It happens too many times. You send someone on an errand, a simple one, and the person gets it wrong. Maybe you are trying to repair a coffee mug, and you ask your spouse to add glue to the shopping list. Your spouse comes back with a glue that does not work on ceramic, and when you try to explain why it is not satisfactory, he or she answers, “You didn’t say anything about a mug. You asked for glue, and that’s what I bought. Was I supposed to read your mind?” Or you ask a child to clean up a room, and he picks up everything that was on the floor so it can be vacuumed. You return and complain that you have company coming in five minutes and the room is a mess; there isn’t even a chair to sit on, and he says, “I did what you asked. You didn’t say anything about company. Was I supposed to read your mind?”
The answer is yes. Whenever we make a request—ask for an errand or give a command—we need the person to read our mind. To make this possible, both parties have to extend themselves. The person making the request can help by specifying the intent behind the request. The person trying to carry out the request has to imagine what the other person really wants, to handle all the details that did not get explained.
“Please pick up some glue at the hardware store since you’ll be going to shop next door anyway” does not sound like a difficult request. But when you walk into the hardware store, you find that there are different types of glues: for wood, for glass, for metal. Some are advertised as superglues that will never fail. Is that good or bad? Might I want to dissolve the glue later? So I really should tell you why I wanted the glue: to repair a ceramic mug, my favorite one. The handle has come off, and it has a small crack, but I believe it can be salvaged as a home for a small plant and stay around like an old friend. You are moved by my devotion to pottery, and agree. Now that you know my intent, you can select the right type of glue and the right amount—the smallest tube the store sells. But wait. Is that economical? How much ceramic glue would I need in the coming months, and what are the chances that this tube will dry up or get lost? You would have to know me and my glue-using habits to make this choice. Perhaps I should have told you the size, but it never occurred to me, and besides, I do not know what sizes the tubes come in.
Then you run into another choice: quick drying or not? That depends on how the mug was broken, and if it was a clean break, whether I have a vise for clamping the pieces together. Is it better to leave me holding the pieces together for hours? Yet if you get me the quick-drying kind and I clamp the pieces together incorrectly, it could be very hard, or impossible, to pull them apart and reconnect them. You also need to know something about my manual dexterity. This is a tougher job than it seemed. You may find yourself using mental simulation to imagine my using the different kinds of glues. The better you know me, the better you will be able to carry out the errand successfully, because you will be able to read my mind and take into account all the things I did not say.
Usually we cannot specify all these details in advance. If you are doing a favor for me, I have to rely on your ability to read my mind and imagine how I would have made all the choices. I do not believe I could anticipate all the relevant details to tell you. There are some people I do not trust, and I do not give them difficult errands.
Being able to read someone’s mind is an important ability for working with others. In example 13.1, the decision makers chose to follow their orders literally, without trying to read the mind of the person who issued the orders.1
The Goeben was a German battleship stationed in the Mediterranean Sea at the start of World War I. The British Navy was supposed to find and destroy it once war was officially declared. The British failed; they had the Goeben surrounded but became confused about the intentions of the admiralty, so they let the Goeben escape through the Dardanelles and into the Black Sea, where it forced the Ottoman government to enter the war on the side of Germany (which resulted in the loss of its empire and the creation of the British Mandate in the Near East). The Goeben also bottled up 95 percent of the Russian shipping (since their only warm water ports were on the Black Sea), helping to create the hardships that led to the Russian Revolution.
In the days preceding the war, Britain knew that it would have to eliminate the Goeben, the only German threat in the Mediterranean. The British ships kept a continual search for it and trailed it as much as possible. They were waiting for the beginning of the war, when they would be permitted to sink the Goeben.
Once war was declared by England, at 11:00 P.M. August 4, 1914, the Goeben managed to stay out of sight for a few days. Eventually it was located and surrounded by a dozen British ships. Unfortunately, the commander of these forces was unsure how to interpret the directions sent on July 31 by Winston Churchill, the head of the Admiralty. The instructions to hunt down the Goeben included the direction, “… do not at this stage be brought to action against superior forces.”
Churchill wanted the Goeben destroyed. He also was aware of the heavy Austrian battle wagons and wanted the British to be careful with these, hence the warning not to become engaged against superior forces. Churchill’s intent in this warning was for his ships to steer clear of the Austrians. However, Rear Admiral Sir Ernest Troubridge, the commander of the ships surrounding the Goeben, felt that the sentence could also be applied to his own situation. Technically, the Goeben was larger than any single one of his own ships and had larger cannon. He debated what to do and decided that the safest course was to regroup rather than run any risks. By letting the Goeben pass, it escaped east, into the Black Sea.
When Churchill heard about what had happened, he was staggered. He had never intended for his message to be interpreted the way that it was.
The commander was unable to read Winston Churchill’s mind, to fathom the intent behind the sentence requesting care in engaging superior forces. In hindsight, Churchill could have avoided the error if he had added more information about the Austrian ships. His eagerness to attack the Goeben should have been sufficient. If he had tried to clarify every point in every order, to prevent any ambiguity, the job would have been impossible. The resulting documentation would have tied up the communications lines, and the important points would have been submerged in the flood of details and clarifications.
The answer is not to pile on the details. That takes too long, and it has its own costs. We may want to pretend that rules and procedures are simple and but they are not. For example, if you give someone the direction to press a button when a green light comes on and the person asks, “What is green?” there is no way to respond. You could explain “green” is light at a wavelength of 530 to 570 millimicrons, but that information is unlikely to be helpful. We assume that living in a shared culture will provide us a basis of common referents. If we have to labor at breaking out all of the assumptions behind requests, teamwork and cooperation would become almost impossible.
One place where we can see the importance of giving and understanding requests is in flying airplanes. When each member of an aircrew does not have the same understanding of what is intended, the results can be tragic. Many taped conversations recovered after crashes document the failure of different members of the aircrew to understand what others want to do. We had a chance to observe this ourselves in a study we did for NASA, using actual aircrews flying missions in the NASA 727 full flight simulator. This was a project to observe team coordination and decision making during difficult conditions. Marvin Thordsen and I watched several aircrews react to simulated malfunctions. The crews were composed of a captain, copilot, and flight engineer.
One of the malfunctions is an unusual fuel leak from the number 3 tank, in the right wing. We watch three crews, and in two of them the captain and the flight engineer disagree.
In addition to the obvious problem of not being able to fly as far because of fuel loss, the captains are concerned about a secondary implication, loss of balance. Since the tank is in the right wing, that wing will get lighter as the fuel leaks out. Therefore, the plane will be difficult to control during landing. In fact, this is the more serious problem since the airplane has sufficient fuel to divert to nearby airports.
In each case, the captains tell the flight engineers to reconfigure the fuel flow so that all engines are fed from the good tank, in the left wing. The intent is to draw down this fuel supply, lighten up the other wing, and restore balance. Nevertheless, in each case, the flight engineers resist or misunderstand the request. Since the flight engineers are responsible only for monitoring and managing the fuel flow and not for handling the controls, they are more worried about a shortage of fuel. In one case, the captain has to swivel completely around in his chair (giving up the task of flying the plane to the copilot) to check on what the flight engineer is doing. He sees that the flight engineer had not carried out his request and is still using the fuel from the number 3 tank. This configuration is not what he wants. The flight engineer explains how this will make sure they do not waste any fuel. The captain is unable to persuade the flight engineer to make the changes until the copilot comes to his rescue. Finally, the flight engineer bows to their wishes. When the airplane lands, the imbalance in weight is 2,000 pounds between the two wings. In a second flight, with a different crew, the captain is never able to explain his wishes to the flight engineer. This airplane lands with a 5,000 pound imbalance. The recommended maximum imbalance for safe flying is officially set at 1,000 pounds.
The problem was not that the flight engineers were simply mishearing what the captains were requesting. Rather, they were not grasping what the captains wanted to do. The captains asked for a reconfiguration and thought the reason would be obvious. The flight engineers reinterpreted the request into something that made more sense to them. If the one captain had not turned to look back, he may never have known what the flight engineer was doing and would not have realized he was being misunderstood. The flight engineer was unable to figure out that the captain was asking for an unusual reconfiguration; he could not read the captain’s mind.
If we can work with people who understand the culture, the task, and what we are trying to accomplish, then we can trust them to read our minds and fill in unspecified details. A team that has much experience working together can outperform a newly assembled team.
A study was once done for NASA (Foushee, Lauber, Baetge, & Acomb, 1986) to investigate the effect of fatigue on performance. Flight crews were asked to fly eight-hour missions in a high-fidelity simulator. The researchers presented the same malfunctions at the beginning of the flights for some crews and at the end for others. They expected that the crews would react better at the beginning, when they were fresh. The results were the opposite: the crew members did better at the end of their missions. Their advantage was all they had learned about working together. They learned to anticipate how each member would react, and they became adept at reading each other’s mind.
Flight crews are not the only teams to wrestle with the problem of guessing intent.2
During an operation, the surgeon decides to lower the patient’s blood pressure. He directs the anesthesiologist to give the patient a drug that will have this effect, but does not explain what he is trying to accomplish. The anesthesiologist gives the drug, notes that the patient’s blood pressure goes down, and boosts the level of another drug that will increase the blood pressure. To the anesthesiologist, this is standard operating procedure to keep the patient’s vital signs stable. The surgeon notes that the blood pressure is higher than he wants and directs the anesthesiologist to increase the dosage of the first drug. The anesthesiologist follows the request, watches for the blood pressure to reduce, and then boosts the drug that will return the blood pressure to its normal level. This cycle continues until the patient ends the game by dying.
When I am at work, no one does a better job of reading my mind than Barbara Law, who has worked with me since 1979, longer than any of my other colleagues. No matter who else has reviewed an article or manuscript, I am uneasy unless she has had a chance to go through it.
I complete a draft of a paper on how to communicate intent and give it to Barbara to edit before we send it out. She notices that I am citing some data from another study we have recently finished. Just to be safe, she goes into the files to make sure all the numbers match. She finds one or two discrepancies, and since I am not around, she brings them to George Kaempf, the lead author of that study, to clarify. She knows that I want to send the draft to Jim Banks, one of our sponsors at the Army Research Institute Field Unit in Monterey. She also knows that Jim had recently seen George Kaempf’s paper. She wants to make sure Jim does not find any mismatches between the two. When George cannot resolve one of the discrepancies, Barbara takes the initiative of holding up the draft. She reasons that it is more important to get the details straight than to rush out my manuscript.
Now compare this example to another incident that took place in my company.
One of Marvin Thordsen’s jobs is to keep our computer systems working. One day there is a malfunction in the mouse used to operate one of our more powerful systems, and it cannot be repaired. Marvin sends in a request to purchase a replacement. To make sure the request is perfectly clear, he tracks down when the original mouse had been ordered and writes that he wants the exact same mouse, and he even includes the date on the earlier purchase order for reference. He believes he has covered all bases. He has done a careful and thorough job. There should be no ambiguity.
To Marvin’s surprise, the replacement mouse doesn’t work; it doesn’t even fit. Somehow the front office has ordered the wrong one. In tracking down the reason, Marvin finds that the hardware company no longer makes the original mouse.
The front office had assumed that Marvin was trying to indicate the company he preferred to order from. It contacted that company and ordered the mouse that was closest in price to the original. They were not aware that there was a compatibility problem, that not all mice fit all machines. Besides, Marvin was traveling when they ordered the mouse, so they couldn’t ask him, and he seemed to want the replacement quickly. They wanted to show how responsive they were.
Here the front office tried to read Marvin’s mind, and failed. They thought they knew what he was after: an inexpensive model from the same company. They did not have the experience to know about hardware compatibility issues. Marvin knew they lacked computer hardware experience, and that was why he tried to make the task easy by digging up the old invoice with the part number. In hindsight, he realized that he should have explicitly told them he needed an inexpensive mouse that he could use on that specific computer model. That way they would have known more about his intent. They could have improvised better when the original mouse was unavailable. By telling the front office exactly what to do but not giving the rationale, he was leaving them vulnerable when the original plan fell apart.
In contrast, look at example 13.4. I gave Barbara Law virtually no instructions other than to review my manuscript. Because of her experience, she could detect a potential problem (the mismatch in numbers), see the bigger picture (a friend and sponsor might also notice the mismatch), anticipate the impact (slight loss of credibility), balance the priorities (fixing the problem versus getting the draft out quickly), and take the initiative to decide on holding up the draft. She could do all these things because she knows my priorities and the things that have bothered me in the past. She could therefore anticipate what I really wanted here.
On a team, a number of things can happen when there is a shared understanding of what the team is intending to do. Table 13.1 lists some outcomes we have seen when the people on the team understand its intent, in contrast to situations where people are told what to do without understanding why. When you communicate intent, you are letting the other team members operate more independently and improvise as necessary. You are giving them a basis for reading your mind more accurately.
Table 13.1 Functions of Communicating Intent
| Promote independence Improve team performance by reducing need for clarifications Detect deviations from the assumptions made by the leader Catch errors in advance and anticipate problems Promote improvisation React to local conditions without having to wait for permission Recognize opportunities that were not part of the plan Set priorities in order to make trade-off decisions Continue beyond the outcome without having to wait for the next order |
One important function is to increase independence so that team members need less attention and monitoring. They can evaluate whether they are making progress. If you just tell me a sequence of steps to carry out, I am always wondering if I am doing them right (“Is this what you wanted?”) and waiting for permission to start the next step. When you have asked me to do something and we both know the goal, I should be able to perform at a higher level. I should be better able to tell if I am making a mistake. Additionally, I can make better use of my experience to notice errors in your planning, along with problems that might arise.
If I run into problems while carrying out the task, I may warn you that the goals were wrong.
The other primary function of communicating intent is to allow better improvisation. We do not want everyone on a team freelancing. That becomes chaotic. Remember how the front office improvised and ordered the wrong mouse. Still, we must accept that few of us can think out all the contingencies in advance, and unless we want to direct every step our teammates take, we are going to have to cede responsibilities, including judgments about how and when to carry out critical tasks. When team members understand the intent and reasoning behind a task, they will be better able to improvise. They will adjust to the field conditions that the planners cannot know about, by finding ways to jury-rig solutions when the plan starts to run into trouble. They will recognize opportunities no one expected. They should be able to understand the goals well enough to set and revise priorities, to decide when to grasp an opportunity and when to let it go. And if they complete a task before receiving more instructions, they will be able to proceed on to the next task rather than wait.
Larry Shattuck (1995), a U. S. Army Lieutenant Colonel currently heading the human factors program at West Point, did a study on Commander’s Intent statements, used to explain the purpose for carrying out a plan. He presented military operations plans to battalion commanders serving in the field. Each battalion commander read and interpreted the plan and communicated his intent to his own company commanders using a variety of methods, including a Commander’s Intent statement, operations orders, and briefings. They relied on the standard procedures that they had been using together for months. Then Shattuck interviewed the battalion commanders to find out how they wanted their company commanders to react to an unexpected development in the scenario. Next, Shattuck videotaped company commanders as they described how they would react. The company commanders’ responses matched those of their battalion commander only 34 percent of the time. When Shattuck told the battalion commanders how their subordinates had adjusted, a common response was, “Why did he do that?”
There have been some documented examples of organizations that have worked to create independence and improvisation. According to the military historian Trevor Dupuy (1977), the Prussian Army created such an organization following their lackluster showings against Napoleon. They regrouped and established a professional force that showed its superiority in war during the next century. Despite what we have heard about German officers excusing themselves by claiming they were only following orders, the reality is quite different. German officers were expected to use independent thinking. The culture of the German Army is illustrated by a famous story described by Dupuy, about the strategy of mission tactics used by Prussian military leaders in the late nineteenth century.
A major, receiving a tongue lashing from Prince Frederick Charles for a tactical blunder, offered the excuse that he had been obeying orders. He reminded the prince that a Prussian officer was taught that an order from a superior was tantamount to an order from the king. Frederick Charles promptly responded, “His Majesty made you a major because he believed you would know when not to obey his orders.”
It is easy to say we want to encourage improvisation and initiative and to make sure that people understand why they have been given certain assignments. In reality, this practice turns out to be difficult, because it means that the people at the higher echelons must give up some of their control. The U.S. Army has sought to accomplish this, with mixed results.
A technique the U.S. Army has tried is to issue Commander’s Intent statements along with its mission orders, or operations orders, which give the details of the plan for the next day. The Commander’s Intent statement helps the soldiers read the commander’s mind if they run into uncertainty about how to carry out the orders under field conditions.
Some research on how well the Commander’s Intent statements work has centered on the National Training Center in the high desert of California. Here, the U.S. Army mounts realistic training exercises using sophisticated laser and computer systems to keep track of opposing forces during battle. A brigade commander might drill his troops for months to prepare them for the National Training Center. Once they arrive, they face a nonstop series of battles for almost two weeks straight, learning to function despite fatigue and stress. Some observers have attributed the army’s success in Operation Desert Storm to experiences gained at the National Training Center.
In one study of field exercises at the center, William Crain (1990) found that only 19 percent of the Commander’s Intent statements said anything about the purpose of the mission and that communication of intent was mediocre. We have found the same thing. George Kaempf’s study investigating the Commander’s Intent statements at the National Training Center found great variation.3 The shortest statement was all of 21 words, and the longest in our sample was 484 words. We asked domain experts to evaluate the effectiveness of a set of thirty-five statements. The average rating was below the midpoint of the scale, closer to very ineffective than to very effective.
A large part of this problem stems from the mysterious nature of intent itself. What does it mean to understand someone else’s intent? If you could interrogate the person giving the orders, what would you ask? What knowledge do we need to understand what someone else wants?4
In observing teams and reviewing their attempts to communicate goals, I have identified a few types of information that are important for describing intent (Klein, 1994). There are seven types of information that a person could present to help the people receiving the request to understand what to do:
All seven types of information are not always necessary. Instead, this list can be used as a checklist, to determine if there are any more details to add. In my company, whenever we begin a new project, we go through the relevant items in the checklist. We try to make sure that everyone working on the project has the same understanding of what we are after.
The first facet is the most obvious: to provide the big picture of why the task is being performed in the first place. When I reviewed the thirty-five Commander’s Intent statements that George Kaempf and I had analyzed, this facet appeared in only fourteen of the statements, usually with just a single comment. This was the facet Crain found only 19 percent of the time. This was the facet Marvin Thordsen did not provide in example 13.5, when he received the wrong mouse. This was the facet Winston Churchill did not offer when he said that British ships were to avoid superior forces.
The second facet presents an image of what the successfully completed request will look like. This is almost always included. In my review of the thirty-five Commander’s Intent statements, it was missing only once. But in about a third of the statements, the facet was described with the bare minimum, for example, “This is a defensive mission to defend at Phase Line Cigar.” It would be like my sending you to the hardware store with the goal of buying me some glue. Your chances of success will go up if I give you a better picture of success: “I expect you to return with a small tube of ceramic glue sometime this afternoon.” This image, together with the contents of the first facet, “so I can repair my coffee mug this evening,” should do the trick.
The third facet is the plan—another place where people run into trouble. They try to pile on the details of how to carry out the assignment rather than giving the big picture. This was where Marvin went wrong in ordering the mouse. In the thirty-five Commander’s Intent statements I reviewed, this facet got the most attention. It was included in thirty-two of the thirty-five statements and was the subject of about 40 percent of the comments. The commanders were telling their troops what steps to take rather than letting the troops understand what was in the minds of the commanders and the planning staff officers. Colonel Hanan Schwartz of the Israeli Defense Forces has argued that the plan should be left to the discretion of the unit carrying it out. Let them figure out how to achieve the mission. If you do not trust them, get others or do a better job of training. Just do not fall into the trap of choreographing each of their movements.
The fourth facet is the reasoning behind the plan. Here is where you can explain the rationale for the plan and even describe the thinking behind drawing up the plan. This is another opportunity to let someone see inside your head, to carry that knowledge along when putting the plan into action. In the formal Commander’s Intent statements, this information was given thirty-one times in the thirty-five examples, usually in combination with 3, the plan itself. If I am sending you to the store for glue, I might explain that my plan is for you to save effort by first doing your grocery shopping since you might find the glue in the hardware section of the supermarket. If that does not work, you can go to the hardware store. Notice that by telling you the rationale, I am giving you a chance to suggest a different plan. You might prefer to pick up the glue first, since it is light and nonperishable, and then go buy your groceries. It does not make sense to buy the groceries first and then carry all the sacks and the frozen items with you into the hardware store.
The fifth facet gets you ready to make key decisions. There might be contingencies to expect: “Sometimes the hardware store gets crowded; if that happens, go to the supermarket first.” There might be priorities to set: “If the hardware store is still jammed when you come out of the supermarket and it looks as if you’ll have to wait fifteen or twenty minutes, skip it. It’s no big deal. I’ll run over there myself tonight.” I found that thirteen of the thirty-five Commander’s Intent statement included some information about key decisions and priorities. Colonel Schwartz has suggested that this could also describe known weaknesses in the plan, to help subordinates know the thinking of the planners and recognize when the plan might be breaking down. None of the statements I reviewed mentioned any weaknesses in the plans.
The sixth facet refers to antigoals, outcomes you do not want to happen. Information is meaningful only in the context of alternatives. If I tell you my intent is to win this battle, that does not add any real information.
Antigoals come in when there are meaningful alternatives to clarify.5 In one of the Commander’s Intent statements, the mission was to delay an enemy advance. The plan was to use artillery to slow the enemy down, and the commander added, “Do not become decisively engaged.” He realized that during a defensive mission, it is easy to take a stand and get into direct combat, and he did not want that to happen. He felt that there was enough of a chance of his troops getting into a battle that it was worth telling them what he did not want. We cannot detail every outcome we do not want to see. Nevertheless, there are times when it can pay off to clarify what is intended by saying what is not intended. The facet of antigoals was added in five of the thirty-five statements.
The seventh facet includes constraints that should be taken into account and extra information—little observations such as “watch out for this” and “you might try to use that.” In sending off a rookie driver on an errand, we might add, “There is a chance it will start raining. We haven’t had rain in awhile, so the roads might get slippery. Be careful.” In the Commander’s Intent statements, these observations were usually about the terrain and about weather.
Besides these seven facets, I was tempted to add some more to cover time and resources. It seemed that a good description of intent should include an instruction about when the task should be completed and at what cost (e.g., dollars, hours, casualties). Hanan Schwartz convinced me to abandon this idea. He explained that a commander or leader who specifies time and resources is micromanaging. The effective commander needs to provide a sufficiently clear picture of the overall mission, including the potential follow-up. This information lets the subordinates make their own decisions about how best to achieve the task and the higherlevel goals. There may be times to adhere to strict timetables and budgets. More often, in chaotic natural settings, schedules will slide, and resources will shrink or expand over the life of a project or mission. Decision makers who believe they can pursue the goals within the initial constraints are too inflexible. You want them listening to what is happening all over so they can adapt, slow down or speed up as is needed. Since one purpose of a Commander’s Intent statement is to allow flexibility and improvisation, you do not want to restrict flexibility to adapt to changing conditions.
Karl Weick (1983) has described a streamlined version of a Commander’s Intent statement. In Weick’s version, these are five facets:
The Commander’s Intent statement is representative of any organizational setting where someone has to describe to others what they are supposed to be doing. Industry, educational settings, health care—all depend on the ability to communicate goals to others and to request clarification of goals where needed.
The art of describing your intent is to give as little information as you can. The more details you pack in, the more you obscure your main points. However, if you leave out an important consideration, you run the risk that the person will become confused at a critical decision point.
In framing the description of your intent, you will probably take into account the expertise of the person or team, the stability of the situation, and your ability to imagine what the outcome might look like.
With more experienced team members, you can concentrate on higherlevel goals. If they are less experienced, you have to specify more of the small steps in the plan. In a fluid and unstable situation, you might not want to state any antigoals because things might shift around, and you might change your mind. When your goals are ill defined (you are not clear what outcome you seek), you might give more attention to the key decision points because your image of the outcome might change depending on how the plan goes.
In teams that work together well, the people carrying out the assignments seem to be able to imagine what the leaders and planners were thinking. The people in the field can carry out the plan without needing continual direction from the higher authorities. The better they can do this, the more decisively they can act. It also helps to have common experience, to have worked with the leader enough to anticipate his or her reactions.
The concept of intent can be applied to equipment as well as to people. Particularly in working with sophisticated computer equipment, we struggle to figure out what the machine is trying to do. For example, as commercial airliners become more technologically sophisticated, they include a range of decision support systems. One of these is the flight management system, which helps track the course that an airplane takes from the time it takes off until it lands. This replaces the autopilot and adds to the computerized capability of keeping the plane on course at the correct speed, vectored in the right direction. George Kaempf managed a project funded by NASA in which we studied Flight Management Systems (Kaempf, Klein, & Thordsen, 1991). We found that computer systems also need to communicate intent.
An airplane is on a routine flight from the West Coast to the East Coast. It is a red eye special and is flying above 30,000 feet. It is 3:30 A.M.
A company employee riding in the jumpseat of the cockpit kicks the rudder control blade by mistake, so that the rudder deflects to an extreme position. At the time of this incident, the switch was located close to the floor, behind a pedestal and thereby shielded from view. No one sees that the control has been displaced, and the flight management system does not notify the crew.
Up to this point, the flight is routine. When the rudder control blade is kicked, the flight management system responds by compensating with other flight controls to keep the aircraft in straight and level flight. The crew notices no change, and the flight management system does not signal to the crew that anything unusual has happened. Because the rudder control blade has been left at an inappropriate setting, the flight management system continues to compensate.
When the flight management system reaches its limit, it gives up. It turns off, handing the controls back to the unsuspecting pilots, with the aircraft in an out-of-tolerance condition. Without warning, the aircraft stalls and begins to plummet. The crew first believes it has an engine problem. To regain control, the crew takes a number of ineffective actions that only make the problem worse. The airplane is falling and increasing in speed.
The story has a happy ending. The pilots are able to wrestle the plane back under control, pulling it out of the steep dive. They eventually figure out what has gone wrong and reset the rudder. But after that, it is hard for the pilots to trust the flight management system fully again.
One problem with the flight management system was that the pilots could not anticipate what the system was going to do. According to Earl Wiener (1989), the questions regarding automated systems most often heard during nonroutine events are: “What is it doing? Why is it doing that? What is it going to do next?” The system was doing a poor job of describing what it was trying to do. The challenge for designers is to figure out how to let systems like these signal their intent. We do not want them spewing forth endless data about every goal and subgoal. These systems have to be able to identify the appropriate times to convey intent and the appropriate level and format for doing it. Only then will the human team members feel that they can read the “minds” of their computer colleagues and feel comfortable working with them.
There are several possible ways to use what we know about describing intent. One is to prepare training so that people who regularly need to direct others can improve their skills. It may only take a few minutes at the beginning of a project to describe the intent clearly. The benefits of providing a clear description of intent—mistakes that are not committed, requests for help that are not made—are not easy to see.
Once we carefully watched an U. S. Army officer lay out his intent at the beginning of an exercise at the Army War College. We were impressed by his thoroughness, but we grew bored watching his team in action. They were not making interesting mistakes. Unlike the other teams, his team was not going off in wrong directions or getting into conflicts. Everyone knew what to do, and the assignment went off uneventfully. That is what seems to happen when intent is clear from the beginning.
To improve the ability to communicate intent, we cannot try to teach a checklist or set of procedures. A more valuable approach is to set up exercises to provide feedback to leaders about how well their intent is understood. Our training is based on the work of Lieutenant Colonel Larry Shattuck. We arrange for team leaders to describe their intent. Then the person running the exercise identifies an unexpected event that might occur. The leader writes down how he or she expects the subordinates to react, and at the same time the subordinates write down how they think they are supposed to react. Then we compare notes. Obviously, there are going to be mismatches and surprises, and we use those to figure out how the intent should have been phrased to provide guidance that would be more robust. We have applied this training with U.S. Marine Corps squad leaders, who felt that it was very helpful in preparing for field exercises.