353 F.3d 552
UNITED STATES of America, Plaintiff-Appellee,
v.
John MITCHELL, Defendant-Appellant.
No. 02-3562.
United States Court of Appeals, Seventh Circuit.
Argued April 4, 2003.
Decided December 23, 2003.
Tu Pham (Argued), Office of the United States Attorney, Chicago, IL, for Plaintiff-Appellee.
Terence F. MacCarthy, Angela F. Milella (argued), Office of the Federal Defender Program, Chicago, IL, for Defendant-Appellant.
Before POSNER, EASTERBROOK, and ROVNER, Circuit Judges.
ILANA DIAMOND 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 communicating 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 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 partner 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 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 appearance, 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 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 everything [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.
Mitchell pled guilty, admitting that he traveled in interstate 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 supported 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 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 presumption, 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 presumption.
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 requires 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 construction. 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, 113 S. Ct. 1913, 123 L. Ed. 2d 598 (1993).
The guideline itself supports Mitchell's argument about the inapplicability of the subsection to attempts. The 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 compromised 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 influencing or compromising. In other words, the enhancement cannot apply where the offender and victim have not engaged in illicit sexual conduct.
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, 121 S. Ct. 361, 148 L. Ed. 2d 213 (2000) ("Whatever merits these and other policy arguments may have, it is not the province of this Court to rewrite the statute to accommodate 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, 102 S. Ct. 1096, 71 L. Ed. 2d 188 (1982) ("we have no authority to rewrite ... legislation based on mere speculation 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 guideline 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 characteristics 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 influence" in the civil context to which we may look for guidance.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, 95 Ill. 2d 452, 69 Ill. Dec. 960, 448 N.E.2d 872, 875 (1983) (internal citation omitted); see also In re Estate of Hoover, 155 Ill. 2d 402, 185 Ill. Dec. 866, 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 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 "influence" where the object of the influence has not acted accordingly. 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 overpowers the dominated party's free will or judgment that he or she cannot act intelligently or voluntarily, but acts, instead, 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 otherwise have done but for such dominion and control." Black's Law Dictionary 1528 (6th ed. 1990). It is this understanding of the term "influence," along with the past tense construction 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.
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 definition of "victim" which ensures that offenders who are apprehended in undercover sting operations are appropriately 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 operations 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 attempted 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 example, Mitchell does not dispute the application of a two-level 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 purpose 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 punishment 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 addressed 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 enhancement at issue in this case, section (b)(2)(B). In any case, as we noted before, this is not a case where we need to decide how the enhancement applies in the case where the defendant is interrupted before the illegal act with a real victim could be completed.
For the skeptic who insists on looking past the plain language 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 presumption of the commentary. Recall that the guidelines create a rebuttable presumption that the participant unduly influenced the victim if he is at least ten years older than the victim. U.S.S.G. § 2A3.2 cmt. n.4. The commentary instructs 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 presumption 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 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 rebuttable 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 Mitchell'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? 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 industry 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, 537 U.S. 1176, 123 S. Ct. 1006, 154 L. Ed. 2d 921 (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 compromised," 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 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, 296 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 enhancement 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 determine 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 examination 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, 296 F.3d at 1234.10 But if the purpose of the enhancement is to mete out punishment based on the defendant's conduct, as the government argues, why would the voluntariness 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 happened 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 susceptible 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, 95 S. Ct. 1255, 43 L. Ed. 2d 541 (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 assaulting 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 influenced 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 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 sentencing 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 compromised the voluntariness of the victim's behavior. U.S.S.G. § 2A3.2, cmt. n.4. The majority in Root, for example, identified specific language from the Internet conversations which demonstrated how the defendant had used his increased knowledge, persuasive powers, and superior resources 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 colloquy between Mitchell's counsel and the district court judge regarding the similarities between this case and the 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.
For the reasons stated above, we reverse and remand for resentencing.
REVERSED AND REMANDED.
Notes:
All references to the United States Sentencing Guidelines are to the 2001 version in effect at the time of Mitchell's sentencing
A chat room is a place on the world wide web where Internet users with common interests can sign on to communicate in real 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
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
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
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.
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 operations. This is not so
The dissent implies that Mitchell might have rebutted the presumption 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
InRoot, 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.
Because this opinion will create a conflict among the Circuit Courts, it has been circulated to all of judges of this court in regular active service pursuant to Circuit Rule 40(e). A majority of judges in active service voted not to rehear this caseen banc. (Judges Easterbrook and Kanne voted to rehear this case en banc.)
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 dispositive 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.
Relying onFeola, 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, 95 S. Ct. 1255). 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.
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 vulnerability. As we describesupra, 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.
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.
EASTERBROOK, Circuit Judge, dissenting.
When calculating the sentence for a person convicted of sexually abusing 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 Application Note 1, which defines "victim" to mean "(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." 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 prohibited 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 operations, and to affirm the judgment of the district court.
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 linguistically: 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 definition 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 informative. 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 misrepresentation of a participant's identity to (i) persuade, induce, entice, or coerce the victim to engage in prohibited sexual conduct; or (ii) facilitate transportation 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 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 subsections (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 characteristics 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 subsections (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.
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 sexual 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 overbearing 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 presumption, 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 Application Note 4 ¶ 4. Mitchell was 50 years old at the time of the events in question; Dena pretended to be 14. The presumption of undue influence therefore 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 unconstitutional. Yet there's nothing wrong with presumptions 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 improbable." Challenges to this presumption have been unsuccessful. 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 predisposition, 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).
And if the presumption turns out to be very hard to defeat, 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 throughout 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 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 doddering 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 (potentially) 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 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 differential, whether or not the defendant succeeds. The Sentencing 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 recognizing the overbearing nature of the offender's tactics or the significance of the difference in age.
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 557-59 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 "influence" 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.