This is an appeal from a final judgment rendered after a trial by jury in a personal-injury lawsuit. The injury occurred on a commercial construction site following the collapse of a crane. The claimant was the general contractor's superintendent, and he sued a subcontractor on claims of negligence, gross negligence, and intentional injury. The jury made findings in favor of the claimant, and the trial court rendered judgment based on the jury's award.
On original submission, we determined that the central issue was whether the Texas Workers' Compensation Act precluded the claimant's recovery of common law damages, and we concluded that it did. We explained that, under the Act, an injured worker's exclusive remedy for work-related injuries is his receipt of workers' compensation benefits. Because we concluded that the Act applied, we held that the subcontractor was immune from any liabilities arising out of its negligence. We recognized that the Act has an exception for gross negligence when the worker suffers a fatal injury, but we held that this exception did not apply because the claimant's injury was non-fatal. We also recognized that the Act has another exception for intentional injuries, but we held that the claimant could not recover under this exception because the evidence did not support a finding that an employee of the subcontractor intended to produce the specific consequences of his conduct. Based on these holdings, we concluded that the claimant's recovery of common law damages could not be sustained, and we reversed and rendered judgment that the claimant take nothing on his claims against the subcontractor.
Following our opinion, the claimant filed a motion for rehearing that challenged our liability analysis. The claimant asked that we revise our opinion to affirm the trial court's judgment or, in the alternative, that we remand for a new trial in the interest of justice. We grant the motion in part, withdraw our previous opinion, and issue this opinion on rehearing in its stead. Our liability analysis remains unchanged, but we remand in the interest of justice.
BACKGROUND
Skanska USA Building, Inc. was the general contractor on a large-scale construction project, and Tyler Lee was Skanska's superintendent. Berkel & Company Contractors, Inc. was one of the subcontractors, and it was responsible for drilling the foundation pilings of an office building that was planned for the project.
Lee suffered his injury as Berkel's crew was attempting to complete one of its pilings. To understand how that injury occurred, we begin with an explanation of how the crew normally installs those pilings.
To make a piling, the crew starts by drilling a hole into the earth using an auger, which is essentially just a large helical drill bit. Once the hole is drilled to a certain depth, the crew pumps grout into the hole through a shaft in the middle of the auger. The crew then gradually extracts the auger, allowing the grout to fill the hole. After the auger is removed, the crew drops a cage made out of steel rebar into the grout. The piling is formed as the grout hardens around the cage.
Heavy machinery is required to complete the pilings, which on this project were between 90 and 110 feet deep. To drill down to those depths, the crew used a 130-foot auger. To operate an auger of that size, the crew used a crawler crane with a 190-foot boom. The crew also staked a 150-foot column known as "the leads" into the *293ground next to the auger to guide the auger in a vertical position. The crane lifted the leads into place along with the auger.
Berkel adopted a number of policies to ensure that its pilings were completed safely. One of those policies provided that a piling should never be started unless sufficient grout is on site to fill an entire hole. This policy was violated on the day of Lee's injury.
After completing a piling, the crew had a few cubic yards of grout to spare, enough to fill a hole to a depth of about fifteen or twenty feet, but well short of the amount needed to fill a hole completely. Berkel's superintendent, Chris Miller, did not want to see this grout go to waste, and because the grout was set to expire soon, he ordered his crew to begin another piling.
Berkel's foreman, Mark Stacy, expressed opposition to Miller's plan. Stacy did not want to drill another hole until a fresh batch of grout had arrived on the jobsite. Miller responded that more grout was on its way, and because the crew was already behind schedule, he told Stacy, "If you don't drill down, I'll find somebody else that will." Stacy acquiesced, and then Miller left the jobsite.
While Miller was away, the crew pumped the old grout into a newly drilled hole, expecting that a fresh delivery of grout would be delivered soon. Against the crew's expectations, the delivery was delayed by traffic, and as the crew was forced to wait, the grout in the hole began to harden.
When the delivery finally arrived, nearly forty minutes had passed. The crew resumed its operation by drilling back down into the old grout, intending for the old grout to mix more evenly with the new grout. But the old grout created a plug around the auger, and when the new grout was pumped into the hole, the pressures beneath the plug began to build, until a sudden release forced the auger to shoot up about five feet. This upward movement caused the cable connected to the auger to backlash. The crew halted the operation so that the cable could be tightened. When the operation was restarted, the crew discovered that the auger was stuck.
Augers occasionally get stuck, and Berkel has a set of guidelines for unsticking them. Stacy tried to unstick the auger without success for nearly ten minutes, and then Miller returned to the jobsite.
Miller was, in his own words, "livid" to discover that the auger was stuck. He knew that if the auger could not be freed, then the auger would have to be cut, and the crew would fall behind schedule at least one or two days.
Miller told Stacy to step aside: "Get the F out of the way. I'm taking over." Miller then positioned himself under the boom, fifteen feet away from the leads, making him closer to the leads than anyone else on the jobsite. For the next twenty-five to forty minutes, Miller gave commands to Berkel's crane operator, Andrew Bennett.
Miller told Bennett to "bump" the auger, or turn it alternately between forward and reverse. When Bennett tried this bumping procedure, no rotational movement could be observed in the auger itself. However, unusual movement could be observed in the hydraulic lines that powered the auger. These lines were hanging from the boom of the crane, and they were heavy, weighing more than a ton. During the bumping procedure, the lines jerked in spasms, or "danced" as they were described by the crew, creating side loads that the boom was never designed to carry. The lines also sprayed so much hydraulic fluid that Bennett had to wipe it clean from the window of his cab.
*294Miller also told Bennett to hoist the auger with the crane. Hoisting caused the crane to become unbalanced. The tracks at the front of the crane dug deeper into the ground, while the tracks at the rear of the crane lifted up. In the industry, this is known as "tipping" the crane, and it is a sign that the crane is attempting to lift a load beyond the crane's rated capacity.
Bennett tried to stop the operation. He left his cab at least five times and told Miller that the crane was overloaded and that continuing would be unsafe. Berkel's excavator operator, Chris Prestridge, also tried to warn Miller, telling him that the tracks of the crane were coming off the ground.
Miller disregarded all of these warnings. He screamed at Bennett, "Pull it out! Pull it out!"
Under Miller's commands, the crane was placed under so much strain that its boom bowed and bent, until it eventually snapped in half and came crashing down.
With the collapse of the boom, the leads toppled over onto one side, away from the crane. Lee, who was standing behind a barricade, far away from the crane, saw the leads plummeting in his direction. He dove, trying to evade their fall, but the leads landed on his left leg, pinning him to the ground.
A track hoe was needed to lift the leads off of Lee. When the leads were finally removed, Lee discovered that his leg had been crushed and severed below the knee. Doctors were unable to save it. Because Lee's knee had been so badly damaged, doctors had to amputate above the knee so that Lee could be fitted with a prosthetic.
Lee was the only person harmed in the incident. Berkel's entire crew of about ten men, including Miller, managed to escape injury.
Lee recovered workers' compensation benefits through a plan administered by Skanska, but he sued Berkel seeking an additional recovery at common law.1 The jury awarded Lee more than $35 million in actual damages.2 The jury also awarded him $8.5 million in exemplary damages, based on Berkel's gross negligence.
Berkel moved for judgment notwithstanding the verdict, but that motion was denied.
ANALYSIS
Berkel raises multiple issues in this appeal, challenging both liability and damages. We only address the liability issues because they are dispositive. These issues focus on the applicability of the Texas Workers' Compensation Act and its exception for injuries that are the result of an intentional tort.
I. Overview of the Act
The Texas Workers' Compensation Act provides reciprocal benefits to covered employees and their subscribing employers. For the covered employee who sustains a work-related injury, the Act guarantees the prompt payment of medical bills and lost wages, without regard to who was at fault for causing the injury. See HCBeck, Ltd. v. Rice , 284 S.W.3d 349, 350 (Tex. 2009). For the subscribing employer, the Act provides immunity from the injured employee's common law claims. Id. This immunity is achieved by the following *295provision, which declares workers' compensation benefits the injured employee's exclusive remedy:
Recovery of workers' compensation benefits is the exclusive remedy of an employee covered by workers' compensation insurance coverage or a legal beneficiary against the employer or an agent or employee of the employer for the death of or a work-related injury sustained by the employee.
Tex. Lab. Code § 408.001(a). Under the plain text of this provision, the employer's immunity extends to his servants, which means that the co-employees of an injured employee are also protected from common law liabilities. Id.
The exclusive-remedy provision is an essential component to the Act, but it has exceptions, one of which is statutory. The Act provides that if an employee suffers a fatal injury because of the intentional act or omission of the employer or because of the employer's gross negligence, then the employee's surviving spouse or heirs are not prohibited from recovering exemplary damages. Id. § 408.001(b).
Another exception is court-made. In 1916, three years after the Act was passed into law, the Texas Supreme Court determined that the Act does not prohibit a worker from seeking common law damages when the worker suffers a non-fatal injury caused by the intentional tort of another. See Middleton v. Tex. Power & Light Co. , 108 Tex. 96, 109, 185 S.W. 556, 560 (1916). This intentional-injury exception is not based on any particular statutory provision. The Court simply explained that the worker's right to seek redress for such injuries is "protected by the Constitution and could not be taken away." Id. , 185 S.W. at 560.
This court-made exception is still a part of our law today. See Medina v. Herrera , 927 S.W.2d 597, 600 (Tex. 1996) (concluding that the intentional-injury exception survived legislative changes to the Act).
II. May Berkel claim the exclusive-remedy defense?
The applicability of the exclusive-remedy provision is the central issue in this case. If the provision does not apply, then there would be no statutory bar to Lee's common law claims.
However, if the provision does apply, then Lee cannot recover on his claim for negligence. The same is true about Lee's claim for gross negligence because Lee did not suffer a fatal injury. Lee would be limited to his workers' compensation benefits, unless his recovery of common law damages can be sustained on a theory that Berkel committed an intentional tort.
To resolve this central issue, we must first answer a threshold question: May Berkel claim the exclusive-remedy defense when Berkel is not Lee's actual employer or co-employee? Berkel argues on appeal that it may, even though Lee worked for the general contractor and Berkel was just a subcontractor. Berkel preserved this point in its motion for JNOV, which the trial court denied.
We review the trial court's ruling on a motion for JNOV under the same standard for any motion that would render judgment as a matter of law. See Cent. Ready Mix Concrete Co., Inc. v. Islas , 228 S.W.3d 649, 651 (Tex. 2007). We consider all of the evidence in the light most favorable to the nonmovant, indulging every reasonable inference and resolving any doubts against the motion. See Union Banks Ins. Co. v. Shelton , 889 S.W.2d 278, 287 (Tex. 1994). To the extent that the trial court's ruling was based on a pure question of law, our review is de novo. See In re Humphreys , 880 S.W.2d 402, 404 (Tex. 1994).
*296Two provisions in the Act govern the employment relationship between general contractors and subcontractors. The first provision establishes the general rule that a subcontractor and its employees are not the employees of the general contractor. See Tex. Lab. Code § 406.122(b). This rule applies only if the subcontractor is operating as an independent contractor and the subcontractor has entered into a written agreement with the general contractor in which the subcontractor has assumed the responsibilities of an employer. Id.
The second provision establishes a permissive exception to the general rule, allowing the general contractor to be deemed the statutory employer of the subcontractor and the subcontractor's employees "for purposes of the workers' compensation laws of this state." Id. § 406.123(e). To create this deemed employer-employee relationship, the Act requires the general contractor and the subcontractor to enter into a written agreement under which the general contractor provides workers' compensation insurance coverage to the subcontractor and the employees of the subcontractor. Id. § 406.123(a).
The undisputed evidence shows that Skanska, the general contractor and Lee's actual employer, agreed to provide workers' compensation insurance to all of its subcontractors and their employees through a contractor-controlled insurance program (CCIP). During the bidding process, Skanska instructed its prospective subcontractors to omit the cost of workers' compensation insurance, as that coverage would be provided through the CCIP. Once the contracts were awarded, Skanska required all of its subcontractors to enroll in the CCIP as a condition to performing work on the jobsite. Berkel enrolled in the CCIP, and Berkel's written agreement with Skanska confirms that Skanska provided insurance coverage to Berkel and its employees through the CCIP. Thus, for purposes of the Act, Skanska is Berkel's statutory employer, and Lee, as Skanska's actual employee, is Berkel's statutory co-employee.
As a co-employee, Berkel is entitled to rely on the Act's exclusive-remedy provision, meaning that the trial court erred by rendering judgment against Berkel on the findings that Berkel was negligent and grossly negligent. See TIC Energy & Chem., Inc. v. Martin , 498 S.W.3d 68, 78 (Tex. 2016) (concluding that the exclusive-remedy provision protected a subcontractor against the common law claims of a general contractor's employee because the subcontractor qualified as a statutory co-employee, even in the subcontractor's corporate capacity).
III. Did the trial court submit jury questions that correctly set forth the law pertaining to the intentional-injury exception?
Even though Berkel is exempt from Lee's common law claims of negligence and gross negligence, Lee could still recover against Berkel if Lee established that Berkel intentionally caused Lee's injury. Lee contends that he met this burden because he secured jury findings that would support a recovery on a theory that Berkel committed an intentional tort. In its answer to Question 2 of the charge, the jury found that a Berkel employee knew that his conduct was substantially certain to result in injury, and in its answer to Question 3, the jury found that this same conduct may be attributable to Berkel, the corporation, because the employee was a vice principal or manager.
Question 2 asked: "Did a Berkel employee acting in the course and scope of his employment believe that injury was substantially certain to result from his conduct on the date in question?"
*297Question 3 asked: "Was the conduct you have found in Question 2 the conduct of a vice principal or manager of Berkel?"
Berkel objected to both submissions. Berkel argued that Question 2 was defective because it only tasked the jury with finding knowledge that a general injury was substantially certain to occur. To meet the standard for an intentional tort, Berkel argued that this knowledge must pertain to the injury of a particular person. Accordingly, Berkel requested that Question 2 ask whether a Berkel employee knew that his conduct was substantially certain to result in injury to Lee.
As for Question 3, Berkel disagreed that corporate liability could be based on the conduct of a vice principal or a manager. Berkel argued that the corporation could act only through an upper-level executive or an alter ego. Berkel requested that Question 3 reflect this more stringent test for corporate liability.
The trial court overruled Berkel's objections, and now Berkel complains of charge error on appeal. Berkel also contends that the evidence is legally insufficient to support the jury's answers to Questions 2 and 3, even under the charge as given.
For reasons of judicial economy, we are bound to begin with appellate complaints that, if meritorious, would result in rendition of judgment. See Nat. Gas Pipeline Co. of Am. v. Pool , 124 S.W.3d 188, 201 (Tex. 2003). One such complaint would be a challenge to the legal sufficiency of the evidence, but our standard of review intertwines sufficiency challenges with alleged charge error: When a party preserves error by objecting to an erroneous charge, we measure the legal sufficiency of the evidence against the charge that should have been given. See St. Joseph Hosp. v. Wolff , 94 S.W.3d 513, 530 (Tex. 2002). Therefore, before we can address Berkel's sufficiency challenge, we must first determine whether Berkel's charge objections were valid.
We begin in reverse order with Berkel's objection to Question 3.
A. Whose conduct may be attributable to a corporate defendant for purposes of applying the intentional-injury exception?
Our law is well-settled that corporations can act only through their agents. See Hammerly Oaks, Inc. v. Edwards , 958 S.W.2d 387, 391 (Tex. 1997). But a corporation may have many different agents at many different levels within its corporate organization. The question presented here is which agents can commit intentional torts attributable to the corporation itself.
Berkel argues that the answer must be limited to only the very highest level of executives and officers-essentially, an alter ego of the corporation-citing Professor Larson's treatise on workers' compensation law. See 2 Lex K. Larson, Larson's Workers' Compensation, Desk Edition § 103.06 (Matthew Bender, Rev. Ed.) (2016) ("When the person who intentionally injures the employee is not the employer in person or a person who is the alter ego of the corporation, but merely a foreman, supervisor or manager, both the legal and the moral reasons for permitting a common-law suit against the employer collapse....").
Professor Larson wrote that the strongest argument in favor of an alter-ego test was a matter of practicality. He explained that if there were a less demanding test that allowed the conduct of an ordinary supervisor to be attributable to the corporation, then an injured employee could recover from his corporate employer simply by showing that the actual tortfeasor "was one notch higher on the totem-pole." Id.
*298Professor Larson feared that, if this were the law, then the workers' compensation system may "go out the window" because many workplace settings (factories, in particular) have layers upon layers of employees at different ranks, and assaults between those employees may be common. Id.
The Texas Supreme Court briefly acknowledged Professor Larson's position in Medina v. Herrera , 927 S.W.2d 597, 601 (Tex. 1996), where the Court cited to an earlier edition of Professor Larson's treatise. Although the issue presented by Professor Larson was deemed "important" in Medina , the parties there had not briefed it, and the Court declined to answer when the conduct of a corporate agent will be attributable to the corporation for purposes of applying the intentional-injury exception. Id.
We had an occasion to consider Medina and its reference to Professor Larson in Urdiales v. Concord Technologies Delaware, Inc. , 120 S.W.3d 400, 406 (Tex. App.-Houston [14th Dist.] 2003, pet. denied). Relying on the historically narrow construction of the intentional-injury exception, we agreed with Professor Larson that a corporate employer should not be held liable for common law damages simply because an employee was intentionally injured at work by another lower-level employee. Id. at 407. However, we stopped short of articulating a test for when the conduct of a corporate agent will be attributable to the corporation. All we said on that issue was that the facts of the case were insufficient to impute liability to the employer, and "something more must be present." Id.
Other courts of appeals have staked out different positions on the alter-ego issue. The Fourth Court of Appeals, relying on Professor Larson, has said that the alter-ego test is "possibl[y]" the law in Texas. See Horton v. Montgomery Ward & Co., Inc. , 827 S.W.2d 361, 370 (Tex. App.-San Antonio 1992, writ denied). The Ninth Court of Appeals has taken a more definitive position in a separate direction, holding that a corporate employer may be liable for the intentional tort of a vice principal, even though a vice principal may have less ownership or control over a corporation than an alter ego. See Stewart v. Lexicon Genetics, Inc. , 279 S.W.3d 364, 370 (Tex. App.-Beaumont 2009, pet. denied).
The Texas Supreme Court appears to favor the vice-principal test. Three years after Medina , in GTE Southwest, Inc. v. Bruce , 998 S.W.2d 605, 617-18 (Tex. 1999), the Court specifically held that the workplace conduct of a vice principal may be imputed to the corporate employer. The Court said, "When actions are taken by a vice-principal of a corporation, those acts may be deemed to be the acts of the corporation itself." Id. at 618.
Like this case, Bruce also involved a corporate employer's assertion of the exclusive-remedy defense, but unlike this case, the claimed injuries were not physical. There, the claim was for intentional infliction of emotional distress, and the claimants' injuries were caused by multiple instances of workplace harassment, which the Court characterized as "repetitive mental trauma." Id. at 611. The Court held that the claimants' injuries were not compensable under the Act, which meant that the exclusive-remedy provision did not apply and the claimants could recover damages under the common law. Id. Once the claims were determined to fall outside the scope of the Act, the Court held that the corporate employer could be held liable because the workplace harassment was intentionally committed by the employer's vice principal. Id. at 617-18.
*299Lee's injury differs from the injuries in Bruce because Lee suffered "damage or harm to the physical structure of the body," meaning that his injury is compensable under the Act. See Tex. Lab. Code § 401.011(26) (defining "injury"). Despite this factual difference, we see no legal reason why Bruce 's rule for attributing the conduct of an agent to the corporation should vary based on the type of injury that the agent inflicted.
Consistent with Bruce , we hold that if a claim falls outside the scope of the Act because the claimant's injury was intentionally inflicted, a corporation may be liable for the injury if the injury was caused by the corporation's vice principal acting in the course and scope of his employment. Question 3 reflects this test. Therefore, the trial court properly overruled Berkel's objection to Question 3.
B. Does the intentional-injury exception have a specific-consequences requirement?
We now consider Berkel's objection to Question 2, which Berkel challenged because it only asked whether a Berkel employee knew that an unspecified injury-rather than the injury to Lee in particular-was substantially certain to result from the employee's conduct. Resolving this issue requires that we examine what proof is needed under the intentional-injury exception. We begin with an overview of "intent."
To establish that an injury was intentional, the plaintiff must prove that the defendant acted with a particular state of mind: "the specific intent to inflict injury." See Reed Tool Co. v. Copelin , 689 S.W.2d 404, 406 (Tex. 1985). We measure this state of mind according to the Restatement's definition of intent. Id.
The Second Restatement defines intent to mean "that the actor desires to cause consequences of his act, or that he believes that the consequences are substantially certain to result from it." See Restatement (Second) of Torts § 8A. That definition, which the Texas Supreme Court applied in Reed Tool , was restyled in the Third Restatement, which provides: "A person acts with the intent to produce a consequence if: (a) the person acts with the purpose of producing that consequence; or (b) the person acts knowing that the consequence is substantially certain to result." Restatement (Third) of Torts: Phys. & Emot. Harm § 1.
The Third Restatement's definition is substantively the same as the Second Restatement's definition. The authors of the Third Restatement explained that they "unblend[ed]" the combined definition of the Second Restatement "to accommodate courts that in particular contexts might want to distinguish between intent in the sense of purpose and intent in the sense of knowledge."See Restatement (Third) of Torts: Phys. & Emot. Harm § 1 cmt. a. We employ those same distinctions here.
The purpose theory of intent is the easier type to recognize. Its paradigm is the "intentional left jab to the chin." See Reed Tool , 689 S.W.2d at 407. When an actor engages in conduct with the purpose or desire of inflicting harm on another, as with a punch or direct assault, the consequences of the actor's conduct are specifically contemplated by the actor, and the actor's conduct is almost always considered wrongful, except in limited situations like self-defense. See Restatement (Third) of Torts: Phys. & Emot. Harm § 1 cmt. a.
The knowledge theory of intent is different because it does not require that the actor have any purpose or desire to inflict harm on another. Id. In fact, "the actor may be engaging in a generally proper activity for generally proper reasons," yet still be held liable for an intentional tort if *300that "activity produces harm as an unavoidable but unwanted byproduct." Id.
Even though an actor need not have a purpose or desire to inflict harm under the knowledge theory of intent, our case law has traditionally required that the actor know that specific consequences are substantially certain to result from his conduct. In Reed Tool , for example, the Texas Supreme Court held that an employer's intentional failure to provide a safe workplace will not rise to the level of an intentional tort "except when the employer believes his conduct is substantially certain to cause the injury. " See Reed Tool , 689 S.W.2d at 407 (emphasis added). By using the definite article "the," instead of the indefinite article "an," the Court implicitly held that knowledge of a general or non-specific consequence would not suffice.
We see the same specific-consequences requirement in other contexts where the Court has applied the Restatement's definition of intent. For example, in its nuisance jurisprudence, the Court has explained that a defendant is liable for intentionally causing a nuisance only if the defendant acted for the purpose of causing the interference, or if he knew or was substantially certain that the interference would result from his conduct. See Crosstex N. Tex. Pipeline, L.P. v. Gardiner , 505 S.W.3d 580, 605 (Tex. 2016) ("To prove an intentional nuisance, the evidence must establish that the defendant intentionally caused the interference that constitutes the nuisance, not just that the defendant intentionally engaged in the conduct that caused the interference.").
The Court has also recognized a specificity requirement in its takings jurisprudence. See Harris Cty. Flood Control Dist. v. Kerr , 499 S.W.3d 793, 800 (Tex. 2016) ("In addition, a specificity element runs through our jurisprudence."). In this context, the Court has said that the government can only be liable for a taking when the government knows that its actions are causing "identifiable harm" or the government knows that "specific property damage is substantially certain to result" from its actions. See City of Dallas v. Jennings , 142 S.W.3d 310, 314 (Tex. 2004). The Court has rejected takings claims where the evidence merely shows that "the government only knows that someday, somewhere, its performance of a general governmental function, such as granting permits or approving plats, will result in damage to some unspecified parcel of land within its jurisdiction." See Kerr , 499 S.W.3d at 800.
This case law is consistent with the plain text of the Restatement, which also uses language of specificity. In both the Second Restatement and the Third Restatement, intent is defined with reference to "the consequences" of the actor's conduct. See Restatement (Second) of Torts § 8A ; Restatement (Third) of Torts: Phys. & Emot. Harm § 1.
The commentary to the Third Restatement likewise confirms that there is a specific-consequences requirement, and that this requirement extends to the identity of victims. The commentary provides: "The applications of the substantial-certainty test should be limited to situations in which the defendant has knowledge to a substantial certainty that the conduct will bring about harm to a particular victim, or to someone within a small class of potential victims within a localized area. "3 Restatement (Third) of Torts:
*301Phys. & Emot. Harm § 1 cmt. e (emphasis added).
The commentary also disapproves of an application of this test that would hold an actor liable when the identity of potential victims cannot be confidently predicted: "The test loses its persuasiveness when the identity of potential victims becomes vaguer and when, in a related way, the time frame involving the actor's conduct expands and the causal sequence connecting conduct and harm becomes more complex." Id.
Against this legal backdrop, the trial court submitted a question that was based on the knowledge theory of intent, but not the purpose theory of intent, because Lee conceded that there was no evidence that a Berkel employee had acted with the purpose or desire of causing Lee harm. However, the trial court's knowledge-based question did not reflect the requirement that the actor must have known that his conduct was substantially certain to result in specific consequences-i.e., in injury to a particular victim, or to someone in a small class of potential victims within a localized area. As submitted, the question merely asked whether the actor had knowledge to a substantial certainty that his conduct would result in "injury," unattached to any particular victim or any particular location.
The trial court's submission did not accurately reflect the law. The question should have mirrored the test articulated in the Third Restatement and asked whether a Berkel employee acting in the course and scope of his employment believed that his conduct was substantially certain to bring about harm to a particular victim, or to someone within a small class of potential victims within a localized area. We will measure the sufficiency of the evidence against this correct standard.
IV. Is the evidence legally sufficient to support a liability finding under the intentional-injury exception?
When analyzing the legal sufficiency of the evidence, we review the record in the light most favorable to the challenged findings, crediting favorable evidence if a reasonable factfinder could and disregarding contrary evidence unless a reasonable factfinder could not. See City of Keller v. Wilson , 168 S.W.3d 802, 827 (Tex. 2005). The evidence is sufficient to support a finding if the evidence rises to a level that would enable reasonable and fair-minded people to differ in their conclusions. Id. at 822. The evidence is insufficient to support a finding only if (a) there is a complete absence of evidence of a vital fact, (b) the court is barred by rules of law or evidence from giving weight to the only evidence offered to prove a vital fact, (c) the evidence offered to prove a vital fact is no more than a mere scintilla, or (d) the evidence conclusively establishes the opposite of the vital fact. Id. at 810.
Berkel asserts that there is no evidence to support the finding that one of its jobsite employees knew that his conduct was substantially certain to result in injury to Lee. Berkel also asserts that there is no evidence to support the finding that any of its jobsite employees qualified as a vice principal.
We begin with the vice-principal issue because our conclusion there will determine whose knowledge should be examined on the substantial-certainty issue.
A. Did one of Berkel's jobsite employees qualify as a vice principal?
*302The term "vice principal" encompasses four classes of agents: (1) corporate officers; (2) those who have authority to employ, direct, and discharge servants of the master; (3) those engaged in the performance of non-delegable or absolute duties of the master; and (4) those to whom the master has confided the management of the whole or a department or a division of the business. See Mobil Oil Corp. v. Ellender , 968 S.W.2d 917, 922 (Tex. 1998). The jury charge tracked this same definition.
During closing arguments, Lee argued that Miller was a vice principal and the person who was specifically contemplated by Question 3. Lee did not argue that any other person fit that definition.
The record amply supports an implied finding that Miller qualified as a vice principal. Even though Miller was not a corporate officer of Berkel, the jury could have reasonably determined that Miller had the authority to employ, direct, and discharge Berkel's employees. Stacy, Berkel's foreman, and Bennett, Berkel's crane operator, directly testified that Miller had the power to fire employees. Other members of the crew testified that they followed Miller's orders out of fear that they would be fired by Miller.
Miller denied having the power to fire the members of his crew, but the jury was free to disbelieve that aspect of his testimony. In other parts of his testimony, Miller admitted that he had the power to remove people from his jobsite and to find their replacements. The crew was not salaried, which meant that if a person was removed by Miller, that person would not be paid until he found another job, either on a different Berkel assignment or outside the Berkel organization entirely. Also, Miller actually threatened to remove certain crew members on this project.
The jury could have also found that Miller was confided with the management of a department or division. The evidence showed that Miller was a superintendent and that he represented "the highest level of operations within Berkel" on the jobsite. Miller described himself as "a boss," and he was "in charge of the overall project from Berkel's perspective." He was responsible for ensuring that supplies were ordered and that all machinery was on site and properly assembled. He was also responsible for workplace safety. Based on this level of authority and responsibility, we conclude that there is legally sufficient evidence to support an implied finding that Miller qualified as Berkel's vice principal. Cf. Purvis v. Prattco, Inc. , 595 S.W.2d 103, 105 (Tex. 1980) (holding that the conduct of an agent could be imputed to his employer where the evidence showed that the agent was a "night manager" of a motel and the "senior employee" on site during his eight-hour shift).
B. Did Miller know to a substantial certainty that his conduct would bring about harm to a particular victim, or to someone within a small class of potential victims within a localized area?
Having determined that Miller qualifies as a vice principal and that his conduct may be imputed to Berkel for liability purposes, we can now consider whether the evidence is legally sufficient to support a finding that Miller knew that the consequences of his conduct were substantially certain to occur.
This issue begs an important question: What exactly is the meaning of "substantially certain"? That phrase was not defined in the jury charge. The Restatement has not defined it either, and our case law has yet to develop a comprehensive definition.
From the body of law addressing this issue, as discussed below, we can identify *303three interrelated principles that will guide our analysis: (1) the substantial-certainty test requires much more than foreseeability; (2) the substantial-certainty test is subjective; and (3) the substantial-certainty test is fact-intensive. We address these principles in turn.
1. The substantial-certainty test requires much more than foreseeability.
Substantial certainty is a species of intent, but its meaning is sometimes gleaned from a comparison to negligence, which is a mutually exclusive theory of liability. The Texas Supreme Court made such a comparison in State Farm Fire & Casualty Co. v. S.S. , 858 S.W.2d 374 (Tex. 1993). There, the Court explained that the difference between negligence and intent is "a matter of degree," and the line that separates these two theories is drawn "where the known danger ceases to be only a foreseeable risk which a reasonable person would avoid, and becomes in the mind of the actor a substantial certainty." Id. at 378.
The Second Restatement makes a similar comparison. It imagines a sort of liability continuum, with intent and negligence occupying opposite ends of the spectrum, separated by degrees of probability:
If the actor knows that the consequences are certain, or substantially certain, to result from his act, and still goes ahead, he is treated by the law as if he had in fact desired to produce the result. As the probability that the consequence will follow decreases, and becomes less than substantial certainty, the actor's conduct loses the character of intent, and becomes mere recklessness.... As the probability decreases further, and amounts only to a risk that the result will follow, it becomes ordinary negligence....
Restatement (Second) of Torts § 8A cmt. b.
A rule emerges from these differences in degrees: An appreciation of risk, without more, does not demonstrate knowledge of a substantial certainty. See Reed Tool , 689 S.W.2d at 406 ("To establish intentional conduct, more than the knowledge and appreciation of risk is necessary...."); see also Tanner v. Nationwide Mut. Fire Ins. Co. , 289 S.W.3d 828, 832-33 (Tex. 2009) (holding that a driver's high-speed and reckless behavior did not establish a substantial certainty of injury as a matter of law).
This rule also means that a substantial-certainty finding cannot be supported when the evidence shows, at most, that the actor knowingly permitted a hazardous work condition to exist; or that he knowingly ordered the claimant to perform an extremely dangerous job; or that he willfully failed to furnish a safe place to work; or even that he willfully violated a safety statute. See Reed Tool , 689 S.W.2d at 406. These actions may set the stage for a finding of negligence or gross negligence, but to make that extra step on the liability continuum-i.e., to move beyond foreseeability to substantial certainty-the actor must know that the injury was "necessarily an incident to, or necessarily a consequential result of the [actor's] action." See City of San Antonio v. Pollock , 284 S.W.3d 809, 821 (Tex. 2009) ("A governmental entity is substantially certain that its actions will damage property only when the damage is 'necessarily an incident to, or necessarily a consequential result of the [entity's] action.' ").
2. The substantial-certainty test is subjective.
Owing to the notion that intent and negligence are separate theories of liability, we measure substantial certainty, *304a species of intent, not by what a reasonable person would have known, which is a negligence standard, but by what the actor subjectively knew. See Crosstex , 505 S.W.3d at 605. Under this subjective test, the evidence must show that the actor knew, not should have known, that the consequences of his actions were substantially certain to occur. See City of Keller , 168 S.W.3d at 829 ("The critical question in this case was the City's state of mind-the Wilsons had to prove the City knew (not should have known) that flooding was substantially certain.").
Even though intent is subjective, its proof is established through "objective indicia." Id. at 830. This proof may be circumstantial, and include such evidence as what the actor did and said, or what he saw and what he was told. See Rodriguez v. Naylor Indus., Inc. , 763 S.W.2d 411, 413 (Tex. 1989) (holding that a fact issue was raised on the issue of substantial certainty based on summary-judgment evidence of what the actor said and what he was told). An actor's statement denying a subjective intent to inflict injury is not considered conclusive. Id.
3. The substantial-certainty test is fact-intensive.
Because intent is established circumstantially, a finding of substantial certainty may turn on any number of factors. Based on one tort law treatise that the Texas Supreme Court has cited often, three factors are pertinent to the substantial-certainty analysis: (1) "the plaintiff (or a particular group of plaintiffs)"; (2) "the source of the harm"; and (3) "a particular time and place." See Dan B. Dobbs et al., The Law of Torts § 29 (2d ed. 2011);4 see also Crosstex , 505 S.W.3d at 605 (citing approvingly to the same section of Professor Dobbs's treatise).
With these principles in mind, we now turn to the evidence adduced at trial.
4. The evidence is legally insufficient.
Miller testified that he had a genuinely good relationship with Lee. He also believed that he was not in any way responsible for what had happened to Lee. When asked to grade his performance as a safe superintendent, Miller gave himself "an A or B."
The jury was not required to credit this rosy characterization. Viewed in the light most favorable to the verdict, the evidence supports a finding that Miller knew that his conduct was not safe. Bennett, the crane operator, told Miller at least five times that he should end the operation to unstick the auger. Bennett also told Miller that the operation was unsafe and that the crane was overloaded.
Miller knew that other members of the crew shared Bennett's concerns. Stacy, Berkel's foreman, urged Miller to cut the auger because he had been trying to unstick it for too long.
Miller also observed signs that the equipment was under stress. He saw that the hydraulic hoses were "dancing." He even saw hydraulic fluid spraying from the hoses.
Most importantly, Miller was warned that the crane was tipping. Prestridge, the excavator operator, told Miller, "You've got that thing coming off the ground." Prestridge also testified that Miller saw the crane tipping, "plain as day."
Miller agreed at trial that tipping the crane was extremely dangerous. He knew *305that Berkel had a stuck-auger policy that expressly forbids operating the crane when its tracks are lifted off of the ground. Miller also conceded that he could not identify a single step in that policy that he had actually followed.
From this evidence, a factfinder could reasonably conclude that Miller knew that the crane was under such severe strain that a collapse was substantially certain to occur. A factfinder could also conclude that Miller consciously disregarded the risks of a collapse because he was determined to unstick the auger, notwithstanding the safety of those around him. This evidence easily supports a finding that Miller was reckless, but it does not support a finding that Miller was substantially certain that Lee would be a particular victim of Miller's conduct. To make that greater finding, there must have been some evidence that Miller at least knew of Lee's particular location. See Dobbs, supra , § 29 ("The substantial certainty test is focused on the plaintiff ... and a particular time and place."). But there is none.
The evidence showed that the jobsite spanned about twelve acres, and Lee, as the general contractor's superintendent, was responsible for all of them. Lee could have been anywhere on those twelve acres, but in the moments preceding the collapse of the crane, he happened to be in the vicinity of Miller's operation.
Lee testified that he was at grade level, behind a barricade, walking with a dirt work contractor and ensuring that dirt was properly being spread and compacted. Berkel's crew, including Miller, was below grade level, in a large pit trying to unstick the auger. Although Miller could have foreseen that Lee might be in the vicinity, no evidence was ever produced that Miller was actually aware of Lee's particular location. Nor was there any evidence that Miller, from his position below grade level, could even see Lee at grade level, standing behind the barricade.
Because there is no evidence that Miller knew of Lee's particular location, or even of Lee's proximity to Miller, a reasonable factfinder could not determine that Miller was substantially certain that Lee would be a particular victim of Miller's conduct. Cf. Restatement (Third) of Torts: Phys. & Emot. Harm § 1 cmt. b, illus. 1 (a hunter in a forest was not substantially certain that his bullet would injure another person after missing a wild deer, the intended target, because the hunter did not know that the other person was in the vicinity).
Lee responds that there was no need for him prove that his particular location had been known to Miller. Relying on the alternative portion of the Third Restatement's commentary, Lee argues that he could satisfy his burden simply by proving that Miller had known that his conduct was substantially certain to bring about harm to "someone within a small class of potential victims within a localized area." See Restatement (Third) of Torts: Phys. & Emot. Harm § 1 cmt. e. Lee then argues that he satisfied this burden because Miller created a "zone of danger" around the crane, and Miller knew that several people were within this "zone of danger" when he engaged in his extremely dangerous operation.
Lee implicitly suggests that the Third Restatement's "localized area" may be recast as a "zone of danger." Even if we were to assume for the sake of argument that these terms were interchangeable, we would still be left with the more difficult matter of how Lee has defined the zone of danger.
Lee imagines a zone of danger that was circular in shape and measured by the fall radius of the crane. He also argues that "anyone within that radius could be injured."
*306We perceive two problems with this argument.
First, the argument does not precisely account for the source of Lee's harm. See Dobbs, supra , § 29 (providing that a court should look to "the source of the harm" when performing a substantial-certainty analysis). Lee was harmed by the leads when they fell over and landed on his leg. Even though the crane played a role in bringing down the leads, Lee was not harmed by the crane itself, as no part of the crane ever made direct contact with him. Therefore, if the zone of danger is to be measured by the fall radius of any object on the jobsite, the more appropriate measure would be the leads, not the crane.
Second, even when we reimagine Lee's zone of danger as the area measured by the fall radius of the leads, as opposed to the fall radius of the crane, this zone of danger merely defines the limits where, in Lee's own words, "anyone ... could be injured." By that description, Lee has only outlined the area where injury was foreseeable , not where injury was substantially certain.
The flaw in Lee's argument is that it equates the probability of a fall with the probability of an injury. If Miller knew to a substantial certainty that the leads were going to fall-and based on the evidence that a collapse of the crane was imminent, the jury could have reasonably concluded that he did-then Miller also knew to a substantial certainty that the leads were going to fall within the limits of their fall radius. However, this knowledge does not necessarily establish that Miller also knew that the leads were going to cause injury to someone within their fall radius. Miller could only be substantially certain of an injury if he knew that the leads would make contact with a person, and to establish that knowledge, Miller would have to know the fall path of the leads.
Lee produced an expert witness, Eric Van Iderstine, who testified about the failure of the boom, but not about the fall path of the leads. Van Iderstine explained that the boom snapped because the crane was severely overloaded. Addressing the leads, Van Iderstine explained, "For the leads to fall over indicates that what was holding it up in place failed. So that would be the boom and the wire rope." This evidence establishes the cause of the fall, not the direction of the fall.
Van Iderstine did not explain whether the leads were capable of falling in any direction (i.e., 360 degrees), or whether they were capable of falling in just one direction (i.e., the direction of Lee).5 There is also a complete absence of evidence that Miller knew, or even could have known, in which direction the leads would fall.
The fall path of the leads is more important than the fall radius of the leads because there is no evidence that the leads actually moved in a radial fashion, like the arms of a clock or the blades of a helicopter. If they had behaved in that manner, then the facts of this case would have been much worse. Berkel's entire crew of about ten men was inside the fall radius of the leads, including Miller, who was closer to the leads than any other person on the jobsite-and yet not a single person on Berkel's crew was injured. If all of these people can escape injury, despite being *307within the fall radius of the leads, then Lee's circular zone of danger is too broad to be an area of substantially certain injury.
We hold that injury was substantially certain only for someone in the fall path of the leads, and there is no evidence that Miller knew what that fall path would be. Without such evidence, a reasonable factfinder could not determine that Miller knew that his conduct was substantially certain to bring about harm to Lee as "someone within a small class of potential victims within a localized area." See Restatement (Third) of Torts: Phys. & Emot. Harm § 1 cmt. e. At best, a reasonable factfinder could only determine that Miller knowingly created a risk of injury, which is insufficient to support liability under the intentional-injury exception. See Restatement (Third) of Torts: Phys. & Emot. Harm § 1 cmt. c, illus. 4 (providing that the substantial-certainty test is not satisfied when the actor merely knows that his conduct produces "a significant likelihood that someone will suffer physical harm").
Lee makes three final points that must be addressed.
First, Lee highlights testimony from Bennett and Miller in which they both agreed that a general, non-specific injury was substantially certain to result from a dangerous operation of the crane.6 This testimony may permit an inference that Miller knew that someone would be injured from his operation, but it does not support the greater finding that Miller (or Berkel) committed an intentional tort against Lee specifically. Cf. Kerr , 499 S.W.3d at 808-09 (rejecting an expansive theory of liability that would make the government accountable for a taking based on the government's knowledge that its actions are substantially certain to cause property damage "somewhere, someday ... to some as yet unidentified private property"); see also Restatement (Third) of Torts: Phys. & Emot. Harm § 1 cmt. e (opposing an application of the substantial-certainty test where the identity of potential victims is vague).
Second, Lee asserts, in a footnote, that the trial court mistakenly excluded computer data showing the degree to which the crane was overloaded. Lee suggests this issue as a cross-point, but he makes no discernible argument supported by citation to authority. We conclude that this cross-point is waived.7 See Tex. R. App. P. 38.2(a) ; Tex. R. App. P. 38.1(i).
Third, Lee argues that Berkel is responsible for the intentional conduct of Bennett, the crane operator. Lee makes no argument, however, that Bennett would qualify as Berkel's vice principal, which is necessary to support Berkel's liability under the intentional-injury exception. Even if we assumed without deciding that Bennett *308was a vice principal, the evidence only shows that Bennett, like Miller, engaged in extremely risky behavior. There is no evidence that Bennett knew to a substantial certainty that Lee would be a particular victim of that behavior. Nor is there any evidence that Bennett knew the fall path of the leads, or that Lee was standing in that fall path.
V. What is the appropriate disposition: rendition or remand?
When there is no evidence in support of a judgment, we generally render judgment against the party with the burden of proof. See Nat. Gas Pipeline Co. of Am. v. Justiss , 397 S.W.3d 150, 162 (Tex. 2012). However, our rules provide that we may choose to remand for another trial in the interest of justice. See Tex. R. App. P. 43.3(b).
Remand is appropriate when there is a probability that a case has, for any reason, not been fully developed. See Union Pac. R.R. v. Seber , 477 S.W.3d 424, 435 (Tex. App.-Houston [14th Dist.] 2015, no pet.). Courts have remanded in the interest of justice when a decision has clarified the way in which a claim should be submitted to the jury. See Torrington Co. v. Stutzman , 46 S.W.3d 829, 840-41 (Tex. 2000) ; Vecellio Ins. Agency, Inc. v. Vanguard Underwriters Ins. Co. , 127 S.W.3d 134, 141 (Tex. App.-Houston [1st Dist.] 2003, no pet.). Courts have also remanded after deciding that a case needs further development because it was tried on an incorrect legal theory or to establish and present evidence regarding an alternate legal theory. See Ahmed v. Ahmed , 261 S.W.3d 190, 196 (Tex. App.-Houston [14th Dist.] 2008, no pet.).
Until our opinion in this case, no appellate decision had examined in detail the requirements of the substantial-certainty test. Indeed, this case appears to be the first involving that test to ever be tried to a verdict in the personal-injury setting. Our opinion clarifies the standard for submitting to the jury a claim that a defendant knew that his conduct was substantially certain to injure the plaintiff. In the interest of justice, we believe that Lee should have an opportunity to establish and present evidence under our clarified standard. See Torrington , 46 S.W.3d at 840-41 ; Vecellio , 127 S.W.3d at 141.8
CONCLUSION
The judgment against Berkel is reversed and the case is remanded for additional proceedings consistent with this opinion.
( Donovan, J., dissenting).
His wife, Leigh Ann, joined the suit too, individually and as next friend of their minor daughter. Where appropriate in this opinion, we use "Lee" in the singular form to refer to both Tyler and Leigh Ann, collectively.
The jury determined that Berkel was responsible for 90% of these damages, leaving the remaining 10% to a separate defendant not involved in this appeal.
The Second Restatement does not articulate a test along these lines, but it does provide an illustration from which the same test can be inferred. The illustration provides as follows: "A throws a bomb into B 's office for the purpose of killing B. A knows that C , B 's stenographer, is in the office. A has no desire to injure C , but knows that his act is substantially certain to do so. C is injured by the explosion. A is subject to liability to C for an intentional tort." Restatement (Second) of Torts § 8A cmt. b, illus. 1.
The Texas Supreme Court has cited to Professor Dobbs dozens of times over the years, sometimes to this treatise where he is the lead author, and other times to the treatise he co-authored with Deans Prosser and Keeton.
The potential range of directions in which the leads could have fallen differentiates this case from another case that Lee has cited, Klepsky v. Dick Enterprises, Inc. , 55 Fed.Appx. 270 (6th Cir. 2003). In Klepsky , the source of the harm was a deck pan attached to the underside of a bridge. Id. at 271-72. Because the deck pan was not already touching the ground, the direction of its fall was predetermined by gravity. The deck pan could only go one way: down.
Bennett's testimony:
Q. So going back, Mr. Bennett, you would agree with me that if you or Mr. Miller knew that the crane was coming off the back of the ground but operated it anyway, that y'all would have been operating knowing a substantial certainty to cause injury or death, right?
A. Correct.
Miller's testimony:
Q. And I asked Mr. Bennett this, and I'll ask you the same thing. You would agree with me that operations are substantially certain to cause serious injury or death if you're operating a crane that's 180 or 200 percent overloaded? You would agree with that?
A. Correct.
Even if this cross-point were not waived, the excluded evidence would only help to establish knowledge to a substantial certainty that the crane would collapse. The evidence would not establish knowledge that the injury to Lee was "necessarily an incident to, or necessarily a consequential result of" the collapse. See Pollock , 284 S.W.3d at 821.
In our opinion on original submission, we did not address Berkel's argument that Lee's recovery of common law damages was precluded by the election-of-remedies doctrine. If meritorious, this argument would provide an alternative basis for rendition. We conclude that the argument is not meritorious and, therefore, it presents no obstacle to remand. See Medina , 927 S.W.2d at 602 (holding that when a person is intentionally injured by a co-employee for work-related reasons, the person is not barred by the election-of-remedies doctrine from recovering compensation benefits under the plan administered by his employer as well as common law damages arising from the intentional tort committed by the co-employee).