In the
 United States Court of Appeals
              For the Seventh Circuit
                        ____________

No. 02-1896
AUDREY A. EDMUNDS,
                                             Petitioner-Appellant,
                               v.

JODINE DEPPISCH,
                                            Respondent-Appellee.
                        ____________
           Appeal from the United States District Court
              for the Eastern District of Wisconsin.
             No. 00-C-1089—Lynn Adelman, Judge.
                        ____________
  ARGUED NOVEMBER 13, 2002—DECIDED DECEMBER 18, 2002
                        ____________


  Before POSNER, COFFEY, and MANION, Circuit Judges.
  POSNER, Circuit Judge. A jury in a state court in Wiscon-
sin convicted Audrey Edmunds of first-degree reckless
homicide in the death of a seven-month-old baby, Natalie
Beard, for whom Edmunds was providing day care. Ed-
munds was sentenced to 18 years in prison. After exhaust-
ing her state remedies, see State v. Edmunds, 

(Wis. 1999); State v. Edmunds, 

(Wis. App.
1999), she sought federal habeas corpus, lost, and now
appeals. She argues that the exclusion of evidence of the
parents’ demeanor on the day of their baby’s death de-
prived her of her constitutional right to present a defense.
Chambers v. Mississippi, 

, 302 (1973).
2                                              No. 02-1896

   An hour after Natalie was dropped off at Edmunds’
house early one morning, Edmunds dialed 911 to report
that Natalie felt limp, appeared to be unconscious, and was
not breathing. Natalie was rushed by paramedics to the
hospital, where she remained in a semicomatose state
until she died that night. There was medical testimony that
death was due to a cerebral hemorrhage probably caused
by violent shaking or hitting within the preceding 24
hours—a period during which the only access to Natalie
except by Edmunds was by Natalie’s parents. Edmunds’
defense was that one or both of the parents, probably the
father, had done the violent shaking or hitting, and she
presented evidence that the father suffered migraine
headaches, that Natalie’s frequent crying and illnesses
had caused him stress, and that he had been alone with
her for 45 minutes the night before her death. The par-
ents testified that they had never shaken or hit Natalie,
and evidence was presented by other witnesses that they
were caring and loving parents. And against Edmunds’
theory was the fact that Edmunds herself testified that
Natalie had appeared to be completely normal when she
was dropped off at Edmunds’ house the morning of her
death. The medical evidence was that it was unlikely that
if she had been shaken so violently the night before as
to induce a cerebral hemorrhage she would have ap-
peared normal the next morning. There was also evi-
dence that Edmunds had once hit another child on the
head with a book. Natalie’s frequent crying, which Ed-
munds speculates had provoked the father to shake her,
may have provoked Edmunds instead.
  The excluded evidence was evidence that would have
been given by three witnesses: (1) the helicopter pilot
who brought Natalie to the hospital, who saw the parents
a few minutes after he arrived walking normally in the
parking lot of the hospital, appearing neither distraught
No. 02-1896                                                 3

nor emotional, and later speaking with an “odd” lack of
panic in their voices; (2) a police officer who talked to
the parents toward evening, and observed that the father
seemed “nervous” and “fidgety”; and (3) a chaplain who
met with the parents twice during the afternoon, and who
thought they displayed “a guarded demeanor, showing
very limited expression of grief,” which was not what
he would have expected in the circumstances; he also
thought that the father had seemed afraid to enter Nata-
lie’s hospital room, and he observed that in the hospital
room the father stood a few feet behind his wife with his
hands in his pockets and then left the room while his
wife remained. The trial judge excluded all this evidence
on the ground that “absent someone who has the expert-
ise to interpret reactions, I don’t think the observations
have any probative value.” No expert testified, and no
scholarly literature was tendered to the judge, concern-
ing the proper interpretation of the parents’ behavior.
  Edmunds raised the constitutional issue throughout the
state court proceedings, but the state courts did not ad-
dress it, treating it instead as a matter purely of state
evidence law. The Antiterrorism and Effective Death
Penalty Act requires us to uphold a state court’s applica-
tion of a rule of constitutional law laid down by the Su-
preme Court (such as the rule of Chambers v. Mississippi)
as long as it is reasonable, 28 U.S.C. § 2254(d)(1); Williams
v. Taylor, 

, 404-05 (2000), but if there is no
application, there is no ruling to defer to. But this is in
general, not in every case. Although the state overstates
the case to say that if a state court finds that the exclusion
of some piece of evidence did not violate the rules of
evidence then a fortiori it could not violate the Constitu-
tion—the state court might as in Chambers itself have ap-
plied an evidentiary rule more restrictive than the Su-
preme Court interprets the Constitution as permitting, see
4                                                  No. 

; Rock v. Arkansas, 

, 62 (1987);
Washington v. Texas, 

, 15-23 (1967)—that reserva-
tion has no application here. The trial judge excluded the
evidence because he thought its probative value was
negligible without a foundation that had not been laid. The
rule thus applied was Wisconsin’s counterpart of Fed. R.
Evid. 403, which no one supposes unconstitutional. If his
ruling was reasonable, there is no basis for deeming it
unconstitutional.
  It may have been incorrect. The judge allowed testi-
mony of Edmunds’ allegedly hysterical demeanor on the
fatal day—so, she asks, why not evidence concerning the
parents’ demeanor?—though it is hard to see how her be-
ing hysterical would, in the circumstances, have indicated
guilt, and the prosecutor did not mention her hysteria in
the closing argument. Evidence concerning witnesses’ de-
meanor, whether on or off the stand, is routinely admit-
ted to establish that a witness is lying, had guilty knowl-
edge, etc. E.g., United States v. Zafiro, 

, 888
(7th Cir. 1991); Rothgeb v. United States, 

, 651
(8th Cir. 1986); Dyer v. MacDougall, 

, 268-69
(2d Cir. 1952) (L. Hand). In United States v. Frappier, 

, 262 (1st Cir. 1986), “testimony that appellant’s behav-
ior at the wake of her deceased husband was emotion-
less and self centered” was admitted as bearing on her
guilt—and that is the type of evidence that Edmunds
wanted to present here. Even closer is the evidence admit-
ted in Commonwealth v. Counterman, 

, 301 (Pa.
1998): “a medical social worker” testified “that when she
met Counterman [the defendant] at the hospital he did
not appear to be grieving over the deaths of his children,
that he expressed concerns about insurance, and that
when she met Mrs. Counterman, she appeared frightened
of her husband.” See also Bagwell v. State, 

,
No. 02-1896                                               5

388-89 (Ga. 1998); State v. Jobe, 

, 417 (Minn.
1992); People v. Snow, 

, 458 (Cal. 1988).
  The validity of the inferences drawn from demeanor
evidence in these settings has been questioned. As Olin
Guy Wellborn III, “Demeanor,” 76 Cornell L. Rev. 1075
(1991), bluntly puts it, summarizing empirical studies:
“According to the empirical evidence, ordinary people
cannot make effective use of demeanor in deciding wheth-
er to believe a witness.” Yet it would be an unreason-
able curtailment of a criminal defendant’s constitutional
right to put on a defense for a judge to forbid the defen-
dant’s lawyer to draw attention to aspects of demeanor
that the lawyer thought undermined an adverse witness’s
testimony. We cannot find a case on the point, but per-
haps only because the suggestion is too outré to have
been litigated. The demeanor evidence at issue in this
case, however, is of a different character. It presupposes
a benchmark consisting of “normal” behavior in the face
of a shocking incident. Is it true that a “normal” (and
innocent) father in the situation of Natalie’s father would
have strode unhesitatingly into Natalie’s hospital room?
Would not have kept his hands in his pocket? Would have
walked at an abnormal pace in the parking lot? Would have
had panic in his voice yet would have been neither nervous
nor fidgety? Maybe so; but these propositions, and the
others necessary to show that one or both parents mani-
fested lack of grief and consciousness of guilt, are not
so obvious that a judge who like Olin Wellborn thought
them devoid of probative value could be thought unrea-
sonable, though again we cannot find a case on the point,
let alone a U.S. Supreme Court case—which cannot how-
ever help Edmunds.
 The trial judge’s assessment of Edmunds’ demeanor
under cross-examination had a somewhat solider basis in
6                                              No. 02-1896

common experience: “[T]he cross-examination of the
defendant [Edmunds] was devastating to the defense. The
transition from a confident, comfortable, organized wit-
ness on direct examination to a halting, uncomfortable,
insecure witness on cross-examination was remarkable. The
tone of voice, the body language (i.e. looking pleadingly
toward her counsel before answering questions), the pace
of the answers changed dramatically from the presum-
ably well-prepared direct to the not totally predictable
cross by a very skilled prosecutor.” Indeed the cross-
examination brought out significant contradictions be-
tween Edmunds’ trial testimony and the statement she
had given the police a few days after Natalie’s death.
For example, she had told the police (according to one of
the officers who had interviewed her) that on the fatal
morning, until shortly before Natalie’s collapse, Natalie
had been “tracking” Edmunds—that is, had been making
eye contact with her, as babies instinctively do. This would
have been unlikely behavior for a severely neurologically
damaged baby, and at trial Edmunds denied that Natalie
had been tracking. Her testimony was further undermined
by such implausible statements as that she had loved
Natalie as much as she loved her own children, even
though Natalie was a fussy and difficult baby who had
been in Edmunds’ care for only 17 days before the child’s
death.
  At the oral argument Edmunds’ able counsel acknowl-
edged to us that if his position were correct, had Natalie’s
father rather than Edmunds been prosecuted for the death
of Natalie the demeanor evidence that Edmunds’ counsel
wanted to present on her behalf would have been admissi-
ble on behalf of the prosecution. What is sauce for the
goose is sauce for the gander. We are dealing with a type
of evidence that, so far as anyone connected with the pres-
ent proceedings seems to know, is of unknown probative
No. 02-1896                                                  7

value, yet likely to be damaging, as much to a defendant
as to a prosecution witness. In these circumstances the
judge did not exceed the bounds of reason when he said
that “absent someone who has the expertise to interpret reac-
tions, I don’t think the observations have any probative
value.”
  Edmunds might have tried to introduce expert evi-
dence that the parents’ demeanor indeed fit a pattern that
reputable research has shown to be indicative of lack of
grief and consciousness of guilt, cf. Krist v. Eli Lilly & Co.,

(7th Cir. 1990), or she might have presented
such evidence to the judge in an effort to persuade him
that the lay testimony that Edmunds wanted to present
would indeed have significant probative value. Cf. United
States v. Hall, 

, 1120 (7th Cir. 1999) (concur-
ring opinion). In the absence of such evidence, the de-
meanor evidence was—or so the trial judge could find
without taking leave of his senses—too speculative to be
admissible.
  Conceivably the chaplain might have been qualified as
an expert witness to offer an opinion on this matter, cf.
State v. Robinson, 

, 167-68 (Wis. 1988); a
chaplain assigned to a hospital is probably as much an
expert on the characteristic demeanor of persons in the
extremity of grief as a worker at a rape crisis center is
an expert on the characteristic demeanor of victims of sex-
ual assault, as held in the Robinson case. See also United
States v. Parish, 

, 1030 (9th Cir. 2002); United
States v. Bighead, 

, 1330 (9th Cir. 1997). But
that wasn’t done either. Edmunds’ trial lawyer did try to
offer expert opinion testimony, similar to the testimony
offered by the chaplain, by an emergency room nurse
who observed Natalie’s parents. We do not know wheth-
er the trial judge even ruled on the matter. The nurse did
8                                              No. 02-1896

not testify, and Edmunds does not mention the matter in
her appeal.
  In the circumstances, we conclude that the exclusion of
the evidence concerning the parents’ demeanor did not
violate Edmunds’ constitutional rights.
                                                AFFIRMED.

A true Copy:
       Teste:

                         _____________________________
                         Clerk of the United States Court of
                           Appeals for the Seventh Circuit




                  USCA-02-C-0072—12-18-02