(FORTUNE Magazine) – IF EVERYBODY THINKS business needs better leadership--and apparently everybody does--then why is the corporate world's understanding of how to teach leadership still Stone Age primitive? Says MIT's Peter Senge, author of the management classic The Fifth Discipline: "We know how to invest in technology and machinery, but we're at a loss when it comes to investing in human capital."

Part of the problem is that as markets evolve, the definition of leadership mutates almost as fast as a 12-year-old's fashion sense. Edward Lawler of the leadership program at the University of Southern California's business school recalls concluding many years ago that IBM was tops at developing executive talent. Wrong. "IBM probably invested the most money of any organization in this," Lawler says now, "but they taught people about a world that doesn't exist anymore. They shrank their gene pool down to people who were very good at managing for the 1970s--so when the 1990s arrived, IBM had lots of people who were very good at the wrong thing."

Today's standard of leadership--influencing human behavior in an environment of uncertainty--is dauntingly difficult to teach. Ronald Heifetz, a professor at Harvard's Kennedy school of government, argues that instead of telling people what to do, real leaders focus on helping people find their own way through "adaptive challenges"--problems without readily apparent solutions. Jim Collins, author of Built to Last, reports that companies that succeed long term stick passionately to a set of values and create systems that get employees to act in accord with those values. Says he: "Companies that take an architectural approach, putting in mechanisms to produce the right kind of behavior, don't need to look outside for leaders."

Indeed, a small group of great companies, esteemed for producing terrific leaders, may have the answer to this management conundrum. Hewlett-Packard, Fuji Xerox of Japan, General Electric, McKinsey, and PepsiCo have been experimenting with an idea powerful enough to transform leadership development--call it making the soft stuff hard. All are building quantifiable processes that command the attention of employees and bring discipline to the mysterious art of human development.

GE's Jack Welch, for instance, built his reputation by melding the seemingly incompatible hard and soft sides of management. GE evaluates people on what it calls boundarylessness to weed out those who obstruct the free flow of ideas--a very unleaderlike thing to do. One unit of Hewlett-Packard has a checklist of 26 leadership characteristics against which to gauge its employees. McKinsey has an elaborate mentoring program that makes sure the best people end up running this $1.5-billion-a-year consulting firm.

Efforts to reduce subjective observations to hard numbers positively invite skepticism. And it's true that procedures like these rarely produce useful results at first. But by remaining committed year after year and learning from mistakes, companies have found they can improve their soft processes until they become wellsprings of competitive advantage. Those that shy away at the first sign of doubt learn nothing.

Serious attempts to get more out of individual human beings necessarily involve the weird, unpredictable, sometimes yucky stuff that has come to be labeled the "soft" side of management. The business world, however, tends to attract and promote hardheaded high jumpers--male MBAs, military vets--who hate facing painful emotions, exploring their feminine side, or becoming more intuitive. These programs help bridge the gap.

"The soft stuff is always harder than the hard stuff," says Roger Enrico, vice chairman of PepsiCo. An engaging marketer with heavy shoulders and thick black hair, he is the former star chief of the company's Pepsi-Cola and Frito-Lay divisions and now heads its troubled $10.5-billion-a-year restaurant business. He continues: "Human interactions are a lot tougher to manage than numbers and P&Ls. So the trick is to make the soft stuff hard, to operationalize it." Notice Enrico's masterful use of a really ugly six-syllable word that conveys an aura of no-nonsense engineering to sell a bunch of warm and fuzzy ideas. This is how you get people to buy into the fruitcake stuff. (For more on Enrico's ideas, see the following article about the private leadership school he runs for PepsiCo.)

Teaching leadership requires defining it. That's no easy task in an era of so much change that, as the old Firesign Theatre comedy troupe once put it, "Everything you know is wrong." Desperate for answers, corporations and consultants are drawing up lists of so-called leadership competencies, from "thinking outside of the box" to "musical listening," which means hearing the emotional content behind someone's words. But how to avoid the IBM trap of indoctrinating people with irrelevant skills?

When you boil it all down, contemporary leadership seems to be a matter of aligning people toward common goals and empowering them to take the actions needed to reach them. Ultimately that means a leader must become worthy of respect. As you start to relinquish the command-and-control model of leadership--something GE chief Welch, PepsiCo CEO Wayne Calloway, and many other admirable leaders haven't completely done yet--your job is to get people to follow you voluntarily. You could do worse than heed the I Ching, a Chinese leadership guide used by Confucius: "Radical changes require adequate authority. A man must have inner strength as well as influential position. What he does must correspond with a higher truth...If a revolution is not founded on such inner truth, the results are bad, and it has no success. For in the end, men will support only those undertakings which they feel instinctively to be just."

At Hewlett-Packard, one model of this new leadership is Jean Kvasnica. She signed on as a secretary 15 years ago, rising since to head a multifunctional sales team serving a major customer. This fall, Kvasnica's group was preparing a complicated pitch to sell hundreds of millions of dollars' worth of computers, meters, and other equipment--a big deal even at a company with annual sales of $25 billion. Yet H-P has given Kvasnica few of the traditional trappings of authority: No one on the team reports to her, so she doesn't formally evaluate performance or give them raises.

How, then, does she motivate them? "They have to believe I know what I'm doing and won't waste their time," says Kvasnica. Says Javier Castelblanco, a team member from the H-P division that makes testing and measurement equipment: "Jean has vision and intense commitment to the successful outcome of a project, but the idea that makes it successful could come from anyone. She's not selfish about it." Forget managing by walking around, a technique that was invented years ago at H-P; this is management by getting out of the way.

One explanation of Kvasnica's effectiveness is that she understands what she herself responds to in others. "The kind of person I would follow, it's like there is a stick down through the center of them that's rooted in the ground," she says. "I can tell when someone has that. When they're not defensive, not egotistical. They're openminded, able to joke and laugh at themselves. They can take a volatile situation and stay focused. They bring out the best in me by making me want to handle myself in the same way. I want to be part of their world. When someone comes into the room with those attributes, it makes everyone in the room feel like we're all contributing."

BUT IF LEADERSHIP is character, it's reasonable to wonder whether it can be taught. Larry Bossidy, Allied-Signal's CEO, believes leadership is at least partly genetic. The consensus view is reflected by Morgan McCall, a professor of business at USC, who studies why high-potential careers sometimes derail. "I don't believe leadership can be taught," he says, "but it can be learned."

Peter Senge objects to the very notion of teaching leadership. "'Teaching' suggests that you have certain concepts you want people to understand, and that's pretty useless in a domain like leadership. Leadership has to do with how people are. You don't teach people a different way of being, you create conditions so they can discover where their natural leadership comes from." PepsiCo's Enrico cuts through the debate with a characteristically pragmatic point of view: "This training is not remedial," he says. "I have not tried to teach leadership to people who aren't already good at it. But it is a skill that can be honed and developed."

Creating the right environment is a symphonic process, involving such disciplines as selection, appraisal, job assignment, and mentoring. Step one is hiring the right people. Asks Jim Collins rhetorically: "How do you get people to share your values? You don't. You find people who share them and eject those who don't."

Just because someone shares your values doesn't mean he or she is leadership material. That's where leadership screening systems can come in. Companies like PepsiCo, McKinsey, and H-P have formal systems to identify tomorrow's leaders. Such systems can help companies avoid overlooking potential talent. With a request for anonymity and a sardonic glance toward the corner office, one person in a company that doesn't have a screening system says, "I'm changing all the time. I hope they realize that around here. I'm not sure they do." Employers who don't perceive the changes in their people tend to lose those people, who move to new employers partly to win recognition of the current state of their identity.

Even the very best screening systems can be infuriatingly imprecise. Paul Russell, director of executive development at PepsiCo, says Enrico set out a couple of years ago to personally train the 20 or 25 PepsiCo executives most likely to become presidents of the company's five operating divisions, or heads of functions such as finance. "Roger knew he was not smart enough to know who those people were, and neither did the division presidents or personnel people," Russell recalls. "So he figured if he trained 75 to 100 people, the 25 he wanted would probably be there."

Although they differ considerably in form, the screening systems at these companies all tend to be rigorous, ruthless in enforcing accountability, and deeply embedded in the culture. For instance, PepsiCo's annual human-resource planning process began more than 20 years ago when President Andrall Pearson, now a partner at the LBO firm of Clayton Dubilier & Rice, institutionalized the up-or-out method he had learned as a consultant at McKinsey. Among other things, he required PepsiCo executives to rank their direct reports into quartiles and describe, in face-to-face meetings, their plans for culling the worst performers.

INSISTENCE on accountability gives backbone to the softer stuff. All the model companies routinely fire people who don't meet their tough standards. Pepsi's Calloway says, "Occasionally it's very important to have a public hanging." And don't think that H-P is all about flower power: People who can't make their numbers or get in sync with the company's values sooner or later are gone.

Only about 20% of McKinsey's entry-level professionals become partners. They are evaluated and selected through a process that relies heavily on informal networking and on evaluations by colleagues. Explains Rajat Gupta, who reigns as first among equals in the McKinsey partnership: "People here are constantly observing each other." Special committees assigned to each of the firm's three tiers of consultants--associates, junior partners, and senior partners--collect anecdotal information, formalize it, and then individually evaluate each employee. "It's very subjective," says Gupta, "but I have a high level of confidence in the process. We may be a year late in promoting someone, but we for sure get it correct over time."

Once you've identified your future stars, it's smart to invest heavily in formal employee training and education--GE happily spends $500 million annually--but don't rely on course work to magically produce leaders. Ultimately the purpose of GE's Crotonville school is to transmit GE's values to employees. As Steve Kerr, the school's director, explains: "I wish I could tell you that courses are the key, but they are not. When we ask our people to write down the outstanding development experience of their lives, only about 10% cite formal training." The majority of peak learning experiences occur on the job--and through serendipity, not planning. But Kerr says you're more likely to get lucky if you give people a carefully thought out series of varied and challenging work assignments characterized bylots of responsibility and real risk of failure.

No matter how radically these model corporations delegate operating authority, they tend to maintain extremely tight central control over key personnel decisions. Both the head of HR and the CEO often directly approve moves within the ranks of their top echelon of executives. The goal is to identify and nurture human potential long before maturity makes it obvious.

"Among the elements of teaching leadership," says Calloway, "80% is experience. Our first line of offense is just to put them in the job." Along with many others, Calloway believes in diversity of experience, so he shuttles promising managers from division to division. CEO Bossidy of AlliedSignal agrees, but adds a caveat: The most fascinating assignment in the world may not teach you much unless your boss gives you a long leash.

GE devotes enormous energy to matching key managers with jobs in a process called Session C that consumes nearly a month of Jack Welch's time each year. Starting in January, 80,000 GEers and their bosses fill out the front and back of one-page "internal resume" forms, filling spaces for skills, career goals, and development needs. Between March and May, Welch and a few other senior executives visit each of GE's 12 operating units to conduct fiercely focused one-day personnel reviews. Meeting mainly with the top leadership at each site, they end up considering the prospects of about 500 GE senior managers. Pairing people with job assignments is part of the process.

In the old days, when companies viewed executives as two-legged bundles of skills, each assignment was expected to teach a new trick or two. Lawler of USC says today's fuzzier, more holistic conception of leadership suggests that what people need most may be career experiences that give them a better understanding of themselves. When that doesn't happen, says Jeffrey Sonnenfeld of the Emory University business school, practices intended to teach leadership teach followership instead.

The personal histories of most of the successful executives FORTUNE interviewed for this article are studded with consciousness-expanding assignments. Gupta got to run McKinsey's Scandinavian office when he was just 32. Jack Welch started out at a plastics skunkworks at GE, which he built into a $1 billion business. Roger Enrico learned to outwit bureaucracy as a Navy logistics officer managing an aviation fuel supply operation in Vietnam. Even H-P's director of education, a psychologist turned MBA named Claudia Davis, had to earn her stripes as a Switzerland-based marketing manager of a $900 million computer service business.

Everybody agrees that failure is a great teacher--but best met early in life. Says Calloway, who takes personal responsibility for Frito-Lay's $16 million flop in the cookie business: "The higher up you get, the more expensive those failures get, so the company will put up some fail-safe mechanisms. That's why we like to give people as much different experience as we can while they are young." As the Outward Bound schools have demonstrated so successfully, the perception of risk helps people learn. But they need support systems too: Uncertainty can be scary.

A PARTICULARLY helpful way to support future leaders is through mentoring. A close relationship with a senior executive of proven leadership skill is likely to keep a young manager open and stretching for growth. Mentoring is what the MBAs call it, but friendship--a soft value if there ever was one--ultimately may be the reason it works. H-P, which stands out from the group of model companies for its egalitarian, West Coast ethos, tries to nurture such relationships by assigning mentors to employees. Employees seem grateful, but mentoring doesn't always work. At a reunion lunch in Palo Alto recently, an H-P executive discussed in chilling terms a mentor who had no time for his protege: "We're on the verge of shutting that relationship down." However, passively trusting luck is no solution either: "One mentor is one more than most people get," says GE's Kerr.

McKinsey's approach may be the best balance of hard and soft. Its consultants work on teams of mixed rank, and senior people are expected to help junior people along. Part of the development process is quite formal: It is customary for young associates, not partners, to make presentations to clients, for instance, and mentoring is an important criterion in partners' appraisals. But long hours, hard work, and plenty of travel to places like West Moose Lung provide opportunities for teammates to forge personal relationships.

"It's like the old craft apprenticeship system," says Joel Bleeke, a senior partner whose clients are mostly in financial services. "The relationship begins as an act of will, not affection, but it becomes much more of an emotional attachment over time. When you are mentoring for leadership, you have to convey much more than problem-solving skills and your personal network--you need to convey aspirations, instill values, excitement, a view that almost anything is possible. You need to instill positive energy."

ONE OF Bleeke's apprentices is David Friedman, a junior partner with deep black circles under his eyes that suggest intense devotion to his work. Friedman says the relationship has made him readier to take risks and try new ideas: "If you know you have someone supporting you, you have much more confidence. You don't have to stay in your comfort zone. You can stay in your mentor's comfort zone, which is probably much bigger than your own."

Yotaro Kobayashi, chairman of Fuji Xerox and successor to Akio Morita as Japan's most prominent international industrialist, describes mentoring relationships that arise from a seemly mix of chance and intent. As a younger man, he was greatly influenced by time spent in what he calls "the presence of greatness." Among the leaders who have served as his personal benchmarks are the late Joe Wilson, grandson of a Xerox founder and the man who built the xerography business. "Only by meeting a person face to face--not once but several times--can you sense their qualities," says Kobayashi. "There are things you only feel. You ask yourself, 'How can I do this?' "

In the course of a long relationship, Wilson induced Kobayashi to take a seminar at the Aspen Institute, where the syllabus included Plato, Aristotle, and the Old Testament. Initially reluctant to mess with such apparently impractical stuff, Kobayashi felt his mind expanding as he read these tomes, and he became an enthusiastic student of religion, history, and culture. "Organizations are made of people," he explains. "As our activities become global, understanding other people in different parts of the world--all the way to the roots of their thinking--is very important. That means understanding the factors on which their values and sense of judgment are based."

Now 62, Kobayashi has become a mentor. He helped create a humanities workshop in Japan called Nidom (green forest), patterned after the Aspen Institute. He participates in a number of groups that meet occasionally for meals and private conversations about subjects as varied as Zen Buddhism and reengineering. And he spends informal time with promising younger executives--not all Fuji Xerox employees--on golf links and in karaoke bars. Although Kobayashi doesn't express it this way, the idea is to lift the veils that give leaders their aura of majesty and mystery; once young people can recognize their heroes as human, they can begin to find heroism in themselves.

Think Kobayashi's story is too Japanese to be relevant to you? Roger Enrico says one of the most important mentoring experiences in his life was a series of long discussions he had as a younger man with Don Kendall, then CEO of PepsiCo. For a couple of hours in the middle of the workday, Kendall would invite Enrico in, close the door, and stop answering the phone. The first time, Enrico showed up with a 300-page fact book and a flip-book of charts, but Kendall spent the meeting talking about opera. In subsequent chats they moved on to Soviet politics and a wide range of other subjects but never viewed Enrico's slides. Instead, Kendall gradually infused Enrico with the broad perspective of a leader.

So you've busted your chops identifying potential leaders, set them up in the right jobs, and provided them with mentors. Now comes the next big challenge: establishing a very clear set of companywide guidelines, a systematic way to measure whether someone is behaving like a leader. In competitive matters, GE's Welch can be brutally tough, yet he believes the way to win is by tapping the emotional energy of employees. Characteristically, he insists on evaluating GE executives according to how well they embody the corporate value of "boundarylessness," among other attributes. That value is about overcoming the barriers that inhibit a free market for ideas within GE--such as overbearing bosses, people unwilling to work with others outside their function, or timidity about speaking openly with customers and sup- pliers. GE makes its inevitably subjective judgments about people's boundarylessness on a rigid one-to-five scale--and sometimes fires those with the worst scores, even if they produce superior financial results. Spend too much time applying shoe leather to the necks of your subordinates, and you're out.

Concedes GE human resources senior VP William Conaty, one of the few HR chiefs with real impact and plenty of access to the CEO: "Ranking boundarylessness on a one-to-five scale is pretty tough. But values are the core ingredient here. The people we are putting in key leadership slots are those we deem to be terrific role models. That means embracing the values, being able to motivate and energize others, and having that infectious enthusiasm to tap people's potential and generate the capacity of the organization beyond what it otherwise would do."

BUT HOW DOES a company know for sure whether its best people are on the right path to leadership? They may think they are living up to a company's values, but are they really? That's where so-called 360-degree evaluations come in. This useful tool has spread well beyond the top-executive ranks where it initially took hold. The 360 process begins with HR devising a detailed questionnaire pegged to the behaviors the corporate culture values most. The questionnaire usually asks employees whether their manager "keeps me informed," "does not interfere with my job," and so forth. Everyone who works with the manager--boss, peers, and subordinates--contributes to the evaluation.

At GE, which uses 360s for more-senior people, the output is a bluntly worded report that the manager discusses behind closed doors with an HR pro and then with his boss. "Development needs"--the euphemism for problems--get plenty of attention.

A classic example of making the soft stuff hard, the 360-degree evaluation can provide great benefits. First, it provides a means of systematically making subjective yet apparently unbiased judgments about people. Is Harry a bully? Well, if six of his ten direct reports say he is, that starts to look like hard data. Second, the 360 process is designed to force supervisors into the sort of candid, face-to-face discussions that most supervisors would prefer to avoid.

Most important for leadership, feedback from 360s can signal opportunities to learn. Growth begins when individuals reach a more objective understanding of their strengths and weaknesses, enabling them to take responsibility for their own development. McCall, the professor who studied derailed careers, cites unwillingness to accept feedback as one of the reasons talented executives fail. That danger increases with rank, says Ellen Hart, head of Gemini Consulting's leadership practice: "The higher executives get in an organization, the less direct feedback they get about their behavior."

Be careful, though: 360s can be harmful if misused. In the wrong hands--say, a person with a grudge--they can become instruments of coercion or enforced conformity. And one HR person can corrupt the whole process, simply by revealing who made which anonymous comment about whom.

Put all these various methods into a pot and let them simmer, and eventually the result will be a corporate culture that encourages the right kind of leadership. Probably nothing more affects an employee's on-the-job development than the subtle messages that radiate from any company's culture and ambiance. Davis, H-P's education director, emphasizes the familiar but little-practiced idea that "you've got to walk the talk. If there is ambiguity about your message or values, people will opt out."

H-P is an example of the potentially overwhelming power of corporate culture. From its founding in 1939 by William Hewlett and David Packard, it has been ruled by a system of values that have come to be known, rather mystically, as the H-P Way. The basic premise, in Hewlett's words, is "the belief that men and women want to do a good job, a creative job, and that if they are provided the proper environment, they will do so." Values don't get much softer than that, but these seem to work: Since 1961, when H-P's stock was publicly listed, it has risen 7,885%.

More fascinating than H-P's famous egalitarianism is the company's systematic synthesis of hard and soft management ideas. H-P began as a provider of testing and measuring equipment for engineers, and the urge to measure and test is deeply ingrained in the culture. So when Davis and her team designed a course on coaching, she built in metrics from the ground up. First, she did research to define a list of 26 behaviors that the course should teach. That list became the questionnaire used for an upward evaluation given to each manager before taking the course. About five months after completing the course, the same evaluation is performed again for each graduate, and the difference in the scores is tabulated. On average, says Davis, scores improved 35%.

Perhaps the simplest, hardest, and most telling measure of any company's leadership development program is the allocation of its CEO's time. Calloway and Gupta say they spend roughly half their professional lives on personnel matters such as appraisals, assigning jobs, coaching, and mentoring. If, as everyone says, people are a company's most precious competitive resource, bosses at all levels have a personal obligation to be role models, visibly investing their own time in tomorrow's leaders.

The companies that are most successful at developing leaders seem to be those that are most successful in general. Does this mean that investing in leadership development is a luxury that only the elite can afford? After all, how can the head of a small, struggling enterprise be expected to fire high-performing managers who happen to be bullies, or to spend afternoons in meandering conversations about opera? What if you're battling for survival? Here's Wayne Calloway's answer: "I'll bet most of the companies that are in life-or-death battles got into that kind of trouble because they didn't pay enough attention to developing their leaders."

Reporter Associate Ani Hadjian