State v. Williams

The opinion of the court was delivered by Johnson, J.: Keshia Denise Williams was convicted of aggravated battery for stabbing her long-time friend, Sandra Kelly, multiple times in the head with a steak knife during an altercation inside of Williams’ home. On appeal, Williams claimed, inter alia, that the trial court should have given certain instructions even though Williams did not request them. We granted review of the instruction issues, in which Williams claims the district court — 508 — should have, sua sponte, given the following instructions: PIK Crim. 3d 54.17-A on no duty to retreat; PIK Crim. 3d 54.18 on use of force in defense of a dwelling; and a lesser included offense instruction on severity level 7 aggravated battery. We find that the failure to give the instructions was not clearly erroneous and therefore the defendant is not entitled to obtain relief on appeal. We affirm the Court of Appeals’ affirmance of the district court. Factual and Procedural overview On April 26, 2008, Kelly and her boyfriend, Steve Jones, had been staying with Williams for 3 days. That night, Cynthia Edwards and Donald Ray McWilliams were also present in Williams’ house. The conflict between Williams and Kelly began when Williams, while seated at the kitchen table, told Jones that he and Kelly, who was then in the bedroom, would have to leave her house. Jones relayed comments between the two women to and from the kitchen and bedroom until Kelly entered the kitchen to directly engage in a verbal altercation with Williams. Kelly repeatedly challenged Williams to “take it outside.” Initially, Williams tried to defuse the situation, but eventually she agreed to go outside. When Kelly went to the bedroom to put on her tennis shoes, Williams feared that Kelly had obtained a weapon. Accordingly, at some point after Williams got up from the kitchen table to follow Kelly outside, she picked up a knife. The two only made it to the door before tire physical fighting began. According to Edwards, Kelly started that altercation by “throwing blows” and pulling two braids out of Williams’ hair. Williams responded by stabbing Kelly in the head. The fight stopped when Kelly realized blood was coming down her face. McWilliams then took the knife from Williams, and Edwards threw it into some trees. Police subsequently recovered a knife north of Williams’ house, near some trees. But evidence was not produced at trial that definitively established that the found knife was the one used to stab Kelly. Williams grabbed a towel and wrapped it around Kelly’s head. She then called 911 and reported that she had stabbed Kelly, albeit — 509 — she would later tell police that she believed that she had acted in self-defense. Jones and another person took Kelly to the hospital, where she received “about a hundred stitches.” Kelly would later say that she had not felt much, if any, pain; that the only subsequent medical attention she required was to have the stitches removed; and that her wounds healed in a couple of months. The police interviewed Kelly at the hospital. She initially told them that she had fallen and cut her head on a piece of glass. After the officers threatened to “get [her] for making a false statement,” Kelly related that she had been cut while fighting with Williams. Eventually, Williams was charged with and tried on one count of severity level 4 aggravated battery, K.S.A. 21-3414(a)(l)(A). Williams moved for a judgment of acquittal at the close of the State’s evidence. The trial court denied the motion, finding that, although there was no evidence connecting the knife the police had found with Kelly’s stabbing injuries, there was evidence from which a jury could find that Kelly suffered great bodily harm. In the process of explaining that ruling, the court opined that “a hundred stitches is more than slight, trivial, minor or moderate harm.” During trial, the prosecutor cross-examined Williams about the fact that she did not try to run from Kelly and hide or “do anything to get away from her.” In closing argument, the prosecutor'returned briefly to the theme that Williams did not try to get away from Kelly. The trial judge gave the jury a general self-defense instruction, but Williams was nevertheless found guilty of the level 4 aggravated battery charge. Upon appeal to the Court of Appeals, a split panel affirmed the conviction. The dissent opined that the trial court should have given a no duty to retreat instruction and a lesser included offense instruction on severity level 7 aggravated battery, notwithstanding the absence of any request for those instructions. See State v. Williams, No. 102,615, 2010 WL 4156759, at *9-13 (Kan. App. 2010) (unpublished opinion) (Leben, J., dissenting). The majority accepted the State’s explanation that it proffered Williams’ failure to run and hide as proof that she was not afraid of — 510 — Kelly and, therefore, resorting to a deadly weapon was unjustified under the subjective prong of self-defense. As noted, we granted review on the three issues involving unrequested instructions. Standard of Review and the Clearly Erroneous Standard Currently, when an instruction issue is being raised for the first time on appeal or has not been properly preserved with an appropriate objection in the trial court, we generally refer to K.S.A. 22-3414(3) and recite simply that “the standard of review is whether the instruction is clearly erroneous.” State v. Adams, 294 Kan. 171, 183, 273 P.3d 718 (2012); see State v. Tully, 293 Kan. 176, 196, 262 P.3d 314 (2011); State v. Magallanez, 290 Kan. 906, 925, 235 P.3d 460 (2010); State v. Ellmaker, 289 Kan. 1132, 1145, 221 P.3d 1105 (2009), cert. denied 130 S. Ct. 3410 (2010). Often then, while still discussing the standard of review, we will explain diat “ ‘[jjuiy instructions are clearly erroneous only if the reviewing court is firmly convinced that the jury would have reached a different verdict had the error not occurred.’ State v. Tully, 293 Kan. 176, 196, 262 P.3d 314 (2011).” Adams, 294 Kan. at 183. That shorthand has tended to blur the distinction between what is to be decided on appeal and how the appellate decision is to be made. The “what” that must be decided on appeal typically follows a three-step process: (1) determining whether the appellate court can or should review the issue, i.e., whether diere is a lack of appellate jurisdiction or a failure to preserve the issue for appeal; (2) considering the merits of the claim to determine whether error occurred below; and (3) assessing whether the error requires reversal, i.e., whether the error can be deemed harmless. How those decisions are made is driven by the applicable standard of review, which “focuses on the deference due a lower court, jury, or agency.” 1 Childress & Davis, Federal Standards of Review § 1.03, p.1-17 (3d ed. 1999). In other words, the standard of review establishes die “framework by which a reviewing court determines whether the trial court erred.” Hall, Hall’s Standards of Review in Texas, 42 St. Mary’s L.J. 3, 13 (2010). — 511 — In that vein, our defining clearly erroneous in terms of the likelihood of a different juiy verdict is more akin to stating a harmless error test than a standard of review, because it “assumes that an error has occurred; it does not guide tire appellate court in determining whether an error has occurred or how much deference to give to a lower court decision.” Hodgkinson, Clear as Mud? “Clearly Erroneous” as a Standard of Review for Instructional Claims, Kansas Bar Association Appellate Practice Newsletter, at p. 2 (Spring 2012). Accordingly, we will pause to take a critical look at how we have historically handled unpreserved appellate challenges to jury instructions with a view to establishing a more precise analytical framework and more accurate standards of review. Because we frequently cite to K.S.A. 22-3414(3) as the source of our standard of review, we look first to that statute. It provides, in relevant part: “No party may assign as error the giving or failure to give an instruction, including a lesser included crime instruction, unless the party objects thereto before the jury retires to consider its verdict stating distinctly the matter to which the party objects and the grounds of the objection unless the instruction or the failure to give an instruction is clearly erroneous.” K.S.A. 22-3414(3). Plainly, the statute sets forth the statutoiy preservation requirement that the complaining party must have objected prior to jury deliberations in order to get an appellate review of a claimed instruction error. A preservation rule for instruction claims, requiring an objection in the trial court, was part of this State’s jurisprudence prior to the statute. For instance, in 1943, we declared that “the general rule is that where no objection is made to the giving of an instruction during the trial and no request was made for its modification or clarification and such instruction is not clearly erroneous a litigant cannot be heard to complain on appeal.” Sams v. Commercial Standard Ins. Co., 157 Kan. 278, 287, 139 P.2d 859 (1943); see also Collett v. Estate of Schnell, 194 Kan. 75, 79, 397 P.2d 402 (1964) (following rule from Sams); Montague v. Burgerhoff, 152 Kan. 124, 128-29, 102 P.2d 1031 (1940) (stating that when no objection is made to any of the jury instructions, instructions become law of the case); Hogan v. Santa Fe Trail Transportation Co., 148 Kan. 720, 724, 85 P.2d 28 (1938) (“No objection was made — 512 — thereto at the time of tire trial and a complaint at this time is too late.”); Jones v. A., T. & S.F. Rly. Co., 148 Kan. 686, 695, 85 P.2d 15 (1938) (stating that instruction became law of the case and jury’s findings must control appeal because appellant “made no complaint of it”); Lambert v. Rhea, 134 Kan. 10, 14-15, 4 P.2d 419 (1931) (stating that absent an objection, jury instruction complaints are unavailing in the Supreme Court where instructions correctly stated the law). Just as plainly, K.S.A. 22-3414(3) includes an exception to the preservation requirement where the instruction or failure to give an instruction was “clearly erroneous.” That exception also predated the statute. In Sams, we “conceded [that] the [preservation] rule does not apply to an instruction which is in itself erroneous and an appellant is not estopped from complaining of it as error by not having objected to it at the time it was given.” Sams, 157 Kan. at 288; see also Collett, 194 Kan. at 79 (“This court adheres to the rule that where the instructions or directions of the trial court are in themselves erroneous, an appellant is not estopped of complaining of them as error by not having objected at the time they were given.”); Richardson v. Business Mens Protective Ass’n., 129 Kan. 700, Syl. ¶ 2, 284 P. 599 (1930) (stating that when an instruction includes statements of law which are “clearly erroneous,” it is not necessaiy, in order for party to predicate error upon the giving of the instruction, to make an objection). The rule was applied in criminal cases, as well. See State v. Ragland, 173 Kan. 265, 269, 246 P.2d 276 (1952) (“it is not necessary in order to predicate error thereon, that a defendant in a criminal action object to the giving of an instruction to the jury, if the instruction is clearly erroneous”). What the statute does not specifically tell us is the meaning of “clearly erroneous.” Likewise, it does not indicate that the district court is to be afforded any deference when the instruction or failure to give the instruction is clearly erroneous. Over time, this court’s description and application of “clearly erroneous” has been fluid. In the earlier cases, the term related to die content of the instruction, with the inquiry being whether the instruction at issue on appeal correctly stated the law. The focus was on the patent — 513 — bona fides of the instruction, rather than the degree of error or prejudice caused by any error. See, e.g., Lambert, 134 Kan. 10, Syl. ¶ 4; Richardson, 129 Kan. 700, Syl. ¶ 2; Foley v. Crawford, 125 Kan. 252, Syl. ¶ 10, 264 P. 59 (1928). State v. Severns, 158 Kan. 453, 456, 148 P.2d 488 (1944), nicely illustrates those early cases that equated “clearly erroneous” with legally invalid. There, the defendant argued for the first time on appeal that the murder instruction misstated the law by indicating that murder at common law included the unlawful killing of a human being committed in the perpetration of a misdemeanor. The Secerns court first noted that the appellant had not objected at trial, but then related that such an objection is “not necessary, in order to predicate error, if the instruction is clearly erroneous.” 158 Kan. at 456 (citing Richardson, 129 Kan. 700, Syl. ¶ 2). The court then held, as a matter of law, that the statutory definition of murder did not include a killing “in perpetration of a misdemeanor,” i.e., the instruction was clearly erroneous. 158 Kan. at 457. The court went on to reverse without engaging in a separate harmless error analysis, concluding summarily that “[t]he instruction as given was erroneous and prejudicial.” 158 Kan. at 458. It appears that clearly erroneous instructions were generally treated as being presumptively prejudicial and, thus, reversible. In 1970, tire legislature enacted K.S.A. 22-3414(3), in the Code of Criminal Procedure, which closely followed the provision regarding waiver of jury instructions found at K.S.A. 60-251(b) in the Code of Civil Procedure. Both required a party to make a proper objection to the giving or failure to give an instruction before the jury retired to consider its verdict, in order to preserve an appellate claim, unless the instruction was clearly erroneous. L. 1970, ch. 129, sec. 22-3414; L. 1963, ch. 303, sec. 60-251. For a time, our appellate courts appeared to continue to equate “clearly erroneous” with legal validity. See State v. Humbolt, 1 Kan. App. 2d 137, Syl. ¶ 4, 562 P.2d 123 (1977) (instruction “not ‘clearly erroneous’ where it correctly states the law and does not mislead the jury”). But later, the focus of the “clearly erroneous” analysis changed from a determination of whether the instruction was patently erroneous to a determination of whether the instruction was clearly — 514 — prejudicial. Perhaps the first case to make that shift was State v. Stafford, 223 Kan. 62, 65, 573 P.2d 970 (1977), which declared: “It is to be noted that none of the objections to instructions now raised were presented to the trial court; therefore, our scope of review is limited to a determination of whether the instructions are ‘clearly erroneous.’ (K.S.A. 22-3414[3]; State v. Birch, 221 Kan. 122, 558 P.2d 119 [1976]; State v. Nesmith, 220 Kan. 146, 551 P.2d 896 [1976]; State v. Suing, 210 Kan. 363, 502 P.2d 718 [1972].) An instruction is clearly erroneous when the reviewing court reaches a firm conviction that if the trial error had not occurred there was a real possibility the jury would have returned a different verdict.” (Emphasis added.) Stafford did not cite to a source for its new definition of “clearly erroneous.” Certainly, there is nothing in the statutory language which would naturally lead one to the conclusion that “clearly erroneous” was meant to be determined by a reversibility standard, i.e., by the errors perceived effect on the trial outcome. In fact, one might find it counterintuitive to make the reviewability of an issue hinge entirely upon its reversibility. Further, Stafford did not explain or acknowledge that it was changing the K.S.A. 22-3414(3) clearly erroneous exception from a certainty-of-error concept to a certainty-of-prejudice concept. In fact, the opinion seems to rely more on the legal validity of the challenged instructions than their impact on the verdict. Nevertheless, subsequent cases picked up Stafford’s definition of clearly erroneous. See State v. Houck, 240 Kan. 130, 139, 727 P.2d 460 (1986) (citing Maxwell and Stafford); State v. Maxwell, 234 Kan. 393, 399, 672 P.2d 590 (1983) (citing Stafford). Eventually, the Stafford version of the meaning of “clearly erroneous” would achieve the status of “well known.” See Ellmaker, 289 Kan. at 1139. Additionally, after Stafford, in 1985, this court clearly and unequivocally assigned the “standard of review” label to the term “clearly erroneous.” State v. Peck, 237 Kan. 756, 764, 703 P.2d 781 (1985), declared: “This ‘clearly erroneous’ standard is the correct standard of appellate review since appellant gave no basis for his objection to the instruction at trial. We have held a party may not assign as error tire giving or failure to give an instruction unless he objects to tire instruction stating the specific grounds for the objection. Absent such specificity an appellate court may reverse only if the instruction is clearly erroneous. State v. Korbel, 231 Kan. 657, Syl. ¶ 4, 647 P.2d 1301 (1982).” — 515 — Again, the origin of the statement designating clearly erroneous as the correct standard of appellate review is unexplained. Peck’s citation to Korbel only refers to syllabus ¶ 4, where that opinion simply repeated tire statutory language of K.S.A. 22-3414(3). State v. Korbel, 231 Kan. 657, 647 P.2d 1301 (1982), did not recite any standard of review on the applicable instruction issue in that case, much less call “clearly erroneous” the “correct standard of appellate review.” But again, subsequent cases fell into step with Peck, consistently referring to “clearly erroneous” as a “standard of review.” See, e.g., State v. Kelly, 262 Kan. 755, 764, 942 P.2d 579 (1997); State v. Hunt, 257 Kan. 388, 394-95, 894 P.2d 178 (1995); State v. Crawford, 253 Kan. 629, Syl. ¶ 1, 861 P.2d 791 (1993); State v. Perez, 251 Kan. 736, 737, 840 P.2d 1118 (1992). Obviously, designating “clearly erroneous” as a standard of review fails to impart the normal information about how the appellate court intends to go about its review. For instance, will the court conduct an unlimited review of the appellate record or look for an abuse of discretion by the trial judge? Likewise, declaring that jury instructions are clearly erroneous only if we are firmly convinced that the jury would have reached a different verdict had the error not occurred does not reveal the process to be used to determine whether there was any error at all. But before we can more accurately describe the standard of review, we need to clarify just what the court must decide when presented with a claim of error for giving or failing to give an instruction that was not requested. As we noted earlier, the first step in the appellate process is normally a reviewability inquiiy. Obviously, K.S.A. 22-3414(3) directly impacts that detennination. Although it purports to withhold appellate jurisdiction in the absence of a proper objection, the statute’s exception effectively conveys such jurisdiction and preserves for appellate review any claim drat the instruction error was clearly erroneous. Moreover, to determine whether it was clearly erroneous to give or fail to give an instruction, the reviewing court would necessarily have to first determine whether it was erroneous. In other words, to determine whether the claim of error is properly renewable, the court must first determine whether there is an error, i.e., perform — 516 — the merits review in the second step of the normal appellate process. That review for error necessarily presents a legal question subject to unlimited review. Only after determining that the district court erred in giving or failing to give a particular instruction would the reviewing court engage in the reversibility inquiry. Given that it has been utilized for decades, the current definition of clearly erroneous sets up the test to determine whether the instruction error requires reversal, i.e., whether tire reviewing court is firmly convinced that the jury would have reached a different verdict had the instruction error not occurred. This assessment of whether there has been injustice would involve a review of the entire record and a de novo determination. Cf. State v. Ward, 292 Kan. 541, Syl. ¶ 8, 256 P.3d 801 (2011) (harmless error analysis performed de novo), cert. denied 132 S. Ct. 1594 (2012). Before moving on, we pause to note that the burden of establishing reversibility differs in this context from the harmless error analyses discussed in Ward. Whereas the burden to show harmlessness generally shifts to the party benefitted by the error, the burden to show clear error under K.S.A. 22-3414(3) remains on the defendant. Cf. United States v. Vonn, 535 U.S. 55, 58-59, 122 S. Ct. 1043, 152 L. Ed. 2d 90 (2002) (a silent, nonobjecting defendant has burden to satisfy plain error rule in context of Federal Rule of Criminal Procedure 11). No Duty to Retreat Instruction Williams’ first complaint is that the trial judge, on his own, did not give the jury the no duty to retreat instruction from PIK Crim. 3d 54.17-A (2007 Supp.). That instruction tracked the statutory provision at K.S.A. 21-3218 and provided: “A person who is not engaged in an unlawful activity and who is attacked in a place where (he)(she) has a right to be has no duty to retreat. (He)(She) has the right to stand (his)(her) ground and to meet force with force.” PIK Crim. 3d 54.17-A (2007 Supp.). Citing to Chambers v. Mississippi, 410 U.S. 284, 294, 93 S. Ct. 1038, 35 L. Ed. 2d 297 (1973), Williams first casts the issue in terms of her fundamental right to present a full and complete de — 517 — fense. But nothing precluded the defense from making the concept of no duty to retreat a part of its theory of defense by requesting the instruction. Requesting an instruction signals the trial judge that its contents are part of the defense theory. Certainly, a defendant does not have a constitutional right to have the trial judge develop the defense theory. Moreover, characterizing the issue as a constitutional claim does not significantly advance Williams’ procedural posture. Even constitutional grounds for reversal are not properly before the appellate court for review if they are being asserted for the first time on appeal. See State v. Gaudina, 284 Kan. 354, 372, 160 P.3d 854 (2007). Accordingly, Williams can pass the first step of establishing that the issue is properly reviewable by this court only if the failure to give a no duty to retreat instruction was clearly erroneous per K.S.A. 22-3414(3). As noted, to get to that determination, we must first consider whether it was error at all to omit the instruction. Williams relies heavily on our decision in State v. Scobee, 242 Kan. 421, 748 P.2d 862 (1988). There, the victim and the defendant had an altercation while in separate vehicles on public roads. After the defendant returned to and parked in the driveway of his parent’s home, the victim, who had followed the defendant home, charged at the defendant. The defendant shot and killed the victim. We reversed the defendant’s involuntary manslaughter conviction based upon the trial court’s refusal to give the no duty to retreat instruction as a supplement to the general self-defense instruction. 242 Kan. at 429. The Scobee decision was heavily influenced by the State’s trial strategy of emphasizing that the defendant had failed to retreat, specifically arguing that he could have driven to a police station or simply remained in his car and honked the horn. Williams suggests that this case is so factually similar to Scobee as to require the same result. Granted, the prosecution here used the same tactic of pointing out to the juiy that the defendant had not reacted to the victim’s aggression by retreating. Unlike the Court of Appeals majority, we do not find the State’s explanation of the purpose for that strategy to be reasonably persuasive. The specific statement during closing was: “She claims self-defense. — 518 — She never tried to get away. She never tried to hide.” The patently obvious inference to be drawn from that argument is that for a defendant to legitimately claim self-defense, he or she must have first tried to get away or tried to hide. A no duty to retreat instruction would have dispelled the State’s implication that diere was a duty to retreat. But the differences in Scobee are more compelling than the similarities. First and foremost, Scobee involved a different procedural posture; Scobee requested that the trial judge give the no duty to retreat instruction. That distinction is a game changer. If Williams had requested the instruction here, it is possible tíiat we would have found it to have been a legally and factually appropriate instruction that was supported by some evidence, i.e., found that the district court had erred in refusing to give the requested instruction. See State v. Plummer, 295 Kan. 156, 162, 283 P.3d 202 (2012) (the district court errs if it refuses to give a requested instruction that is legally appropriate and factually supported by some evidence). In that event, we would have been looking at whether the State had carried its burden to establish harmless error, i.e., “if there is a reasonable probability that the error will or did affect the outcome of the trial in light of the entire record.” Ward, 292 Kan. at 569. Instead, the defendant, who must assume at least some of the responsibility for the omitted instruction by failing to request it, bears the burden of firmly convincing us that the jury would have reached a different verdict had the error not occurred. For the reasons discussed below, Williams did not clear that high hurdle. Factually, this case is more akin to what occurred in State v. Saleem, 267 Kan. 100, 977 P.2d 921 (1999). There, the defendant and the victim started an argument inside the defendant’s sister’s house and then agreed to step outside to settle their differences. Once outside, a shoring match ensued, albeit the evidence was conflicting as to which person was the aggressor in the shoring match. After the shoring caused the defendant to fall to the ground, he pulled a weapon and fatally shot the victim, firing four shots in the process. This court found that a no duty to retreat instruction was not required in that scenario. 267 Kan. at 113-15. — 519 — In reaching its decision, Saleem analyzed both Scobee and State v. Ricks, 257 Kan. 435, 437, 894 P.2d 191 (1995), but found Ricks to be more applicable. Ricks involved a confrontation in a public parking lot. The Ricks opinion limited Scobee1s holding to the "infrequent factual situations ... with such elements as a nonaggressor defendant being followed to and menaced on home ground.” Ricks, 257 Kan. at 437. In this case, Williams agreed to go outside with Kelly to settle their argument. She must have thought that a physical altercation was likely, because she armed herself enroute. In essence, the evidence would support a reasonable inference that Williams agreed to engage in mutual combat with Kelly and took a knife to a fistfight. Granted, the testimony was that Kelly threw the first punch. But it seems strained to say that a person who has agreed to fight another is nevertheless a “nonaggressor” because the other combatant beat that person to the first punch. Moreover, Williams aggressively responded to the commencement of the fight by stabbing Kelly repeatedly in the head. Except for the fact that the trial court instructed the jury on self-defense, it is questionable whether a no duty to retreat instruction was factually appropriate in this case. Nevertheless, even if we were to find that omitting the instruction was erroneous, Williams has failed to establish the requisite level of prejudice. At the time Kelly threw the first punch, Williams had agreed to “take it outside” and had armed herself with a deadly weapon. The jury heard and rejected the general self-defense instruction. Even Scobee said that ordinarily the general self-defense instruction, then contained in PIK Crim. 2d 54.17, would be sufficient. 242 Kan. at 428. Accordingly, we are not firmly convinced that the instruction would have made a difference in the jury verdict. Use of Force in Defense of Dwelling Instruction Next, Williams contends that the trial judge should have given the defense of dwelling instruction, set forth in PIK Crim. 3d 54.18 (2007 Supp.), which provides: “Defendant claims (his)(her) conduct was permitted as a lawful defense of [(his)(her)] (dwelling) (occupied vehicle). — 520 — “Defendant is permitted to use force to the extent that it appears to (him)(her) and (he)(she) reasonably believes that such force is necessary to prevent another person from unlawfully (entering into) (remaining in) (damaging) [(his)(her)] (dwelling) (occupied vehicle). Reasonable belief requires both a belief by defendant and the existence of facts that would persuade a reasonable person to that belief. “[Defendant is permitted to use deadly force to prevent another person from unlawfully (entering into) (remaining in) (damaging) [(his)(her)] (dwelling) (occupied vehicle) only when (he)(she) reasonably believes deadly force is necessary to prevent death or great bodily harm to (himself)(herself) (someone else). Reasonable belief requires both a belief by defendant and the existence of facts that would persuade a reasonable person to that belief.] “When use of force is permitted as a lawful defense of [(bis)(her)] (dwelling) (occupied vehicle), there is no requirement to retreat.” Again, Williams did not request the instruction, so our review is precluded unless the failure to give the instruction was clearly erroneous. Williams is so far from reaching that standard that we will make short work of this issue. Simply put, Williams’ dwelling was not under attack. Her argument is that because she had told Jones that he and Kelly could no longer stay at her house, Kelly’s status transformed from a houseguest to a trespasser. Therefore, the argument continues, Williams could use force to prevent Kelly from remaining in the dwelling. If Williams wanted to be creative in crafting an argument that would have justified the giving of the defense of dwelling instruction, she should have done so in the district court. We will certainly not require trial judges to divine unique defense theories, especially where such theories are not revealed by the evidence. Williams testified that she tried to defuse the argument by asking Kelly to sit down at the kitchen table to talk. One who has been invited to sit at the kitchen table is really not a trespasser, even if that person’s status as a houseguest has been terminated. Moreover, when the two were at the front door, it was for the purpose of going outside to fight. Williams was not using force to prevent Kelly from remaining in the house; Kelly was voluntarily headed out the door. In short, the defense of dwelling instruction was not applicable to the facts of this case, and it was not clearly erroneous to fail to give that instruction. — 521 — Lesser Included Offense Instruction Finally, Williams contends that the trial judge should have, on his own, given a lesser included offense instruction for severity level 7 aggravated battery. That version of aggravated battery is defined as “intentionally causing bodily harm to another person with a deadly weapon, or in any manner whereby great bodily harm, disfigurement or death can be inflicted.” K.S.A. 21-3414(a)(1)(B). In contrast, the severity level 4 version with which Williams was charged, tried, and convicted is defined in relevant part as “[ijntentionally causing great bodily harm to another person.” K.S.A. 21-3414(a)(l)(A). In ruling on Williams’ motion for judgment of acquittal, which is referred to in the record as a motion for directed verdict, the district court found that the State had not proved that the knife the police found was used to inflict the injuries on Kelly, indicating a failure of proof that bodily harm was caused with a deadly weapon. Therefore, the relevant difference between the two crimes in this case is whether Williams actually caused great bodily harm to Kelly or whether she merely caused bodily harm but did so in a manner that could have inflicted great bodily harm. Again, the reviewability of the issue is governed by the clearly erroneous exception in K.S.A. 22-3414(3), which specifically refers to the “failure to give an instruction, including a lesser included crime instruction.” On the way to making a clearly erroneous determination, we must necessarily look first at whether it was legally and factually appropriate for the district court to give a lesser included offense instruction on severity level 7 aggravated battery. See Plummer, 295 Kan. 156, Syl. ¶ 1. The instruction would have been legally appropriate because “level 7 aggravated battery is a lesser included offense of level 4 aggravated battery.” State v. Hernandez, 294 Kan. 200, 205, 273 P.3d 774 (2012); accord State v. Winters, 276 Kan. 34, 38, 72 P.3d 564 (2003). Further, the giving of lesser included crime instructions is not a matter of discretion with the trial judge. K.S.A. 22-3414(3) directs that “where there is some evidence which would reasonably justify a conviction of some lesser included crime . . . , — 522 — the judge shall instruct the jury as to the crime charged and any such lesser included crime.” (Emphasis added.) Here, there was obviously evidence of bodily harm. Eyewitnesses testified that Williams stabbed Kelly in the head, causing blood to flow down her face. Kelly testified that it took approximately 100 stitches to close her head wounds. Further, a reasonable jury could certainly find that stabbing someone in the head with a steak knife is a “manner whereby great bodily harm, disfigurement or death can be inflicted” under K.S.A. 21-3414(a)(l)(B). Accordingly, there was the requisite evidence to support the giving of an instruction on severity level 7 aggravated battery. Williams attempts to portray the trial judge’s statement that “a hundred stitches is more than slight, trivial, minor or moderate harm” as being a finding that the injuries suffered by Kelly were “great bodily harm” as a matter of law, thus precluding the lesser included offense instruction on merely “bodily harm.” But that takes the judge’s remarks completely out of context. He was ruling on a defense motion for directed verdict that required the court to determine whether the State had presented a prima facie case before resting, i.e., whether it had presented sufficient evidence from which a rational jury could find each of the crime’s elements beyond a reasonable doubt. One of those elements was that Williams had caused “great bodily harm.” That term has been defined as “more than slight, trivial, minor, or moderate harm, that does not include mere bruising, which is likely to be sustained by simple battery.” State v. Green, 280 Kan. 758, 765, 127 P.3d 241, cert. denied 549 U.S. 913 (2006). Therefore, the district court was merely opining that tire evidence that Kelly required 100 stitches would be sufficient to support a jury finding of great bodily harm, not that the court was making that finding as a matter of law. Nevertheless, we must consider the question, even if the trial judge did not. If the evidence established, as a matter of law, that Kelly suffered great bodily harm, then the evidence could not reasonably justify a conviction for causing mere bodily harm, and a lesser included offense instruction on the severity level 7 version of aggravated battery was not required. — 523 — Ordinarily, whether a victim has suffered great bodily harm is a question of fact for the jury to decide. Green, 280 Kan. at 765. But that seemingly straightforward proposition has become muddled by some opinions that chose to take the decision away from the jury. See, e.g., State v. Moore, 271 Kan. 416, 420-21, 23 P.3d 815 (2001) (holding bums and scarring from hot iron on victim’s legs, breast, and inner thighs constituted great bodily harm as matter of law); State v. Valentine, 260 Kan. 431, 435, 921 P.2d 770 (1996) (holding that bullet wound which severed spinal cord and caused paralysis constituted great bodily harm as matter of law); State v. Gideon, 257 Kan. 591, 614, 894 P.2d 850 (1995) (holding that rape or aggravated criminal sodomy constituted great bodily harm as matter of law); Doolin v. State, 24 Kan. App. 2d 500, 503-04, 947 P.2d 454 (1997) (holding that bullet wound which required hip bone and hip socket to be replaced constituted great bodily harm as matter of law); but cf. State v. Brice, 276 Kan. 758, 773-74, 80 P.3d 1113 (2003) (bullet that goes through body without hitting any vital organs may or may not cause great bodily harm); State v. Vessels, No. 96,421, 2008 WL 1847374, at *5 (Kan. App. 2008) (unpublished opinion) (broken bones do not necessarily constitute “great bodily harm” as matter of law). Here, Kelly’s own testimony sent mixed signals to the jury as to whether her injuries were great bodily harm or mere bodily harm. Although she related that her wounds required a large number of stitches, she also minimized the pain she had suffered and said that she did not require any follow-up medical services other than to remove the stitches. A reasonable and rational jury could have gone either way: bodily harm in a manner that could have caused great bodily harm, or great bodily harm. Accordingly, it was error not to give the lesser included offense instruction on severity level 7 aggravated battery. But our determination that the omission of the instruction was erroneous does not answer the question of whether the failure to give the unrequested instruction was clearly erroneous. In other words, just because we find that a rational jury could have found Williams guilty of the lesser included offense does not necessarily mean that we believe that the jury would have convicted her of — 524 — the lesser offense. Here, the evidence is such that we simply cannot be firmly convinced of which crime the jury might have chosen, as between the severity level 4 and severity level 7 versions. That degree of certainty, or perhaps more accurately, that degree of uncertainty falls short of what is required to meet the clearly erroneous standard. Accordingly, we affirm Williams’ conviction. Affirmed.