Mr. PRESIDING JUSTICE SULLIVAN

delivered the opinion of the court:

This is an appeal by plaintiff from summary judgments in favor of both defendants in a strict products liability action seeking damages because of personal injuries. In granting summary judgments, the trial court found as a matter of law that plaintiff assumed the risk of injury and that the product was not a proximate cause of his injuries.

The facts before the trial court were contained in excerpts from the depositions of occurrence witnesses which had been attached to the motions, memoranda of the parties, answers to interrogatories, and the affidavits of plaintiff and his expert witness. These documents revealed that injury occurred when the undercarriage of a moving railroad car crushed the legs of plaintiff, who was working underneath the car. The car had been manufactured by Pullman and was designed to transport products such as cement, which is loaded by cascading the material into the car through an opening on the top of the car and is unloaded by the force of gravity in a tunneling motion when the discharge gate at the bottom is opened.

For the most part, the car had a rectangular shape but, to facilitate unloading, a portion of the car’s floor was cut away and an undercarriage attached. The undercarriage consisted of two metal panels which were situated perpendicular to the ground and parallel to the tracks and two other panels which sloped toward each other as they approached the ground. This funnel configuration was cut off approximately eight inches from the ground by a discharge gate which was horizontal to the ground and slid back and forth on ledges to the opened and closed positions. When the gate was opened, the force of gravity and the funnel configuration of the undercarriage caused the cement to empty. Prior to loading, and throughout the car’s transportation, the gate must be closed. Plaintiff was injured while attempting to close such a gate preparatory to the loading of the car.

The record discloses that plaintiffs employer, Medusa, Inc., leased cars of the aforementioned type from Northwestern, including the car involved here, and that Northwestern crews delivered the cars to the edge of Medusa’s yard. Thereafter, the cars were moved by Medusa’s crews into its yard, using one of the two locomotives owned by it. To ready the cars for loading, a Medusa engine would push them into one of the seven tracks in the yard, where normally they would stand until workmen had closed the discharge gates.

Plaintiff and his immediate supervisors, foreman Huyett and assistant foreman Santos, outlined the usual practice in closing these gates. First, a man would stand next to the car and insert a four- to five-foot-long iron bar into a knuckle joint on one of the ledges which supported the discharge gate. Manipulation of the bar in the knuckle was supposed to close the gate. However, the men stated that very frequently this operation would not close the gate, in which event a man would climb under that portion of the car which was several feet above track level, and he would then face the undercarriage and apply direct impact force to the gate with a 30-40 pound bar (bump bar). Santos said that in 80-90% of the cars, the gates could only be closed by using the bump bar. Should this method fail to close the gate, the car would be rejected by the foreman unless the car was badly needed — in which event a man would climb inside of the car in an attempt, with a hammer and chisel, to clean the grooves of caked cement.

On the day of his injuries, plaintiff — with Huyett and Santos — was preparing railroad'cars for loading. At approximately 7:30 a.m., they approached a string of five cars which were standing at the northern end of track #3. The track curved away from the men at the southern end of the yard and this curve, together with the presence of cars on tracks #2 and #3, obscured their view of the southern end of track #3 so that they were unaware of the presence of one of Medusa’s locomotives on track #3. They set about preparing the five cars for loading. Huyett attached seals to the ladders of the cars while Santos attempted to close the discharge gate on the northernmost car by using the long bar to turn the knuckle, but he was unable to close the gate. Plaintiff then crawled under the car and struck the gate several times with the bump bar, when suddenly and without warning the car began to move. As the undercarriage, which was only about eight inches above ground, came towards him, he reached up to grip the metalwork of the car in an attempt to move with it before the undercarriage reached him. He was not successful, and it came in contact with his legs, causing serious injury to them. As soon as Huyett noticed the movement of the cars, he ran southward and signaled to the crew in the locomotive to stop. Neither plaintiff nor his supervisors saw or heard the locomotive before the movement of the cars. Plaintiff was able to crawl out from beneath the car but almost immediately lost consciousness.

It appears to have been the practice for a switchman to have set blocks to prevent movement of the cars while workmen were preparing them for loading, but this was not done on the date in question. Further, plaintiff and his supervisors stated they had no advance warning of the movement in question, either by a bell or whistle. The record also discloses that plaintiff knew of no occurrence in the past where a workman was injured while cleaning a car; that he had seen the engine push the cars onto track #3, but that once they were stopped he could no longer see the engine and did not see it at any time thereafter; that he believed the engine had left to get other cars; that he assumed the cars were ready for cleaning; that he asked his supervisors whether they wished work to be done on them; and that he walked to the cars with his supervisors to ready them for loading. Huyett also testified that when the cars were pushed onto a loading track, the practice was to uncouple the engine; that the switch crew would leave the cars on any available track to be cleaned, and they were cleaned wherever they had been left; that it was a practice to use chocks or steel derails while the cars were on the cleaning tracks, and it was the responsibility of the switchman to set them out; that the switch crew would check for the presence of workmen before recoupling the engine. Further, he testified that before the accident, plaintiff told him the cars had arrived and asked whether they were to be cleaned — to which he replied, “Yes, I’ll be right with you”; that he did not see the locomotive at the southern end of the track; that he heard a whistle and the tightening of couplings simultaneously with his observation of movement; that cars were to have been placed on track #1, but the switch to that track was frozen, and the cars were moved onto track #3.

Santos, plaintiffs assistant foreman, also testified that when the cars are pushed onto a track, the engine is generally uncoupled; that 80-90% of the gates on Northwestern cars could only be closed by use of the bump bar; that Huyett told him and plaintiff to clean the cars on track #3; that he tried to close the gate on the car in question but couldn’t and, when he told plaintiff that he needed help, the latter climbed under the car and that there was no warning before the car was moved. Frank Conrad, engineer of the train in question, stated that a verbal radio message from the foreman was the usual method for an engineer to leam that workmen were cleaning the cars; that he did not remember any other injury to any other employee trying to move the discharge gate with a bump bar; that lack of communication was the cause of the injury — that is, someone gave the wrong radio signal to the switch crew, so that they were unaware of the cleaning of the cars at the time of the movement.

Plaintiff’s expert witness, Jeremy Davis, stated that in his professional career he had examined more than a hundred different railroad cars for the purpose of evaluating design defects therein. In his experience, he has noted that cement has a tendency to accumulate and cake in the grooves of the ledges upon which the discharge gate was to slide. It was his opinion, based upon a reasonable degree of scientific certainty, that the manufacturer should have foreseen not only the occurrence of this condition but also that direct impact force to the gate would be necessary to close it. Indicative of Pullman’s prior knowledge that the cement would accumulate was a legend it affixed to the cars, which directed the user to clean the inside grooves when the gate was stuck. In addition, Davis stated that a design involving rollers would have actually eliminated the accumulation problem, so that the workmen could close the gate from the outside of the car without difficulty.

Plaintiff and his supervisors stated that they would climb into the cars to clean the grooves only as a last resort if Medusa needed cars very badly. In fact, they could not describe how the grooves were cleaned, except that plaintiff stated that to do so a workman would have to stand inside the car with his feet protruding through the open gate to the ground below.

Opinion

Plaintiff first contends that a genuine question of material fact existed on the issue of proximate cause and that the trial court erred in determining that issue as a matter of law. Section 57 of the Civil Practice Act (Ill. Rev. Stat. 1975, ch. 110, par. 57) provides for summary judgments, and subsection (3) thereof contains the following:

“The judgment or decree sought shall be rendered forthwith if the pleadings, depositions, and admissions on file, together with the affidavits, if any, show that there is no genuine issue as to any material fact and that the moving party is entitled to a judgment or decree as a matter of law.”

Summary judgment proceedings test for the existence of any genuine triable issue of fact. (Clausen v. Ed Fanning Chevrolet, Inc. (1972), 8 Ill. App. 3d 1053, 291 N.E.2d 202.) However, the right of the movant to summary judgment must be free from doubt, and the affidavit and attachments supporting the motion should be strictly construed and must leave no question as to the movant’s right to judgment. Contrarywise, the counteraffidavits should be liberally construed. (Hendricks v. Deterts (1973), 13 Ill. App. 3d 976, 301 N.E.2d 625.) Its purpose is not to try an issue of fact and, although inferences may be drawn from undisputed facts, the motion should not be granted unless those facts are susceptible of but a single inference. (Stanfield v. Medalist Industries, Inc. (1975), 34 Ill. App. 3d 635, 340 N.E.2d 276.) Where reasonable men could arrive at different results, questions cannot be determined as a matter of law. Stanfield.

Thus, we must first ascertain whether from the facts of this case reasonable men could arrive at diverse results as to whether plaintiffs injuries were proximately caused by the unreasonably dangerous condition of the discharge gate. If such results could be reached, then a triable issue of fact exists which should not be decided as a matter of law.

Solely to facilitate the trial court’s determination of their motions for summary judgment, defendants have stipulated that the slide discharge gate of the car involved was defective and unreasonably dangerous. However, they contended in the trial court, as they do here, that the defective condition was not a proximate cause of plaintiffs injury. Chicago and Northwestern asserts, “The defect, which prevented the discharge gate from closing, played no role in, or produced no other consequence independent of the car movement.” It takes the position that the movement of the car resulted from acts completely unrelated to the defect in the car and it argues that the defect was a condition, rather than a cause, referring to Merlo v. Public Service Co. (1942), 381 Ill. 300, 45 N.E.2d 665, where it was stated, at page 316:

* e * [ijf negligence charged does nothing more than furnish a condition by which the injury is made possible and that condition causes an injury by the subsequent, independent act of a third person, the creation of the condition is not the proximate cause of the injury where the subsequent act is an intervening efficient cause which breaks the causal connection between the original wrong and the injury, and itself becomes the proximate or immediate cause. The cause of an injury is that which actually produces it, while the occasion is that which provides an opportunity for the causal agencies to act.”

Pullman takes a slightly different course in reaching the same result. It argues that the defect did not inflict the injury but, rather, that it was the result of a force beyond its control — the movement of the train.

It is the rule that where other causes combine to produce injury, the causal connection between the defective product and the injury will be broken only if the acts or omissions of others were improbable or unforeseeable. (Lewis v. Stran Steel Corp. (1974), 57 Ill. 2d 94, 311 N.E.2d 128.) Foreseeability is usually a jury question but may be decided as a matter of law where the facts demonstrate that plaintiff could never be entitled to recovery. (Winnett v. Winnett (1974), 57 Ill. 2d 7, 310 N.E.2d 1.) Thus, whether there was causal connection between the defective condition and plaintiffs injury initially depends upon foreseeability by defendants that workmen might go under the car to clear the grooves. In this regard, by their stipulation and from the record, it appears that defendants knew cement dust would accumulate in the grooves, and they were aware that this would impede the closing of the gate from outside the car. Of course, they also knew that the undercarriage had a clearance of only about eight inches. An alternate design used rollers which apparently would have avoided the accumulation problem so that the gates could more readily be closed from the outside. Rollers were not included in the design of the car involved, but Pullman did affix a legend which stated, “Clean inside grooves.” However, even if defendant’s suggested method of closing the gates had been pursued; that is, that the workmen climb inside the car to clean the grooves, it appears from an uncontroverted statement of plaintiff in his affidavit that in order to clean from inside the car, the workman’s legs and feet would still have extended through the undercarriage onto the right-of-way and thus would be about eight inches below the bottom of the undercarriage. In case of movement of the car, it seems to us that a workman in this position would be in danger of injury similar to that experienced by plaintiff.

In Wilson v. American Chain and Cable Co. (3d Cir. 1966), 364 F.2d 558, 562, the court after stating that a negligent act, such as the movement of the car in the instant case, will not relieve a defendant of liability unless it was so extraordinary as not to have been reasonably foreseeable, then went on to say:

“ * * * [W]hether the act was reasonably foreseeable [is] to be determined by following retrospectively the sequence of events and looking back from the harm to the negligent act rather than by considering whether the defendant should prospectively have envisaged the events which unfolded and caused the accident.”

Significant on the question of foreseeability is the recent case of Troszynski v. Commonwealth Edison Company (1976), 42 ill. App. 3d 925,931,356 N.E.2d 926,931; appeal denied (January 28,1977), where the court held it to be objectively reasonable that defendant should have foreseen a person could come into contact with exposed wires in a meter box, where the facts were set forth in the opinion as follows:

“The evidence revealed that a middle-aged man of average intelligence and experience thrust his hand into a meterbox to retrieve a piece of glass accidentally broken by his son while throwing a tennis ball. The entire apparatus was constructed so that once the glass was gone a dangerous uninsulated electrical wire lay exposed inches from the window. The cover contained no warning that this dangerous situation existed within the box. The reader merely was admonished not to break the seal as no fuses were contained within. The box itself was held shut only by a wire hasp which could be opened by anyone in possession of a pair of wire cutters.”

We note also that, here, plaintiffs expert witness testified defendants should have foreseen the caking of cement in the grooves of the ledges upon which the discharge gate was to slide, and he also gave his opinion that direct impact force to the gate would be necessary to close it under those circumstances.

Thus, we think reasonable persons, having knowledge that accumulations in the grooves would prevent the gates from closing, would reach different opinions as to whether it was foreseeable that workmen, f aped with inoperable gates, would find it quicker and perhaps more practical to climb underneath the car to clean the grooves. Further, it is unquestioned in the record that plaintiff was under the car solely because of the inability to close the discharge gate from outside the car, and that this failure was the result of the stipulated defective condition of the gate. In view of our belief that a question of fact exists as to whether defendants could have foreseen that workmen would go under the car to force open the gates, we think it follows naturally that reasonable persons would also arrive at different results on the question of the foreseeability of movement of the car while a workman was under it. In any event, we cannot say as a matter of law from the record presented here that the act and omissions of those responsible for the movement in the instant case were unforeseeable by defendants. For the reasons stated, it is our opinion that a question of fact exists as to whether there was causal connection between the defective condition of the car involved here and the injury to plaintiff.

Plaintiff also contends that summary judgment was improperly granted on the theory that he assumed the risk of his injury. Assumption of risk, which is a defense to strict liability, involves conscious exposure to danger, a deliberate decision to encounter a known risk, or the willingness to take a chance and is measured subjectively. (Sweeney v. Matthews (1968), 94 Ill. App. 2d 6, 236 N.E.2d 439, aff'd (1970), 46 Ill. 2d 64, 264 N.E.2d 170.) In Williams v. Brown Manufacturing Co. (1970), 45 Ill. 2d 418, 430, 261 N.E.2d 305, 312, the court stated:

“ ° ° ° [Wjhether a user has assumed the risk of using a product known to be dangerously defective is fundamentally a subjective test, in the sense that it is his knowledge, understanding and appreciation of the danger which must be assessed, rather than that of the reasonably prudent person [citation]. [I]t must also be remembered that this is ordinarily a question to be determined by the jury.”

The Williams court went on to state that the determination of assumption of risk “is not to be solely on the basis of the user’s own statements,” but should include a consideration of “the user’s age, experience, knowledge and understanding as well as the obviousness of the defect and the danger it poses 6 *

It is the position of Pullman that no risk of injury resulted from the defective condition of the car, and it contends that the danger of being beneath the car in the event it was moved was obvious and apparent— particularly where no precautions were taken beforehand. Northwestern urges that plaintiff assumed the risk of his injury by trying to clear the gate from underneath the car when he knew that the proper procedure was to go down into the car to do so, and by going beneath the car when he could not see whether an engine was hooked to it and when he did not otherwise determine whether the car could be moved, it urges that he made a deliberate decision to encounter a known risk.

Assumption of risk, as used in Comment n to the Restatement (Second) of Torts § 402A (1965), is referred to as “ " # * voluntarily and unreasonably proceeding to encounter a known danger ° Whether the character of plaintiff’s conduct is such that he will be considered to have voluntarily and unreasonably incurred a known risk must necessarily depend upon the particular circumstances of each case. Scott v. Dreis & Krump Mfg. Co. (1975), 26 Ill. App. 3d 971, 326 N.E.2d 74.

The danger here, as contended by defendants, was in the act of plaintiff going beneath the car without ascertaining whether it could be moved. The record indicates that plaintiff acted voluntarily, and it appears that he should have known that the car might be moved. There remains the crucial question, however, as to whether he acted unreasonably in going under the car and, from our examination of the record, it appears to us that the matrix of logic does not compel the acceptance of the conclusion that the facts here are susceptible of the single inference that plaintiff unreasonably encountered a known danger. The documents presented indicate that he knew or reasonably could have assumed that the car would not be moved because (1) track #3 was used regularly for cleaning cars; (2) when cars were placed on such a track, it meant they were ready for cleaning, and the usual practice was to uncouple the engine; (3) after an engine was uncoupled, it would leave to work elsewhere; (4) his employer used only one engine at a time; (5) before an engine was coupled to cars on the cleaning tracks, the switch crew was required to check for the presence of workmen; (6) one or more of the cars on track #3 was chocked, as it was the responsibility of the switch crew to chock the cars after they were placed on a cleaning track; (7) he was told by his supervisor to clean the cars on that track (testimony of Santos); (8) in the past, most of the cars (80-90%) could only be cleaned by going underneath and using a bump bar; (9) no prior injury had occurred to a workman cleaning a gate from beneath a car; and (10) a bell or whistle would be sounded to warn him before cars were moved, and no such warning was given at the time of the occurrence in question.

Furthermore, because the accident in question included a combination of unusual circumstances, the unreasonableness of plaintiff*s conduct is further in doubt. The cars were to have been placed on track #1 for cleaning, in which event the engine would have been uncoupled and left; however, because the switch to track #1 was frozen, the cars were temporarily placed on track #3 until the switch to track #1 was cleared. When that switch was opened, the engine started but no whistle was blown and the bell malfunctioned. In addition, the engineer stated his unawareness of the cleaning of the car at the time of the movement, because someone gave the wrong radio signal to the switch crew.

It appears to us that defendants ignored a vital aspect of the law on assumption of risk. In addition to realizing that existence of the defect or danger in voluntarily exposing himself to it, plaintiff must perceive and appreciate the risk involved. (See Prosser, Torts § 67 (3d ed. 1964).) In other words, although plaintiff knew that the car might be moved after he voluntarily climbed under the car, under the circumstances he may not have appreciated the unreasonable character of the risk.

Section 496D of the Restatement (Second) states this sometimes overlooked contour of assumption of risk as follows:

“Except where he expressly so agrees, a plaintiff does not assume a risk of harm arising from the defendant’s conduct unless he then knows of the existence of the risk and appreciates its unreasonable character.”

Moreover, as Comment (c) explains, the question is not whether he should have realized the risk but whether, in fact, he did realize it:

“The standard to be applied is a subjective one, of what the particular plaintiff in fact sees, knows, understands and appreciates.”

Here, while it is true that plaintiff climbed under the car without ascertaining whether a locomotive was attached to the cars on the track or whether any of those cars were chocked, we must focus on what he actually knew, understood and appreciated. In this regard, we see that for many years he and other workmen had been closing gates by using impact force from under cars and considering, as we have pointed out above, that it was not unreasonable for him to assume that the car under which he was working would not be moved without some warning to him, we believe that the facts fall far short of the subjective perception and appreciation of the particular risk of harm that eventually resulted.

We are cognizant that summary judgment plays a necessary and crucial role in civil litigation and that the courts have looked upon granting it as a salutatory development in the administration of justice. (Pawlowski v. Damall (1973), 15 Ill. App. 3d 390, 304 N.E.2d 478.) Nevertheless, considering the totality of the evidence presented in its aspect most favorable to plaintiff, as we are required to do (Williams; Scott), we are of the opinion that a question of fact exists on the issues of causation and assumption of risk and that summary judgment should not have been granted. The judgment is reversed, and the case is remanded for further proceedings.

Reversed and remanded.

LORENZ and MEJDA, JJ., concur.