

have previously held that fundamental error occurred when the

prosecution presented testimony from a law enforcement officer

improperly bolstering the testimony of a witness for the State. Davis,

360 So. 3d at 811. Similarly here, the egregious bolstering from multiple

State's witnesses, including two law enforcement officers, went "to the

foundation of the case . . . and [were] equivalent to a denial of due

process." Id. at 814 (quoting F.B. v. State, 852 So. 2d 226, 229 (Fla.

2003)). And building upon the improper bolstering, the State expressly

relied upon the problematic testimony in closing and thereby "engaged in

such . . . impermissible argument as to vitiate the fairness of the

proceedings." See Fuller, 540 So. 2d at 184.

We also reject the State's argument that the testimony regarding

the child victim's reliability was appropriately presented to the jury

pursuant to the hearsay exception for child victim statements set forth in

section 90.803(23), Florida Statutes (2024). This subsection provides for

the admissibility of hearsay statements of a child victim under certain

circumstances and directs the trial court to hold "a hearing conducted

outside the presence of the jury" to determine whether "the time, content,

and circumstances of the statement provide sufficient safeguards of

reliability." See § 90.803(23)(a)(1) (emphasis added). This subsection

does not permit witnesses to vouch for the child's credibility in front of

the jury—which is what occurred here. Nevertheless, the State

8

incorrectly and unwaveringly maintained throughout this appeal that the

testimony in question was appropriate under this hearsay exception.

Indeed, at oral argument in this court the State could not even provide a

valid purpose for Officer Williams' testimony.

Finally, we address Cosme's argument that he should have been

allowed to offer in support of his defense of fabrication that the victim

had viewed sexually explicit material. We agree. The record shows that

during his own testimony, Cosme proffered the following exchange:

Q: Did you see a video on [the victim's] phone?

A: Yes, I did.

Q: And what did you see?

....

A: So I saw nude photos and videos of her mom, you know,

at the time with her then boyfriend doing things that, you

know, adults would do, having sex, basically.

And I told her it wasn't a phone that she needed to have in

her hand. So I called the mom, I told the mom this is not a

phone she should have. The mom then said, okay.

I apologized, took the phone back, and I bought [the victim]

her own cell phone.

The court excluded this testimony on the basis that what the victim

allegedly saw on the video was not sufficiently similar to her allegations

against Cosme. But the victim alleged that Cosme put her on his lap

while he was lying down and rocked her back and forth over his erect

penis—essentially, mimicking sexual intercourse.

This court has held that testimony that a child victim knew about

certain sexual activities "was relevant—in fact, critical—to [the

defendant]'s claim of fabrication" and that "[h]owever light the weight of

9

[the] testimony, the jury should have been given the opportunity to

evaluate it." Dixon v. State, 605 So. 2d 960, 961-62 (Fla. 2d DCA 1992).

On remand, Cosme should be allowed to present his proffered testimony

in support of his theory of defense. See Mateo v. State, 932 So. 2d 376,

379 (Fla. 2d DCA 2006) ("[A]s a general proposition, any evidence that

tends to support the defendant's theory of defense is admissible, and it is

error to exclude it.").

We recognize the great burden of conducting a new trial, especially

in a molestation case where the child victim will need to take the stand

again. But where the State has violated the defendant's due process

rights so clearly and egregiously, we have no choice other than to order a

new trial. We trust that the State will be mindful to avoid the same error

again on remand and in other similar cases in the future.

Reversed and remanded for a new trial.

NORTHCUTT and VILLANTI, JJ., Concur.

__________________________

Opinion subject to revision prior to official publication.

10