OPINION
J. WOODFIN JONES, Chief Justice.
The University of Texas at Austin (the “University”) appeals the county court at law’s order denying its plea to the jurisdiction. See Tex. Civ. Prac. & Rem.Code Ann. § 51.014(a)(8) (West 2008). The University asserts that sovereign immunity bars the negligence claims brought by the ap-pellee, Robert Hayes. Hayes sued the University for injuries he sustained when he rode his bicycle onto campus at night [882]*882and collided with a metal chain stretched across the throughway on which he was traveling.1 He alleged both a premises defect and a special defect under the Texas Tort Claims Act. See id. §§ 101.001-.109 (West 2005 & Supp 2008) (the “Act”). We will affirm the trial court’s order.
FACTUAL AND PROCEDURAL BACKGROUND
Hayes alleged that at approximately 8:30 p.m. on the evening of September 12, 2003, he and his wife were riding their bicycles from their home north of the University campus to a restaurant southwest of campus. Entering the campus on San Jacinto Boulevard, Hayes and his wife began to ride up the road that proceeds uphill from Winship Circle to the Gregory Gym parking lot. Hayes alleged that he had taken this route many times before and was very familiar with it. As he was riding, his front bicycle tire struck a metal chain that was stretched across the throughway. As a result of the collision, Hayes was thrown from his bicycle and suffered personal injuries, including a broken jaw. According to Hayes, his bicycle was equipped with a functioning light and reflectors, and his light was turned on at the time of the accident. He further alleged that the chain was rusted and dark in color and that, given the advancing darkness and the poor lighting in the area, he could not see it even though he was looking straight ahead before the collision occurred. He pleaded that there was no reflective material or any other reflectors or lights on the chain itself.
According to the University, the chain had been placed across a “closed” service driveway earlier that day for the purpose of configuring university parking for a football game scheduled for the following day.2 Positioned in front of the chain was a large, orange and white reflective sawhorse barricade that Hayes had tried to bypass in an attempt to continue on the throughway. Hayes does not dispute the existence of the barricade, but stated that it blocked only the center of a throughway that was several feet wider than the barricade, and that he thought he could simply ride his bicycle around it.
In his first amended petition, Hayes alleged that the presence of the chain constituted either a premises defect or a special defect. The University filed an amended plea to the jurisdiction and motion to dismiss. Under its primary theory, the University asserted that the jurisdictional evidence conclusively established that Hayes was a trespasser at the time of his accident, and therefore he was required to allege that the University’s conduct was grossly negligent, willful, or wanton. His petition contained no such allegation. In the alternative, the University asserted that even if Hayes was not a trespasser for purposes of the Act, he was not an invitee, but only a licensee, and that the evidence conclusively established that it had discharged its duty to warn of the allegedly dangerous condition by putting up the reflective barricade. These are the only arguments asserted in the University’s amended plea to the jurisdiction.
[883]*883For his part, Hayes filed a motion for partial summary judgment asserting that the University’s immunity had been waived pursuant to section 101.021 of the Act. After a hearing, the trial court denied the University’s amended plea to the jurisdiction. In a separate order, the trial court granted Hayes’s motion for partial summary judgment, finding that the University’s immunity had been waived and that “[a] final ruling on immunity to liability is not yet ripe for determination.”
The University now brings this interlocutory appeal of the denial of its amended plea to the jurisdiction. See Tex. Civ. Prac. & Rem.Code Ann. § 51.014(a)(8).
STANDARD OF REVIEW
In Texas, sovereign immunity deprives a trial court of subject-matter jurisdiction for lawsuits in which the state or certain governmental units have been sued unless the state consents to suit. Texas Dep’t of Parks & Wildlife v. Miranda, 138 S.W.3d 217, 224 (Tex.2004). Thus, a governmental unit such as the University properly raises the issue by a plea to the jurisdiction. See id. at 225-26. We review the denial of a plea to the jurisdiction de novo. Id. at 226. In deciding a plea to the jurisdiction, we may not weigh the merits of the plaintiffs claims, but must consider only the plaintiffs pleadings, construed in the plaintiffs favor, as well as any evidence pertinent to the jurisdictional inquiry. County of Cameron v. Brown, 80 S.W.8d 549, 555 (Tex.2002). However, in some cases, disputed evidence of jurisdictional facts that also implicate the merits of the case may require resolution by the finder of fact. Miranda, 133 S.W.3d at 226. Only if the pleadings and jurisdictional evidence affirmatively and conclusively negate the existence of jurisdiction should a plea to the jurisdiction be granted. Id. at 227.
When a plea to the jurisdiction challenges the pleadings, we first determine if the pleader has alleged facts that affirmatively demonstrate the court’s jurisdiction to hear the cause. Id. at 226 (citing Texas Ass’n of Bus. v. Texas Air Control Bd., 852 S.W.2d 440, 446 (Tex.1993)). We construe the pleadings liberally in favor of the plaintiff and look to the pleader’s intent. Id. If the pleadings do not contain sufficient facts to affirmatively demonstrate the trial court’s jurisdiction, but do not affirmatively demonstrate incurable defects in jurisdiction, the issue is one of pleading sufficiency and the plaintiff should be afforded the opportunity to amend. Id. at 226-27 (citing Brown, 80 S.W.3d at 555). If the pleadings affirmatively negate the existence of jurisdiction, then a plea to the jurisdiction may be granted without allowing the plaintiff an opportunity to amend. Id. at 227.
If a plea to the jurisdiction challenges the existence of jurisdictional facts, we consider relevant evidence submitted by the parties when necessary to resolve the jurisdictional issues raised, as the trial court is also required to do. Id. If the evidence creates a fact question regarding the jurisdictional issue, the trial court cannot grant the plea to the jurisdiction, and the fact issue will be resolved by the fact finder. Id. at 227-28. If, however, the relevant evidence is undisputed or fails to raise a fact question on the jurisdictional issue, the trial court rules on the plea to the jurisdiction as a matter of law. Id. at 228. This standard generally mirrors that of a traditional summary judgment under Texas Rule of Civil Procedure 166a(c). Id. Therefore, when reviewing a plea to the jurisdiction in which the pleading requirement has been met and the plea is supported by evidence that implicates the merits of the case, we take as true all evidence favorable to the nonmovant, and [884]*884we indulge every reasonable inference and resolve any doubts in the nonmovant’s favor. Id.
DISCUSSION
On appeal, the University argues in two issues that Hayes’s pleadings and jurisdictional evidence failed to establish that its immunity had been waived pursuant to the Act. See Act § 101.021. Under the Act, a governmental unit’s immunity from suit is waived to the extent that the Act creates liability. Miranda, 133 S.W.3d at 224. The Act expressly waives sovereign immunity for premises-defect and special-defect claims, subjecting the governmental unit to liability for personal injury caused by a condition or use of real property if the governmental unit would, were it a private person, be liable to the claimant under Texas law. Act §§ 101.021, .022; Brown, 80 S.W.3d at 554. Under Texas law, whether a condition of real property is classified as a premises defect or a special defect controls the entrant’s legal status, which in turn determines the duty owed by the landowner. See Act § 101.022; Brown, 80 S.W.3d at 557-58; State Dep’t of Highways v. Payne, 838 S.W.2d 235, 237 (Tex.1992); City of Austin v. Rangel, 184 S.W.3d 377, 382-83 (Tex.App.-Austin 2006, no pet.).
In the present case, if the chain across the throughway was a premises defect and not a special defect, then the University owed Hayes the same duty that a private landowner owes a licensee. See Payne, 838 S.W.2d at 237 (citing Act § 101.022(a)). That duty would require the University (1) not to injure a licensee by willful, wanton, or grossly negligent conduct, and (2) to use ordinary care to warn of or make reasonably safe a condition that poses an unreasonable risk of harm of which the University is aware and the licensee is not. See id. If, instead, the condition was a special defect, the University’s duty is the same as that owed by a private landowner to an invitee. See id. (citing Act § 101.022(b)). With respect to an invitee, a governmental unit is required to use ordinary care to reduce or eliminate an unreasonable risk of harm of which the governmental unit knew or reasonably should have known. Id. A third entrant status, that of trespasser, is not expressly addressed under the Act, but exists at common law. The common law duty of care owed to a trespasser is merely to refrain from injuring the trespasser willfully, wantonly, or through gross negligence. State v. Shumake, 199 S.W.3d 279, 287 (Tex.2006); City of Bellmead v. Torres, 89 S.W.3d 611, 613 (Tex.2002).
Because the question of whether the University’s immunity from suit has been waived relates to Hayes’s legal status and the resulting duty of care owed by the University, this is a case, like Miranda, in which the jurisdictional challenge squarely implicates the merits of the plaintiffs cause of action. In these circumstances, “after the state asserts and supports with evidence that the trial court lacks subject matter jurisdiction, we simply require the plaintiffs, when the facts underlying the merits and subject matter jurisdiction are intertwined, to show that there is a disputed material fact regarding the jurisdictional issue.”3 Miranda, 133 S.W.3d at 228. This protects plaintiffs from having to put on their case simply to establish jurisdiction. Id. Unless the evidence conclusively [885]*885shows an absence of jurisdiction, the disputed evidence regarding the jurisdictional facts that implicates the merits of the case will usually require resolution by the trier of fact, thereby giving the trial court the discretion to defer resolution of the jurisdictional issue. Id. at 226. In this case, the trial court made an express finding that a final ruling on the issue of the University’s immunity “is not yet ripe for determination.” Therefore, under Miranda, the trial court properly acted within its discretion by deferring resolution of the ultimate jurisdictional facts unless the University conclusively established that it retained its immunity under the Act.
Hayes’s Status as a Trespasser
In its first issue, the University argues that because Hayes was a trespasser at the time of the accident, and because he failed to plead that the University acted willfully, wantonly, or with gross negligence, his pleadings affirmatively negate the existence of jurisdiction. Thus, while framed as a challenge to the pleadings, the University essentially argues that the jurisdictional evidence conclusively established that Hayes was a trespasser.
In support of its plea, the University points to the fact that Hayes was not a current student, faculty member, or employee and that he was riding across the campus for the purpose of recreation and convenience in order to reach an off-campus destination. It further urges that Hayes was cycling through campus in violation of the University’s traffic regulations: upon entering campus, he took a road marked, “No Thru Traffic”; he rode his bicycle in the “wrong direction” on the circle driveway; and he admitted to ingesting marijuana and alcohol before departing from his house, contrary to the University’s regulations prohibiting the operation of motor vehicles and bicycles while under the influence of alcohol or other controlled substances.
Hayes responds that the University’s assertion that he was a trespasser because he is not currently enrolled or employed at the University is groundless, pointing to evidence that the University permits and encourages bicycle riding and acquiesces to the use of the throughway by the public, including cyclists such as himself who have routinely ridden through the campus “countless times over many years.” He also argues that the sole traffic regulations relevant to the question of whether he was a trespasser are those rules governing the entry of vehicles, which only limit access to the campus during regular business hours, from 7:30 a.m. until 5:45 p.m. It is undisputed that Hayes entered campus after 8:30 p.m.
We conclude that the jurisdictional evidence creates a fact question as to whether Hayes was a trespasser. The University cites no authority for the proposition that violation of its traffic regulations renders a person a trespasser as a matter of law for purposes of civil actions, and we find none. The provisions of the parking and traffic regulations manual do not define and are not relevant to Hayes’s legal status at the time he entered the University campus; they simply set forth the rules that a motor vehicle operator or cyclist must follow while present on University property. Cf. Park v. Troy Dodson Constr. Co., 761 S.W.2d 98, 101 (Tex.App.-Beaumont 1988, writ denied) (holding that violation of water code provisions prohibiting use of motor vehicles on levees did not establish trespass as matter of law). For example, the manual provides that the director of parking and transportation services may restrict or reroute traffic as necessitated by special events, including on public streets, and that “it shall be prohibited to violate such regulations.” However, it does not state that by violating the [886]*886regulations, the person becomes a trespasser.
Nor does evidence that Hayes entered the campus for his own convenience in order to reach an off-campus destination conclusively establish that he was trespassing. See, e.g., Mellon Mortg. Co. v. Holder, 5 S.W.3d 654, 671 (Tex.1999) (O’Neill, J., dissenting) (noting cases where “ [licensees have been found to include: those taking short cuts across the property ...; loafers, loiterers and people who come in only to get out of the weather; [...] spectators and sightseers not in any way invited to come; those who enter for social visits-”) (citations omitted). Specifically, Hayes has raised a fact issue concerning whether the University impliedly consents to the public’s entry by producing evidence that members of the public routinely enter the campus, that the University makes no attempt to regulate vehicle traffic outside of normal business hours, and that no signs inform the general public that they are prohibited from entering the campus or using the inner-campus roadways. Cf. Murphy v. Lower Neches Valley Auth., 529 S.W.2d 816, 820 (Tex.Civ.App.-Beaumont 1975) (plaintiffs not held to be trespassers where summary-judgment evidence showed that children swam in canal every day, defendant knew yet never asked them to leave, and no signs prohibited their activity), rev’d on other grounds, 586 S.W.2d 561 (Tex.1976). Because the University did not conclusively establish that Hayes was a trespasser, we hold that the University’s plea to the jurisdiction was properly denied on this basis.4 See City of Corsicana v. Stewart, 249 S.W.3d 412, 414 (Tex.2008). We overrule the University’s first issue.
Nature of the Defect and the Duty Owed to Hayes
In its second issue, the University argues that even if Hayes was not a trespasser, its immunity has not been waived because it conclusively established that it discharged the duties it owed to Hayes under the standard of care owed to a licensee. In so arguing, the University maintains that the chain across the driveway was a mere premises defect, rather than a special defect, and thus Hayes cannot claim the elevated standard of care owed to invitees. See City of Dallas v. Reed, 258 S.W.3d 620, 622 (Tex.2008) (“The legal distinction between a premises and special defect lies in the duty owed by the State to the person injured.”).5 But [887]*887before we can analyze the University’s specific contention that it discharged any duty it might have owed to Hayes as a licensee, we must address the issues precedent to determining that Hayes was shown to be a licensee as a matter of law.
Unreasonable Risk of Harm
As a threshold question, we first consider whether the University proved as a matter of law that chain in the throughway was no “defect” at all. To bring a claim for either a special defect or a premises defect, the plaintiff must plead and prove that the condition of the property posed “an unreasonable risk of harm.” See City of Dallas v. Giraldo, 262 S.W.3d 864, 869 (Tex.App.-Dallas 2008, no pet.) (“Regardless, under either theory, the plaintiff must prove the condition of the premises created an unreasonable risk of harm to the licensee or invitee.”); see also Brinson Ford, Inc. v. Alger, 228 S.W.3d 161, 162 (Tex.2007); Payne, 838 S.W.2d at 237. The University argues on appeal that the chain did not present an unreasonable risk of harm because the presence of the reflective barricade in front of the chain adequately warned of the danger presented by the chain. We disagree that the University has negated the unreasonable-risk-of-harm element as a matter of law.
“A condition presenting an unreasonable risk of harm is defined as one in which there is a sufficient probability of a harmful event occurring that a reasonably prudent person would have foreseen it or some similar event as likely to happen.” Seideneck v. Cal Bayreuther Assocs., 451 S.W.2d 752, 754 (Tex.1970). Therefore, whether a condition constitutes an unreasonable risk of harm is, by definition, a function of reasonableness: “That is, if the ordinarily prudent man could foresee that harm was a likely result of a condition, then it is a danger.” Rosas v. Buddies Food Store, 518 S.W.2d 534, 537 (Tex.1975). “It follows that an owner or occupier of land can be charged with knowledge and appreciation of a dangerous condition on his premises only if from a reasonable inspection a reasonably prudent person should have foreseen a probability that the condition would result in injury to another.” Seideneck, 451 S.W.2d at 754 (citing Restatement (Second) of Torts § 343(a) (1965)).
“As might be expected with a matter involving the ‘reasonable man’ concept, there seems to be no definitive, objective test which may be applied to determine whether a condition presents an unreasonable risk of harm.” Seideneck, 451 S.W.2d at 754. Thus, the question of whether a condition poses an unreasonable risk of harm is usually an inherently fact-intensive question determined from the standpoint of a reasonably prudent person. See, e.g., Hall v. Sonic Drive-In of Angleton, Inc., 177 S.W.3d 636, 645 (Tex.App.-Houston [1st Dist.] 2005, pet. denied) (“The determination of whether a particular condition poses an unreasonable risk of harm is generally fact specific.”); Brooks v. First Assembly of God Church, 86 S.W.3d 793, 797 (Tex.App.-Waco 2002, no pet.) (“Whether a particular risk is unreasonable or not is a fact question for the jury.”); Reliable Consultants, Inc. v. Jaquez, 25 S.W.3d 336, 342 (Tex.App.-Austin 2000, pet. denied) (“It is important to note that reasonableness determinations such as the one here are fact-intensive inquiries and, as such, are issues well-suited for a jury’s determination.”).
While the reasonable-person and “unexpectedness” standards are designed to be objective, see City of Austin v. Leggett, 257 S.W.3d 456, 472 (Tex.App.-Austin 2008, pet. denied), reasonable minds may, in many cases, differ as to the dangerous character of a condition, see Rosas, 518 S.W.2d at 537, and the determination will [888]*888depend on the facts and circumstances of each individual case, see, e.g., Corbin v. City of Keller, 1 S.W.3d 743, 747 (Tex.App.-Fort Worth 1999, pet. denied) (flash flood at low-water crossing is entirely predictable to ordinary motorist traveling in inclement weather); Harding v. Kaufman County, 119 S.W.3d 428, 434 (Tex.App.Tyler 2003, no pet.) (potholes and ruts in unpaved road in creek bottom after recent rain are neither unusual nor unexpected). Under some circumstances, of course, a court can determine as a matter of law that a particular condition does not pose an unreasonable risk of harm. For example, the supreme court has held that an icy bridge during cold, wet weather does not pose an unreasonable risk of harm because it is not unexpected or unusual; rather, it is “entirely predictable.” See State Dept. of Highways & Pub. Transp. v. Kitchen, 867 S.W.2d 784, 786 (Tex.1993). Similarly, the court has determined that the accumulation of mud on an outdoor concrete slab is “nothing more than dirt in its natural state and, therefore, is not a condition posing an unreasonable risk of harm.” M.O. Dental Lab v. Rape, 139 S.W.3d 671, 676 (Tex.2004). Yet the court has also held that a wet floor on a rainy day does not pose a danger so “open and obvious as to relieve [the store owner] of any duty or to charge [the entrant] with knowledge and full appreciation of the nature and extent of the danger,” Rosas, 518 S.W.2d at 538, despite the fact that the accumulation of water inside the automatic doors of a grocery store after a day of rain is arguably foreseeable and expected.
Construing the facts before us in Hayes’s favor, as we must, we cannot conclude as a matter of law that a dark metal chain stretched at night across the entire width of a throughway and sidewalk often used by bicyclists and pedestrians, without cones or other reflective material on the chain, did not pose an unreasonable risk of harm. Users of this throughway cannot be charged with anticipating the sort of danger posed by the chain in the same sense that drivers should expect that bridges will ice over in cold, wet weather, see Kitchen, 867 S.W.2d at 786, or that pedestrians should know that after rainfall, the dirt on a sidewalk will create a “slippery mud surface,” M.O. Dental Lab, 139 S.W.3d at 672. Allegations that Hayes failed to exercise care by trying to ride past the reflective barricade or because of his alleged intoxication, to the limited extent that these matters are even raised in the University’s plea, are issues of comparative responsibility for the fact-finder to decide and do not render the danger posed by the chain in this case expected or foreseeable as a matter of law. See Brown, 80 S.W.3d at 557. We hold that Hayes has not failed to plead, nor do his pleadings affirmatively negate, the unreasonable-risk-of-harm element necessary to both his special-defect and premises-defect claims. We further hold that the jurisdictional evidence does not conclusively show that the chain did not present an unreasonable risk of harm.
Special Defect
We now address the University’s argument that Hayes cannot claim the status of an invitee because, as a matter of law, the chain was not a special defect. Hayes, who pleaded that the chain constituted both a premises defect and a special defect, asserted in his petition that the University “had either actual or constructive knowledge of a condition that posed an unreasonable risk of harm,” that it “did not exercise reasonable care to reduce or eliminate the risk,” and that “its failure to use such care proximately caused Plaintiff, who had no knowledge of the condition, to suffer damages.”
It is undisputed that Hayes’s pleadings allege a special-defect claim for [889]*889which immunity has been waived; however, the University asserts that the jurisdictional evidence conclusively established that the condition posed by the chain was not a special defect.6 The determination of whether a condition is a special defect is a question of law for the court to decide. City of Grapevine v. Roberts, 946 S.W.2d 841, 843 (Tex.1997); Payne, 838 S.W.2d at 238. Under the Act, special defects include “defects such as excavations or obstructions on highways, roads, or streets.” Act § 101.022(b). Although the statute does not define “special defect,” the Texas Supreme Court has explained that, under the ejusdem generis rule, “we are to construe ‘special defect’ to include those defects of the same kind or class as excavations or obstructions.” Texas Dep’t of Transp. v. York, No. 07-0743, 2008 WL 5105254, at *2, 2008 Tex. LEXIS 1003, at *5-6, — S.W.3d-(Tex. Dec. 5, 2008) (citing County of Harris v. Eaton, 573 5.W.2d 177, 179 (Tex.1978)). A condition may be a special defect only if it is an excavation, obstruction, or some other condition that presents “an unexpected and unusual danger to ordinary users of roadways.” Kitchen, 867 S.W.2d at 786. Moreover, a special defect must pose a threat to the ordinary users of a roadway. Payne, 838 S.W.2d at 238 n. 3; see Reed, 258 S.W.3d at 622.
When determining whether a special defect exists, courts often look to factors such as the size, nature, location, and permanence of the condition. See, e.g., York, 2008 WL 5105254, at *2, 2008 Tex. LEXIS 1003, at *7, — S.W.3d at - (loose gravel on roadway does not share characteristics of excavation or obstruction and is therefore not special defect); City of Grapevine, 946 S.W.2d at 843 (partially crumbled and cracked sidewalk step is not special defect); Payne, 838 S.W.2d at 239 (culvert was too far from road to be encountered by normal users of roadway); Eaton, 573 S.W.2d at 179-80 (abnormally large hole in highway was special defect); Harris County v. Smoker, 934 S.W.2d 714, 718-19 (Tex.App.-Houston [1st Dist.] 1996, writ denied) (longstanding, routine, or permanent conditions are not special defects).
The University does not dispute that the chain in this case was an “obstruction.” See Eaton, 573 S.W.2d at 181 (Steakley, J., dissenting) (“ ‘Obstruction’ is defined as a thing that obstructs or impedes; an obstacle, impediment, or hindrance.”) (quoting Webster’s New International Dictionary (2d ed.1961)). Nor does it challenge whether Hayes was within the scope of the “ordinary users” of the throughway to whom the chain posed a danger. The University contends, however, that the chain was stretched across an area that is not a “highway, road, or street” under section 101.022(b) of the Act and therefore as a matter of law was not a special defect. According to the University, the “inner-campus service driveway” where Hayes was injured was instead a “private road or driveway,” arguing that it was located “inside the guard shack perimeter of the University’s campus,” that “it did not directly connect to City of Austin streets,” and that it was not used by the general public; consequently, the University asserts, “the service driveway and Winship Circle were restricted to those who were authorized to [890]*890enter campus and who followed the University’s traffic regulations.”
The Act itself does not define the terms highway, road, or street. In these instances, courts have often looked to the definitions provided by the transportation code. See, e.g., Rangel, 184 S.W.3d at 383; City of El Paso v. Chacon, 148 S.W.3d 417, 423 (Tex.App.-El Paso 2004, pet. denied); City of Wichita Falls v. Ramos, 596 S.W.2d 654, 657 (Tex.Civ.App.-Fort Worth 1980, writ ref'd n.r.e.). Under the transportation code, “ ‘Highway or street’ means the width between the boundary lines of a publicly maintained way any part of which is open to the public for vehicular travel.”7 Tex. Transp. Code Ann. § 541.302(5) (West 1999). “ ‘Private road or driveway’ means a privately owned way or place used for vehicular travel and used only by the owner and persons who have the owner’s express or implied permission.” Id. § 541.302(9). Finally, “ ‘Roadway’ means the portion of a highway, other than the berm or shoulder, that is improved, designed, or ordinarily used for vehicular travel.” Id. § 541.302(11).
The jurisdictional evidence included testimony from the University’s designated representative from its office of parking and transportation, David Kazmirski, who stated that the driveway was paved; that he had seen motor vehicles, bicycles, and pedestrians on the driveway; that there were no signs restricting cycling or pedestrian traffic; and that he would expect that people would ride their bicycles up and down this road. Kazmirski further stated that the drive connects to the Texas Exes Alumni Center and the east side of the ROTC building, and then continues to the parking area next to Gregory Gymnasium. He answered affirmatively when asked whether vehicles that are looking to park in the Gregory Gym parking lot come up that road, as well as vehicles going to the Texas Exes center and the ROTC building.
The record does not include any photographs of the area in question, but it can be discerned from the parties’ pleadings that the chain was stretched across the entire width of the throughway, as well as the sidewalks on either side of it. It is not clear, however, how the throughway or Winship Circle connect to San Jacinto Boulevard, the public street that Hayes turned onto when he entered the campus, or to any of the other paved streets that run through the campus. Nor is there any evidence or argument concerning whether the throughway is “publicly maintained,” apart from Hayes’s unchallenged assertion that the University is a public institution, financed by the taxpayers of the State.
As evidence that the driveway is functionally a “private road or driveway” because the University restricted the public’s access to it, the University emphasizes that the service driveway is within the perimeter of the guard shacks. Although there is some evidence that the guard shacks were manned during regular business hours, it does not appear that the University continues, after 5:45 p.m., to attempt to regulate which vehicles may enter campus.
On this evidence, a reasonable inference could be drawn that the throughway was open to the public, was ordinarily used for vehicle traffic, and was therefore a “highway, street, or roadway” as those terms have been defined in the transportation code. Kazmirski’s statements in particular suggest that, at least after 5:45 p.m., the University knew of and acceded to the use of campus roadways, including [891]*891the throughway in question, by vehicles without regard to whether they were operated by current University students and employees. The University bore the burden of conclusively establishing that the throughway where Hayes was injured was not a highway, street, or road in order to show as a matter of law that the chain draped across it could not have been a special defect; we conclude that it failed to meet its burden. See Miranda, 133 S.W.3d at 228 (holding that state is required to meet summary-judgment standard of proof). In light of the foregoing, we conclude that a fact question remains as to whether the throughway where Hayes was injured was a “highway, road, or street.” Because the jurisdictional inquiry with respect to Hayes’s special-defect claim may require farther factual development, we hold that the trial court did not err in denying the University’s plea to the jurisdiction on this ground.
In so holding, we recognize that the trier of fact might ultimately conclude that the chain was a mere premises defect, rather than a special defect. Nonetheless, even if this question is resolved in the University’s favor, such that Hayes was entitled only to the duty of care owed to a licensee, we conclude, for the reasons that follow, that the plea to the jurisdiction was properly denied on the basis that the University failed to conclusively establish that it discharged its duty to Hayes as a licensee.
Premises Defect: Duty to Warn of or Make Safe the Dangerous Condition
The University contends that even if Hayes was a licensee rather than a trespasser, it conclusively established that, by erecting the barricade, it warned of or made reasonably safe the dangerous condition posed by the chain. See Tex. Civ. Prac. & Rem.Code Ann. § 101.022; Payne, 838 S.W.2d at 237 (duty owed to licensee is not to injure him by willful, wanton, or grossly negligent conduct and to use ordinary care to warn or make reasonably safe dangerous condition of which property owner is aware and licensee is not).8
The University’s jurisdictional evidence pertaining to this issue consists solely of Hayes’s deposition testimony. Hayes testified that he saw “the reflective sawhorse barricade, orange and white colored, placed in the middle of [the] road,” blocking “about four feet of that section of the road.” He stated that there were “about ten feet on either side of the barricade that appeared to be clear.” He continued, “So without braking, without slowing down significantly, you know, [I] veered to the left-hand side of that barricade and then was stopped short by the chain.” Hayes stated that after the accident, he observed that the chain was suspended in line with the barricade, but placed just behind it. He described the chain as a rusty brown color “with no reflectors on it” and noted that there were no cones in the area.
Hayes also presented jurisdictional evidence, attaching to his response Kazmir-ski’s deposition testimony. Like Hayes, Kazmirski testified that the barricade did not span the width of the driveway, stating that the intent of the barricade was only to deter motor vehicle traffic. Because “we don’t expect the cyclists to disturb parking,” Kazmirski stated, the barricade was not in place to prevent cyclists or pedestrians from going around it. He further testified that it would be normal for the University to place two to four traffic cones to either side of the barricade in [892]*892order to warn of the presence of the chain,9 and that the University possessed reflective tape and metal reflectors that it could affix directly to the chains, making the chains visible at night.
We hold that the University did not conclusively prove that it used ordinary care to warn of or make safe the dangerous condition posed by the chain. A reasonable fact-finder could determine, contrary to the University’s contention, that the barricade was not put in place to warn of the existence of the chain, that it did not adequately signal the presence of the chain, and that it did not safeguard users of the driveway and sidewalk from the potential dangers that the chain posed.10 See TXI Operations, L.P. v. Perry, 278 S.W.3d 763 (Tex., 2009), available at http://www. supreme.eourts.state.tx.us/opinions/HTMLo-pinion.asp?OpinionID=2001330 (holding that posted speed-limit sign was merely “[a] ‘be careful’ warning” that “might be some evidence that the premises owner was not negligent,” but was not conclusive in situation where sign “was only a general instruction; it neither informed the driver of road hazards generally, nor did it identify the particular hazard” in question). We overrule the University’s second issue.
CONCLUSION
Because the University has not conclusively established that it retained its immunity under the Act, we affirm the trial court’s order denying the University’s amended plea to the jurisdiction.
DISSENTING OPINION