United States v. Mitchell, John ( 2003 )

In the
United States Court of Appeals
For the Seventh Circuit
____________
No. 02-3562
UNITED STATES OF AMERICA,
Plaintiff-Appellee,
v.
JOHN MITCHELL,
Defendant-Appellant.
____________
Appeal from the United States District Court
for the Northern District of Illinois, Eastern Division.
No. 01 CR 1121—William T. Hart, Judge.
____________
ARGUED APRIL 4, 2003—DECIDED DECEMBER 23, 2003
____________
Before POSNER, EASTERBROOK, and ROVNER, Circuit
Judges.
ROVNER, Circuit Judge. The Internet has opened the
doors for many to transact business and personal affairs
with almost complete anonymity. For fifty-year-old John
Mitchell, it allowed him to initiate a relationship with
fourteen-year-old Dena Hugh. After two weeks of communi-
cating with Dena about a variety of topics, but mostly about
sex, he arranged to drive from Indiana to Illinois to meet
her at a hotel near her home for the purpose of engaging in
sexual activity. But the anonymity of the Internet works in
both directions, and unfortunately for Mitchell, “Dena” was
actually an undercover Cook County Sheriff’s Detective
2 No. 02-3562
posing as a fourteen-year-old girl. Mitchell was arrested at
the Illinois hotel and pled guilty to traveling in interstate
commerce with the intent to engage in a sexual act with an
undercover agent whom he believed to be a fourteen-year-
old girl. During sentencing the district court increased his
offense level by two based on the United States Sentencing
Commission Guidelines Manual (U.S.S.G.) § 2A3.2(b)(2)(B)
(2001)1 which provides for a two-level enhancement where
the defendant unduly influenced a minor under the age of
sixteen to engage in prohibited sexual conduct. Mitchell
argues that this enhancement cannot apply when the victim
is an imaginary teenager and where no sexual conduct has
occurred. Because we agree that the plain language of the
sentencing guideline cannot apply in the case of an attempt
where the victim is an undercover police officer, we reverse
and remand for re-sentencing.
I.
At the time of the crime, according to Mitchell, he was
down on his luck. His marriage had failed, his appraisal
business was failing, his stepfather and business part-
ner had recently died, and his ex-wife had just married
a former neighbor. As a result of these facts, he claims, this
otherwise upstanding veteran, Lion’s Club President,
Habitat for Humanity volunteer, and father of three turned
to Internet sites where older men can meet young women,
and in many cases too young girls. On December 3, 2001, he
entered a chat room2 entitled “I love Older Men!!:2,” using
1
All references to the United States Sentencing Guidelines are
to the 2001 version in effect at the time of Mitchell’s sentencing.
2
A chat room is a place on the world wide web where Internet
users with common interests can sign on to communicate in real
(continued...)
No. 02-3562 3
the screen name3 “hoosiermale50.” Cook County Sheriff’s
Detective William Plahm, posing as “Dena” was also in the
room, having signed on with the screen name “ilgirl4u.”
Mitchell initiated a conversation with Dena and then the
two exited the chat room to begin a private conversation
using Yahoo!Messenger, a service that allows parties to
send private messages to each other in real time. Dena
informed Mitchell that she was a fourteen-year-old girl
living near Chicago, Illinois.
Mitchell and Dena had several private instant messenger
conversations on the Internet and exchanged a few e-mails
over the course of eleven days. They conversed about
Mitchell’s job, children, divorce, old relationships and about
Dena’s school, parents, and softball. But mostly they wrote
about sex. Mitchell appears to have initiated the topic of
conversation by asking Dena about her physical appear-
ance, whether she had had sex yet, and whether she was
interested in older men. He then proceeded to “educate” her
about things she needed to know about sex. He lectured her
on losing her virginity, performing and receiving oral sex,
masturbation, being naked with another person, and
pleasing men. He also talked reassuringly about sex, telling
her, “I would never force you to do anything,” (R. at 23, ex.
A, p.9); “you have a lot of fun ahead of you,” (R. at 23, ex. A,
p.10); “I like to do whatever my lover enjoys—we try
2
(...continued)
time. Generally, when users enter a chat room they see a list of
other persons (usually known by pseudonyms) who have also
signed on to the chat room site. To “chat,” users type a message
which can be seen almost immediately by all of the other persons
“present” in the chat room. They may, in turn, respond.
3
A screen name is an appellation used to identify oneself in a
chat room or when sending instant messages to another computer
user. Although it can be the user’s real name, it is more often a
pseudonym.
4 No. 02-3562
something and you don’t like it we stop and do something
else,”4 (R. at 23, ex. A, p.12); “just don’t be scared—I will be
gentle and we will get used to each other” (R. at 23, ex. A,
p.57). He also reassured her that he had had a vasectomy
and could not get her pregnant. (R. at 23, ex. A, p.11).
On the other side of the screen, the Sheriff’s Deputy, as
“Dena,” expressed interest in Mitchell stating, “we like each
other lots, I think,” (R. at 23, ex. A, p.37); “I think ev-
erything [we do] will b cool,” (R. at 23, ex. A, p.39); “U will
teach me stuff the rite way rite?” (R. at 23, ex. A, p.29);
“wow its sounds so way awesum john,” (R. at 23, ex. A,
p.11); “Ok it sounds way fun to do,” (R. at 23, ex. A, p.12);
and responding to many of Mitchell’s statements with a
reciprocal “I want u to,” (R. at 23, ex. A, p.40) or “I want to
do same u.” (R. at 23, ex. A, p.78).
After a few rounds of chatting, Mitchell and Dena made
plans to meet at a hotel near Dena’s house in Hillside,
Illinois. Dena e-mailed Mitchell with information regarding
hotels located near her house. According to the plan to
which the two had agreed, Mitchell left his home in
Elkhart, Indiana on December 15, 2001, and drove to the
pre-arranged meeting spot in the parking lot of a Holiday
Inn in Hillside. Once there, he called Dena to let her know
that he had arrived. He told Dena that he would probably
get a room, but he entered the lobby of the Holiday Inn and
then exited without having booked the room. Shortly
thereafter, a Sheriff’s Deputy posing as Dena approached
Mitchell and he was arrested.
4
Real time conversations in Internet chat rooms and through
instant messenger services are most often informal and involve
typographical errors, shorthand, symbols and abbreviations. For
this reason we have not used “[sic]” to indicate each error or
mistake in the original text of the messages.
No. 02-3562 5
Mitchell pled guilty, admitting that he traveled in inter-
state commerce with the intent to engage in a prohibited
sexual act with an undercover agent whom he believed to be
a fourteen-year-old girl, in violation of 18 U.S.C. § 2423(b).
In the plea agreement, both parties agreed that U.S.S.G. §
2A3.2 entitled “Criminal Sexual Abuse of a Minor Under
the Age of Sixteen Years (Statutory Rape) or Attempt to
Commit Such Acts” applied. The parties also agreed that
the base offense level could be increased by two levels under
subsection (b)(3) because Mitchell had used a computer and
an Internet access device during the commission of the
crime. This was the only enhancement mentioned in the
plea agreement. Nevertheless, the probation officer who
drafted the pre-sentence investigation report applied a two-
level enhancement to the defendant’s sentence based on a
conclusion that Mitchell had unduly influenced his victim.
U.S.S.G. § 2A3.2(b)(2)(B). Although the government had not
sought this enhancement in the plea agreement, it sup-
ported the pre-sentence investigation report’s use of the
enhancement. Over Mitchell’s objections, the district court
applied the two-level enhancement for behavior which
“unduly influenced the victim to engage in prohibited
sexual conduct.” Mitchell appeals.
II.
Mitchell objects to the two-level sentencing enhancement
imposed by the court pursuant to U.S.S.G. § 2A3.2(b) (2)(B).
This enhancement—applied to persons convicted of criminal
sexual abuse of a minor under the age of sixteen years or
attempts to commit such acts—increases a participant’s
sentencing level by two where, “a participant otherwise
unduly influenced the victim to engage in prohibited sexual
conduct.” Id. The commentary to the guideline states,
In determining whether subsection (b)(2)(B) applies, the
court should closely consider the facts of the case to
6 No. 02-3562
determine whether a participant’s influence over the
victim compromised the voluntariness of the victim’s
behavior.
In a case in which a participant is at least 10 years
older than the victim, there shall be a rebuttable pre-
sumption, for purposes of subsection (b)(2)(B), that such
participant unduly influenced the victim to engage in
prohibited sexual conduct. In such a case, some degree
of undue influence can be presumed because of the
substantial difference in age between the participant
and the victim.
U.S.S.G. § 2A3.2, cmt. n.4. Because Mitchell was more than
ten years older than the fictitious Dena, the court applied
the presumption that he had unduly influenced his victim
and found that he had not successfully rebutted the pre-
sumption.
Mitchell argues first that the enhancement applies only
when a participant successfully influences a victim to
engage in prohibited sexual contact—not in the case of an
attempt, and second, that because the enhancement re-
quires a court to consider the voluntariness of the victim’s
behavior, it cannot apply in the case of a sting operation
where there is no real victim.
We review the district court’s interpretation of sentencing
guidelines de novo. United States v. Smith, 332 F.3d 455,
457-58 (7th Cir. 2003). When construing federal sentencing
guidelines, we turn to the general rules of statutory con-
struction. United States v. Lewis, 93 F.3d 1075, 1080 (2d
Cir. 1996). Accordingly, we must begin by looking at the
plain language of the guideline. United States v. Twieg, 238
F.3d 930, 931 (7th Cir. 2001). We treat the commentary to
the guideline as authoritative as well. Stinson v. United
States, 508 U.S. 36, 38 (1993).
The guideline itself supports Mitchell’s argument
about the inapplicability of the subsection to attempts. The
No. 02-3562 7
guideline requires the two level enhancement where “a
participant otherwise unduly influenced the victim to
engage in prohibited sexual conduct.” U.S.S.G. § 2A3.2(b)
(2)(B) (emphasis supplied). The government, using brackets
and ellipses to alter the tense of the guideline language,
states that the plain language only requires “undu[e]
influence[ ] . . . to engage in prohibited sexual conduct.”
(Brief of the United States at 10). But the government’s
editorial contortions merely highlight the problem—the
language of the enhancement is in the past tense; an honest
reading of the plain language of the guideline would
indicate that it could not apply where the participant had
either failed in his attempt to influence the victim or where
the two otherwise had not engaged in prohibited sexual
conduct. See United States v. Chastain, 198 F.3d 1338, 1353
(11th Cir. 1999) (“the plain language of the [sentencing]
guideline that uses the past tense . . . cannot be ignored . .
. [and] clearly contemplates a completed event”); United
States v. Lacy, 119 F.3d 742, 749 (9th Cir. 1997) (finding
that the government’s attempt to alter the verb tense was
significant and changed the meaning of the sentencing
guideline). Similarly, the commentary instructs courts to
closely consider the particular facts of the case to determine
whether a “participant’s influence over the victim compro-
mised the voluntariness of the victim’s behavior.” U.S.S.G.
§ 2A3.2, cmt. n.4 (emphasis supplied); see also U.S.S.G. §
2A3.2, cmt. background (noting that the enhancement
applies where “the voluntariness of the victim’s behavior
was compromised”). Again, the past tense of the verb
“compromised” indicates that the participant must have
succeeded in compromising the voluntariness of the victim’s
behavior. In both instances—the guideline itself and the
commentary—the offender must have succeeded in influenc-
ing or compromising. In other words, the enhancement
cannot apply where the offender and victim have not
engaged in illicit sexual conduct.
8 No. 02-3562
The only way to make the language applicable in the case
of an attempt is to use a grammatical shoehorn and rewrite
the guideline and its commentary in the present tense. But
no matter what the policy reason for doing so, a court may
not rewrite a statute or guideline to suit its or any other
needs. Artuz v. Bennett, 531 U.S. 4, 10 (2000) (“Whatever
merits these and other policy arguments may have, it is not
the province of this Court to rewrite the statute to accom-
modate them.”); Carter v. Litscher, 275 F.3d 663, 665 (7th
Cir. 2001) (same). Nor can we rewrite guidelines based on
speculation as to what the Sentencing Commission may
have intended the statute to say. United States v. Joseph,
50 F.3d 401, 403 (7th Cir. 1995) (a court may not re-write
sentencing guidelines). See also New England Power Co. v.
New Hampshire, 455 U.S. 331, 343 (1982) (“we have no
authority to rewrite . . . legislation based on mere specula-
tion as to what Congress ‘probably had in mind.’ ”). If the
Sentencing Commission intended a different result it must
rewrite the guideline accordingly.
In addition to the grammatical construction of the guide-
line and commentary, our conclusion is further bolstered by
the context of the enhancement and commentary. Although
most sentencing guidelines and their enhancements are
written to focus on the conduct of the defendant, this
enhancement, instead, focuses on the behavior and charac-
teristics of the victim. Certainly the Sentencing Commission
could have written the enhancement to focus solely on the
defendant’s actions. Instead, the Sentencing Commission
chose to look at the voluntariness of the victim’s behavior in
determining undue influence. U.S.S.G. § 2A3.2, cmt. n.4.
For this reason, the dissent’s discussion of the past tense of
other enhancements is irrelevant. None of these other
enhancements is written to focus on the effect that the
offending behavior has on the victim.
The guidelines themselves offer no definition of undue
influence, but there are myriad definitions of “undue in-
fluence” in the civil context to which we may look for guid-
No. 02-3562 9
ance.5 In contracts, wills, and trusts, for example, undue
influence has been defined as “any improper urgency of
persuasion whereby the will of a person is overpowered and
he is induced to do or forbear an act which he would not do
or would do if left to act freely.” Franciscan Sisters Health
Care Corp. v. Dean, 448 N.E.2d 872, 875 (Ill. 1983) (internal
citation omitted); see also In re Estate of Hoover, 615 N.E.2d
736, 740 (Il. 1993) (“[U]ndue influence may be inferred in
cases where the power of another has been so exercised
upon the mind of the testator as to have induced him to
make a devise or confer a benefit contrary to his deliberate
judgment and reason.”); Rice v. Office of Servicemembers’
Group Life Ins., 260 F.3d 1240, 1250 (10th Cir. 2001)
(“[U]ndue influence is generally defined as influence that is
sufficient to overpower volition, destroy free agency, and
impel the grantor to act against the grantor’s inclination
and free will.”); Lyle v. Bentley, 406 F.2d 325, 328 (5th Cir.
1969) (“[U]ndue influence is such influence or dominion as
exercised at the time, under the facts and circumstances of
the case, which destroys the free agency of the testator, and
substitutes in the place the will of another”) (internal
citation omitted); John Calamari & Jospeh M. Perillo, The
Law of Contracts § 9.9 (4th ed. 1998) (“undue influence” is
“any improper or wrongful constraint, machination, or
urgency of persuasion, whereby the will of a person is
overpowered, and he is induced to do or forbear an act
5
One scholar has argued for the use of the “undue influence” test
in the criminal context to determine whether someone has
committed the crime of causing another to commit suicide. Sue
Woolf Brenner, Undue Influence in the Criminal Law: A Proposed
Analysis of the Criminal Offense of Causing Suicide, 47 Alb. L.
Rev. 62 (1982). She suggests that under her theory the inquiry
would have to consider whether or not the perpetrator actually
influenced the victim to an act of physical destruction. Id. at 93.
The actual completion of the act would, of course, be necessary
under this proposed criminal law scheme.
10 No. 02-3562
which he would not do, or would do if left to act freely.”)
(internal citations omitted). 25 Am. Jur. 2d Duress and
Undue Influence § 31 (2002) (describing the four elements
of undue influence as requiring: “(1) a person who is
susceptible to influence; (2) another person who had an
opportunity to exert undue influence; (3) the exertion of
improper influence; and (4) the production of the desired
effect as a result of the influence.”) (emphasis supplied).
Each of these definitions requires an actual target of
influence and contemplates a situation where the
“influencer” has succeeded in altering the behavior of the
target.
Furthermore, the term “influence” especially, but not only
in the past denotes causal efficacy. To say that Mitchell
“influenced Dena to have sex” implies that the two did,
indeed, have sex. This is what the dictionary definitions
(both legal and lay) instruct: that there can be no “influ-
ence” where the object of the influence has not acted ac-
cordingly. The Oxford English Dictionary defines influence
as “[t]o affect the mind or action of; to move or induce by
influence; sometimes esp. to move by improper or undue
influence.” Oxford English Dictionary (2d ed. 1989). The
word influence, therefore, in and of itself requires that the
target of the influence act accordingly. Likewise, Black’s
Law Dictionary contains several paragraphs on “undue
influence” each of which defines the term, in part, based
on the effect of the influence on the target. For example, it
describes undue influence as influence which “so overpow-
ers the dominated party’s free will or judgment that he
or she cannot act intelligently or voluntarily, but acts, in-
stead, subject to the will or purposes of the dominating
party,” or “urgency of persuasion whereby the will of a
person is overpowered and he is induced to do or forbear an
act which he would not do or would do if left to act freely,”
or that which causes one “to do what he would not other-
wise have done but for such dominion and control.” Black’s
No. 02-3562 11
Law Dictionary 1528 (6th ed. 1990). It is this understanding
of the term “influence,” along with the past tense construc-
tion of the enhancement, the reference to engaging in illicit
sexual conduct, and the focus on the victim rather than the
offender that allows us to conclude that the enhancement is
not applicable to attempts in a case such as this one.6 The
dissent makes much of the situation in which a real victim
is rescued from the hands of a sexual predator just in the
nick of time. These examples are undoubtedly compelling,
but are not particularly helpful in this case where Dena was
not a real victim but an undercover police officer. Our task
is to interpret the guidelines as they are written and then
apply them to the facts at hand. Although our reading of
the guideline points toward a conclusion that a sexual act
must have occurred in all cases, we need only decide today
that the enhancement cannot apply in the case where no
sexual act has occurred and could not have occurred
because the “victim” was not real.
Mitchell’s second argument, that the guideline cannot
apply in the case of a sting operation, collapses somewhat
with the first argument, for in a case where there is no real
victim but only an undercover police officer, there will never
be completed action on the part of the victim. But here, we
must meet head on the guideline’s definition of victim as:
(A) an individual who, except as provided in subdivision
(B), had not attained the age of 16 years; or (B) an
undercover law enforcement officer who represented to
a participant that the officer had not attained the age
of 16 years.
U.S.S.G. § 2A3.2, cmt. n.1.
6
The dissent incorrectly implies that our opinion relies primarily
on the past tense of the verb “influence” as support for the
proposition that the enhancement does not apply to sting oper-
ations. This is not so.
12 No. 02-3562
The government argues that not only does the definition
specifically include undercover law enforcement officers, but
that the application of the enhancement in sting operations
jibes with the underlying purpose of the expanded defini-
tion of “victim” which ensures that offenders who are
apprehended in undercover sting operations are appropri-
ately punished.
This argument does not get the government far. Even
were we to consider the undercover agent as a victim for
purposes of this enhancement, we still must follow the
instructions of the guideline to see whether “Dena” was
unduly influenced to engage in prohibited sexual conduct,
and we are back where we began: looking to see whether
Mitchell unduly influenced Dena to engage in illicit sexual
conduct. Where no prohibited sexual conduct has occurred,
there has been no undue influence (after all, a real Dena
may not have been influenced at all, and may not have
appeared at the hotel).
As for the government’s argument that application of the
enhancement ensures that those apprehended in sting op-
erations are appropriately punished, we emphasize that by
refusing to apply this enhancement to undercover probes,
we are not ignoring Congress’ intent to punish offenders
who are caught in sting operations, nor are we allowing sex
offenders to walk away scot free. Those offenders will
receive a substantial sentence pursuant to this guideline
which sets the base offense level at twenty-one for at-
tempted criminal sexual abuse of a minor under the age
of sixteen (the base offense level for actual commission
of the act is only three levels higher, at twenty-four).
U.S.S.G. § 2A3.2(a). A base offense level of twenty-one
translates into a sentencing range of 37-46 months. In
addition, a court may increase the base offense level under
other enhancement provisions of the guideline. For exam-
ple, Mitchell does not dispute the application of a two-level
No. 02-3562 13
enhancement for the use of a computer or Internet access
device under U.S.S.G. § 2A3.2(b)(3). Likewise, his sentence
could have been enhanced by two levels had he knowingly
misrepresented his identity during the course of the crime.
U.S.S.G. § 2A3.2(b)(2)(A). Each of these enhancements is
written to focus on the conduct of the defendant and not the
victim, and each is applicable in both the case of sting
operation and in the case of an attempt. A court may not
contort the plain language of an enhancement to make it
applicable to a sting operation simply because it believes
that such an application would be consistent with the pur-
pose of the guideline.
The dissent argues that it would somehow be a “boon” to
Mitchell were he not to receive the two-level enhancement
for undue influence. For some reason the dissent thinks
that Mitchell already received his “lucky break” by being
sentenced for an attempt rather than engaging in actual
illicit sexual conduct and therefore he should not receive an
additional break under section (b)(2)(B). The dissent paints
a horrifying picture of the offender just moments away from
the illicit act, interrupted purely fortuitously by the police,
parents, a snowstorm, or a miscommunication about the
rendezvous spot. We are meant to ask ourselves why such
an offender would be spared the maximum possible punish-
ment merely because of an unforeseen interruption. But
this is what the guidelines do all the time— they punish
offenders less severely for attempts than for completed acts
even though the act would have been completed but for the
interruption. Furthermore, the dissent mixes apples
(attempts versus completed action) with oranges (undue
influence versus no undue influence). The former is ad-
dressed by section (a) of the guideline which assigns
different base offense levels for attempts versus completed
conduct. The latter is, of course, addressed by the enhance-
ment at issue in this case, section (b)(2)(B). In any case, as
14 No. 02-3562
we noted before, this is not a case where we need to decide
how the enhancement applies in the case where the defen-
dant is interrupted before the illegal act with a real victim
could be completed.
For the skeptic who insists on looking past the plain lan-
guage to the intent of the Sentencing Commission (which,
for the reasons discussed above, we believe we should not
do), we need look no further than the rebuttable presump-
tion of the commentary. Recall that the guidelines create a
rebuttable presumption that the participant unduly influ-
enced the victim if he is at least ten years older than the
victim. U.S.S.G. § 2A3.2 cmt. n.4. The commentary in-
structs that a court should closely consider the facts of the
case “to determine whether a participant’s influence over
the victim compromised the voluntariness of the victim’s
behavior.” Id. But how would a defendant rebut the pre-
sumption in the case where the “victim” is an undercover
agent? What facts would the court consider? All parties
agree that it would be nonsensical to look to see whether
the participant compromised the voluntariness of the
undercover agent’s behavior. We must, therefore, look at
the specific characteristics of the imaginary minor. If this
is the case, then it will be virtually impossible for the
defendant to rebut the presumption of undue influence. The
government controls every fact of the imaginary minor from
her age to her mental state. Of course the government will
always ensure that the imaginary victim is more than ten
years younger than the offender, and that his or her will
can be overcome readily by the offender’s influences. If the
Sentencing Commission intended to allow a defendant to
rebut the presumption of undue influence, it cannot have
meant to apply the presumption in the case of a sting
operation where the government can manipulate the
characteristics and actions of the victim to create undue
influence in every single case. If we were to follow the
No. 02-3562 15
reasoning of the dissent, there would never be a case
involving a sting operation in which the enhancement does
not apply.7
Indeed, as the dissent points out, there are many rebut-
table presumptions within the guidelines that are difficult
to rebut, for example, the presumption that weapons found
near drugs are connected to the offense. See, e.g., U.S.S.G.
2D1.1(b)(1); United States v. Johnson, 289 F.3d 1034, 1041-
42 (7th Cir. 2002). But these presumptions do not involve
factual scenarios that are created and manipulated by the
police. In these Internet sting operations, the police create
the victim, decide what characteristics he or she will have,
and what actions he or she will take. The guidelines then
instruct a court to look at those characteristics and
actions—the very same ones created by the police—to
determine whether a two-level enhancement applies. The
Sentencing Commission surely cannot have contemplated
that the rebuttable presumption can be made irrebuttable
by the manipulations of the government.
The absurdity of this result is highlighted by the dissent’s
comment that “if the sort of things [Mitchell] wrote would
have unduly influenced a real 14-year-old girl, then the
enhancement applies.” How would anyone know if Mitch-
ell’s messages would have unduly influenced a real
fourteen-year-old girl, and how could Mitchell rebut the
presumption that his messages would have influenced her?
7
The dissent implies that Mitchell might have rebutted the pre-
sumption of undue influence by pointing out that Dena used the
screen name “ilgirl4u” and that they met in a chat room entitled
“I Love Older Men.” Our colleague cannot possibly believe that
such innocuous details would be sufficient to rebut a presumption
of undue influence. Even if they were, however, the police could
simply manipulate these details as well to make certain that the
victim was the ideal candidate to be unduly influenced.
16 No. 02-3562
Of course there is no prototypical fourteen-year-old girl to
whom we might refer for the answers. We can only know if
a real fourteen-year-old girl would be influenced if we, in
fact, have a real fourteen-year-old girl on the receiving end
of the influence. Perhaps if Dena had previously had many
affairs with older men or had been involved in the sex in-
dustry he might be able to rebut the presumption, but no
police officer would ever create a fictional victim with such
a profile. The dissent’s argument merely highlights the
need to look at the response of an actual fourteen-year-old
victim.
The Eleventh Circuit, in United States v. Root, 296 F.3d
1222 (11th Cir. 2002), cert. denied, 123 S. Ct. 1006 (2003),
considered the identical enhancement in a case with similar
facts,8 and a majority of the panel determined that the
enhancement could, indeed, be applied in the case of a sting
operation. The majority decision in Root, however, ignored
the plain meaning of “unduly influenced” and “was compro-
mised,” and ignored the clear language of the commentary
requiring a court to closely consider the voluntariness of the
victim’s behavior. We decline to follow the holding in Root.9
The Root court first explained that the enhancement
8
In Root, the defendant Root met an undercover Sheriff ’s
Department investigator posing as thirteen-year-old “Jenny” in
the chat room “I Love Older Men.” Root and Jenny chatted about
sex over the course of three days. Root then traveled from North
Carolina to meet the fictional “Jenny” at a mall in Georgia for the
purpose of engaging in prohibited sexual conduct.
9
Because this opinion will create a conflict among the Circuit
Courts, it has been circulated to all of judges of this court in reg-
ular active service pursuant to Circuit Rule 40(e). A majority of
judges in active service voted not to rehear this case en banc.
(Judges Easterbrook and Kanne voted to rehear this case en
banc.)
No. 02-3562 17
under § 2A3.2(b)(2)(B) can apply in the case of an attempt.
As discussed earlier, we disagree with this proposition. The
plain language of the statute simply does not apply where
no sexual act has occurred. The Root court used the same
grammatical machinations as the government in this case
to get to its result. Root, 122 F.3d at 1234 (“The offender
need only have exerted undue influence aimed at convincing
the victim ‘to engage’ in future improper sexual activities.”).
Confounding the issue, from this conclusion, (that the
enhancement applies in the case of attempts), the court
concluded that “the identity of the victim of an attempt
conviction is irrelevant for purposes of a sentence enhance-
ment under § 2A3.2(b)(2)(B).” Id. The Root majority,
however, provided no explanation for why this would be
true. In fact, as discussed above, the identity of the victim
is always relevant because the guideline requires it to be so:
“In determining whether subsection (b)(2)(B) applies, the
court should closely consider the facts of the case to deter-
mine whether a participant’s influence over the victim
compromised the voluntariness of the victim’s behavior.”
U.S.S.G. § 2A3.2, cmt. n.4. The Root majority offered no
explanation for contravening this unambiguous and direct
instruction from the Sentencing Commission and focusing
instead on the conduct of the offender. Although it is true
that generally the sentencing guidelines focus on the
conduct of the defendant in doling out punishment, this
enhancement was specifically written to require an exami-
nation of the effect on the victim. See Root, 296 F.3d at 1237(Kennedy, J., concurring in part, and dissenting in part).
Both the Root majority and the government agree that
“the voluntariness of a real child victim’s actions would be
dispositive if an undercover agent were not involved.” Root,
18 No. 02-3562
296 F.3d at 1234.10 But if the purpose of the enhancement
is to mete out punishment based on the defendant’s con-
duct, as the government argues, why would the voluntar-
iness of the victim’s behavior ever be dispositive? Clearly
the state of mind of the victim is critical in the inquiry. It is
critical because the guidelines tell us it must be so. Where
the state of mind of the victim is critical, and perhaps
dispositive, it simply cannot apply in the case where the
victim has no state of mind whatsoever because she does
not exist. The government and the Root court’s equivocation
on this point exposes the fallacy of the claim that only the
defendant’s conduct matters.
Finally, the government argues that it would be unfair to
treat two defendants differently simply because one hap-
pened to victimize a mature fourteen-year-old girl who
made a truly voluntary decision and another victimized
a naive fourteen-year-old who was particularly suscept-
ible to influence. The law, however, frequently punishes
offenders differently based on the individual characteristics
of the victim. The criminal offender takes his victim as he
finds him. United States v. Feola, 420 U.S. 671, 685 (1975)
(drug dealing defendants who believed they were robbing
and assaulting drug buyers but actually assaulted federal
narcotics agents could be convicted of the crime of assault-
10
At oral argument the government agreed with this proposition:
The Court: In Root, the majority specifically stated that the
voluntariness of a real child victim’s actions would be dispo-
sitive if an undercover agent were not involved.
Government: And that is correct your honor. And that’s
right, because the commentary to the Guidelines indicates
that.
No. 02-3562 19
ing a federal officer);11 Brackett v. Peters, 11 F.3d 78, 81 (7th
Cir. 1993) (“in criminal law as in tort law, the injurer takes
his victim as he finds him”). In Brackett, the court upheld
the murder conviction of a man who raped an eighty-five-
year old widow. Id. at 82. As a result of the rape, the
widow’s physical and psychological health deteriorated to
the point where she had to be fed via syringe. She died a
month after the rape when food from the syringe became
lodged in her trachea, asphyxiating her. Id. at 79. This
court upheld the rapists’ conviction for murder despite the
fact that her age and fragile condition contributed, in part,
to her death. In short, criminal law does indeed recognize
the impact of individual characteristics of victims in
assessing the severity of the punishment. The relevant
guideline in this matter instructs the sentencing judge to do
just that—to closely consider the facts of the case to
determine the extent of the voluntariness of the victim’s
behavior. U.S.S.G. § 2A3.2, cmt. n.4. For these reasons, we
do not believe it unfair, in light of the guideline language,
to consider the particular characteristics of the victim
rather than focus solely on the defendant’s conduct in
deciding whether a particular enhancement applies. The
result may be that an offender who attempts to have sex
with a mature fourteen-year-old who is not unduly influ-
enced will be punished slightly less severely than one who
attempts to have sex with a naive fourteen-year-old. But
this result is not necessarily unfair. Persons who commit
11
Relying on Feola, the Seventh Circuit has explained that in
some cases the defendant’s state of knowledge about the identity
or status of a victim may be required before the requisite intent
can be found as in the case “where legitimate conduct becomes
unlawful solely because of the identity of the individual or agency
affected.” United States v. Hillsman, 522 F.2d 454, 460 (7th Cir.
1975) (citing Feola, 420 U.S. at 685). This is not such a case. In
the case of this enhancement, the offender knows that his conduct
is illegal whether he chooses a mature or immature victim.
20 No. 02-3562
crimes take their chances that the harm to the victim will
be worse than they had intended. In this case they take
their chances that the sexual conduct will look to a sentenc-
ing court to be less consensual and more forced based on the
characteristics of the victim.12
Finally, even were we to decide that the enhancement
for “undue influence” can apply to sting operations, the
district court in this case failed to make the necessary
factual findings. See United States v. Goines, 988 F.2d 750,
775 (7th Cir. 1993) (“the district court must be precise in
explaining the basis for specific factual findings, including
any adjustments increasing or decreasing the base offense
level and criminal history category”). In determining
whether the enhancement applies, the court must closely
consider the facts of the particular case to determine
whether a participant’s influence over the victim compro-
mised the voluntariness of the victim’s behavior. U.S.S.G.
§ 2A3.2, cmt. n.4. The majority in Root, for example, iden-
tified specific language from the Internet conversations
which demonstrated how the defendant had used his in-
creased knowledge, persuasive powers, and superior re-
sources to unduly influence his victim. Root, 296 F.3d at
1235-36. No such factual findings were made here. In its
brief on appeal, the government appears to argue that a col-
loquy between Mitchell’s counsel and the district court
judge regarding the similarities between this case and the
12
The dissent’s comparison to the vulnerable victim enhancement
in this instance is not particularly helpful. The vulnerable victim
enhancement, unlike the one at issue here, focuses on the conduct
of the defendant—punishing those who knowingly choose their
victims specifically because they possess a particular vulnerabil-
ity. As we describe supra, for purposes of this enhancement, the
defendant takes his victim as he finds her and receives the
enhancement based on conduct and characteristics of his victim.
No. 02-3562 21
Root case constituted findings of fact by the court.13 It does
not. The government conceded so during oral argument.
Although the facts are similar (indeed, many of these
Internet sting cases involve similar factual scenarios, see
United States v. Raney, 342 F.3d 551 (7th Cir. 2003)), there
are significant differences. A judge may not import the
findings of fact from one case into a different case where
there are different defendants, different victims, and of
course, different conversations between the offender and the
victim. The dissent argues that the burden should be on the
defendant to rebut the presumption of undue influence. We
can leave that question for another day, for irrespective of
where the burden lies, the district court must make a
factual finding explaining the basis of the enhancement. It
failed to do so.
Even if there were a circumstance where undue influence
could be determined merely by looking at the perpetrator’s
behavior without regard to the victim’s response, this could
not be such a case. The district court never made any
findings that Mitchell’s words or actions were so influential
as to unduly influence any victim—regardless of her
individual characteristics.
III.
13
During sentencing the following exchange occurred:
Court: The majority opinion in Root is directly on point here.
Ms. Milella: It is direct—it is the same factual situation,
your honor.
Court: Yes.
Ms. Milella: Yes; we disagree with the Eleventh Circuit’s
holding.
R. at 30, p. 19.
22 No. 02-3562
For the reasons stated above, we reverse and remand for
resentencing.
REVERSED AND REMANDED.
EASTERBROOK, Circuit Judge, dissenting. When calcu-
lating the sentence for a person convicted of sexually abus-
ing a minor, the district judge must apply the following
rule: if “a participant . . . unduly influenced the victim to
engage in prohibited sexual conduct”, then add two levels.
U.S.S.G. §2A3.2(b)(2)(B). Mitchell contends that this rule
does not apply to sting operations, for agents masquerading
as underage girls cannot be “unduly influenced” and do not
“engage in prohibited sexual conduct”. But §2A3.2 as a
whole applies to both attempts and completed offenses, so
lack of success is not dispositive. To determine the proper
scope of this enhancement, one has only to consult Applica-
tion Note 1, which defines “victim” to mean “(A) an indi-
vidual who, except as provided in subdivision (B), had not
attained the age of 16 years; or (B) an undercover law
enforcement officer who represented to a participant that
the officer had not attained the age of 16 years.” Plugging
the definition into the guideline produces this rule: if “a
participant . . . unduly influenced a law enforcement agent,
posing as someone under 16 years old, to engage in pro-
hibited sexual conduct”, then add two levels. This requires
us to agree with United States v. Root, 296 F.3d 1222 (11th
Cir. 2002), that subsection (b)(2)(B) applies in sting oper-
ations, and to affirm the judgment of the district court.
No. 02-3562 23
The principal contrary argument in the majority’s opinion
depends on the word “influenced”. It is in the past tense,
which means, my colleagues believe, that the influence
must have succeeded—which is impossible if the object of
the defendant’s wiles is an agent. This is wrong linguisti-
cally: it may mean that influence lies in the past, but it does
not mean that sexual relations occurred in the past. The
latter portion of subsection (b)(2)(B) specifies an object but
not a time: “to engage in prohibited sexual relations”, not
“so that the victim engaged in prohibited sexual relations”
or an equivalent phrase. To the extent there is doubt, we
should read “influenced” to be compatible with the defini-
tion of “victim” rather than declare the two at loggerheads
and chuck the definition of “victim” into the waste basket.
Actually the premise of this reliance on tense is wrong.
The majority treats “influenced” as the only verb in §2A3.2
that looks backward. If this is so, then tense may be infor-
mative. But it is not so. Here is the complete list of specific
offense characteristics in §2A3.2(b):
(1) If the victim was in the custody, care, or supervisory
control of the defendant, increase by 2 levels.
(2) If subsection (b)(1) does not apply; and—
(A) the offense involved the knowing misrepre-
sentation of a participant’s identity to (i) persuade,
induce, entice, or coerce the victim to engage in pro-
hibited sexual conduct; or (ii) facilitate transporta-
tion or travel, by the victim or a participant, to
engage in prohibited sexual conduct; or
(B) a participant otherwise unduly influenced the
victim to engage in prohibited sexual conduct,
increase by 2 levels.
(3) If a computer or an Internet-access device was used
to (A) persuade, induce, entice, or coerce the victim to
engage in prohibited sexual conduct; or (B) facilitate
24 No. 02-3562
transportation or travel, by the victim or a participant,
to engage in prohibited sexual conduct, increase by 2
levels.
(4) If (A) subsection (a)(1) applies; and (B) none of sub-
sections (b)(1) through (b)(3) applies, decrease by 6
levels.
All of these are in the past tense. The verb in subsection (1)
is “was”; the verb in subsection (2)(A) is “involved”; the verb
in subsection (2)(B) is “influenced”; the verb in subsection
(3) is “was used”. (The similar list of specific offense charac-
teristics in U.S.S.G. §2G1.1(b), which deals with commercial
sex, likewise is in the past tense.) Past tense makes sense
from the perspective of judges (the addressees of the
Guidelines) because, by the time of sentencing, all salient
events lie in the past. Consider subsection (3), which adds
two levels if a computer “was used” to persuade the victim
to engage in sexual conduct. Mitchell received a two-level
increase under this subsection, without protest. Yet subsec-
tions (2)(B) and (3) are parallel: where is the material
difference between “influenced the victim to engage in
prohibited sexual conduct” (subsection (2)(B)) and “used to
. . . induce . . . the victim to engage in prohibited sexual
conduct” (subsection (b)(3))? If the latter applies to Mitchell,
so does the former.
Seeing significance in the tense of “influenced” is the
majority’s novelty. Mitchell’s lawyer made nothing of the
conjugations or inflections of these verbs, so the United
States did not have occasion to respond. Mitchell’s choice
was understandable, not only because the past tense runs
throughout §2A3.2 but also because it is the norm of the
Sentencing Guidelines. Open the Guidelines Manual at
random and read some specific offense characteristics.
You’ll find that verbs are in the past tense. This makes it
impossible to deduce from the past tense of “influence” in
§2A3.2(b)(2)(B) that any particular reading is called for.
No. 02-3562 25
What Mitchell did argue is that a completed sexual act is
essential because the subsection uses the phrase “engage in
prohibited sexual conduct”. This argument (which the
majority does not accept) is wrong not only because of the
definition of “victim” but also because all of §2A3.2 applies
to attempts as well as completed offenses. Many crimes can
be committed, and enhancements applied for aggravating
events, even though success was impossible. Think of
fraudulent statements made to obtain a loan at a bank
whose staff knows the truth, or perjury before a grand jury
armed with accurate information, or attempts to purchase
cocaine from federal agents who have only sugar to sell.
These are real crimes, and the guidelines for each specify
enhancements that apply to those aspects of the conduct
that the offender did complete, even though success in the
venture as a whole was impossible. Given the definition of
“victim” in Application Note 1, §2A3.2 must be included
among these guidelines. Mitchell sent his blandishments to
“Dena” over the Internet; this was completed conduct, and
if the sort of things he wrote would have unduly influenced
a real 14-year-old girl, then the enhancement applies.
Assessing “undue inducement” can be hard even if a sex-
ual act took place. An offender might argue that the victim
was eager for sexual relations and that the inducements
thus were unnecessary, mutually desired, or not so over-
bearing as to be “undue” given the circumstances. So the
Sentencing Commission included a presumption:
In a case in which a participant is at least 10 years
older than the victim, there shall be a rebuttable pre-
sumption, for purposes of subsection (b)(2)(B), that such
participant unduly influenced the victim to engage in
prohibited sexual conduct. In such a case, some degree
of undue influence can be presumed because of the
substantial difference in age between the participant
and the victim.
26 No. 02-3562
U.S.S.G. §2A3.2 Application Note 4 ¶4. Mitchell was 50
years old at the time of the events in question; Dena pre-
tended to be 14. The presumption of undue influence there-
fore kicks in—which makes it a puzzle why the majority
suggests at the end of its opinion that the district judge
neglected to develop the record; given this presumption, the
burden was on Mitchell, not the prosecutor.
Mitchell’s brief takes on Application Note 4, calling the
presumption “irrebuttable” and hence suspect, if not un-
constitutional. Yet there’s nothing wrong with presump-
tions in sentencing; we see lots of sentences affected by
the presumption in U.S.S.G. §2D1.1(b)(1) Application Note
3, that a gun found near drugs was connected to the offense
unless the defendant shows that a link was “clearly improba-
ble.” Challenges to this presumption have been unsuccess-
ful. See, e.g., United States v. Johnson, 289 F.3d 1034, 1041-
42 (7th Cir. 2002); United States v. Bjorkman, 270 F.3d 482,
492 (7th Cir. 2001). It is easier to rebut the age-difference
presumption for sex offenses than the gun presumption for
drug offenses. The court must decide whether the defendant
did or said things that would compromise the voluntariness
of a child’s decision. Here, for example, Mitchell might have
argued that he did not presume on Dena, that she led him
on (her screen name was “ilgirl4u”, indicating predisposi-
tion, and they “met” in a chat room with the title “I LOVE
OLDER MEN!!” that she had joined independently). That Dena
was an agent rather than a youngster does not affect this
process: assessment depends on the words exchanged, not
on the contents of the victim’s head. The presumption might
be rebutted by the language the agent used to describe her
sexual history and desires, by the older person’s failure to
make any effort to overcome real or feigned resistance, and
so on. But Mitchell makes a legal rather than a factual
objection (perhaps because his messages imply that he was
using his age and sexual experience to inveigle Dena and
overcome reluctance).
No. 02-3562 27
And if the presumption turns out to be very hard to de-
feat, why should that be dispositive? The presumption that
a gun found with drugs is used in the crime is so hard to
rebut that it might as well be conclusive. My colleagues
imply that agents may too readily manipulate things to
bring this enhancement into play, but one might as well say
therefore that stings cannot be used in drug cases, because
the agents can manipulate the weight of the drugs they
offer to buy or sell. Yet this circuit has disparaged
the doctrine of “sentencing entrapment,” see, e.g., United
States v. Estrada, 256 F.3d 466, 473-77 (7th Cir. 2002), and
it is no more compelling in a sex case than in a drug case.
See United States v. Richardson, 238 F.3d 837, 839-40 (7th
Cir. 2001) (sentence proper even though agent sent the
defendant sadomasochistic images that he arguably did not
want). Although a defendant who is wheedled or tricked
into committing a more serious version of an offense may
seek a downward departure, this possibility is no reason to
change the way the offense level is determined. Mitchell did
not ask for a downward departure on this ground, nor could
he: there was plenty of predisposition on his part. Anyway,
the majority’s rationale is not that the 10-year presumption
does not apply to decoys but that subsection (b)(2)(B) itself
may never be used unless sexual relations occur between
the perpetrator and the victim—a rule that applies equally
whether the victim is a real youngster or a pretend one. (My
colleagues deny that they have held this, and that is
technically correct, but they do not deny that this is the
logical consequence of the opinion’s reasoning.)
The undue-influence enhancement in §2A3.2(b)(2)(B) is
kin to the vulnerable-victim enhancements found through-
out the Sentencing Guidelines. Although the question
whether vulnerable-victim enhancements apply to sting
operations has not been much litigated, the sole appellate
decision on the subject holds that they may be so applied.
See United States v. Shenberg, 89 F.3d 1461, 1475-76 (11th
28 No. 02-3562
Cir. 1996). This is sensible. If a person sets out to prey on
the weak and infirm (say, selling securities to residents of
nursing homes), then the enhancement should apply even
if the “victim” in a given case is an agent posing as a dod-
dering 85-year-old woman with more money than caution.
Just so when an agent poses as a suggestible 14-year-old
girl.
Treating the enhancement for undue influence as (po-
tentially) applicable to sting operations makes functional as
well as linguistic sense. The preceding example shows why.
The difference between attempted and completed sexual
abuse lies not in subsection (b)(2)(B) but in subsection (a):
the base level is 24 if sexual relations occurred, 21 if no
sexual act was committed but the crime violated chapter
117 of Title 18 (§§ 2421-27), and 18 otherwise. Section 2423,
the crime to which Mitchell pleaded guilty, is in chapter
117, so his base offense level is 21. Having received a
3-level reduction to account for the difference between an
attempt and a completed offense, Mitchell should not
receive an additional boon by an interpretation that
restricts subsection (b)(2)(B) to completed offenses. That
would be double counting. (Even with the undue-influence
enhancement, Mitchell received a sentence of only 41
months. The district judge chose the bottom of the 41-51
month range. The statutory maximum is 15 years. On the
majority’s view, his sentence must be selected from the
range 33-41 months.)
Guideline §2A3.2 follows the norm in the Sentencing
Guidelines. Subsection (a) specifies a base offense level that
establishes a difference between attempt and the completed
offense. (Each guideline does this either directly, as in
§2A3.2, or indirectly through application of §2X1.1.)
Subsection (b) then lists aggravating and mitigating
circumstances that adjust the base offense level in response
to conduct showing that the defendant is more (or less)
dangerous than the plain-vanilla offender who commits
No. 02-3562 29
each element of the crime but does nothing else. A sexual
predator who tries to overbear the youngster’s will is more
dangerous than one who does not; a sexual predator more
than 10 years older than the victim is more dangerous than
one close to the victim’s age. Subsection (b)(2)(B) and the
presumption in Application Note 4 reflect these
things—which obtain, and thus justify a sentencing differ-
ential, whether or not the defendant succeeds. The Sentenc-
ing Commission concluded that failure was worth a three
(or six) level discount; if it set out to create a five (or eight)
level difference between success and failure, it picked an
odd way to do so. Why not be more direct?
Suppose that Dena actually had been 14 years old, as she
claimed, and that Mitchell had persuaded her to show up at
the motel. Dena’s parents arrive, having found out from her
computer where she had gone, and prevent Mitchell from
having sexual relations with their daughter. Language,
logic, and public policy all allow the application of the
two-level undue-influence enhancement to such a situation.
The police found condoms and a camera in Mitchell’s car; he
admitted, after being arrested on his way to the arranged
motel room, what he planned to do there with Dena. These
preparations were no less real than the messages Mitchell
sent over the net. Applying subsection (b)(2)(B) to Mitchell
sentences him for what he actually did, and for the risks
that such conduct creates, as sound sentencing policy
should. On the majority’s view, by contrast, even if the
victim is a 14-year-old girl rather than an agent, and even
if the defendant undermines the voluntariness of the child’s
decisions, no enhancement is proper unless sexual relations
occur. Anything that thwarts success (intervention of
parents, misunderstanding of the time or place for the
rendezvous, a snowstorm) prevents the judge from recogniz-
ing the overbearing nature of the offender’s tactics or the
significance of the difference in age.
30 No. 02-3562
At times, the majority suggests that the word “influence”
even in the present tense implies success. In many contexts
this is so; in others it is not. Consider a testator, the subject
of pages 9-11 in the majority’s opinion: A senile person may
be “unduly influenced” to make a new will, and the fact that
someone snatches the pen from his hand before he can sign
does not imply that no impropriety occurred. A lawyer who
influenced the testator unduly could (and should) be
disciplined even though a relative thwarted the scheme. To
determine whether the Sentencing Commission uses the
word “influenced” to denote success, we must see how it is
employed throughout the Guidelines. It appears repeatedly
in contexts showing that the Sentencing Commission
understands “influence” to refer to the accused’s conduct
independent of the effect on the object. Take the bribery
guidelines beginning with U.S.S.G. §2C1.1. Background
Notes to each of these describe the offense as an effort “to
influence . . . official action”, a crime that can be committed
without actually altering the official’s conduct. A person
who pays off an official in order to influence that official’s
action is culpable whether or not the official alters his
conduct as a result. Likewise it obstructs justice to wield
influence over a juror or witness, see U.S.S.G. §2J1.2
Background Note, even if the attempt fails; and efforts to
influence governmental action lead to enhancements under
U.S.S.G. §§ 2Q1.4(b)(4) and 2Q1.5(c)(1) even if the official
does his duty honestly. Many other uses of the word “influ-
ence” in the Guidelines are to the same effect.
This is a simple case. Application Note 1 defines “victim”
to include an agent posing as an underage child. Under the
banner of giving “influenced” a plain meaning, the majority
declines to apply this definition. I would follow all of the
text promulgated by the Sentencing Commission.
No. 02-3562 31
A true Copy:
Teste:
________________________________
Clerk of the United States Court of
Appeals for the Seventh Circuit
USCA-02-C-0072—12-23-03