(FORTUNE Magazine) – "Management"--the word--traces its origins to the arts of horsemanship--from the Latin manus, "hand," via the Italian maneggiare, "to train horses." But since most of us call down to the stable to have the coach and four readied even less often than we wear white tie, let's talk cars. When you're driving, you do several potentially contradictory things at once. You want to get to your destination quickly, and also safely. You must watch the road and also look behind and around you. You can do one--you can do the other--but you want to do both. These are dilemmas: a word from the Greek, "two assumptions or premises." Dilemmas are what your boss talks about when he says, "You're in charge, Fosdick, but make sure Susannah is on board."

Managing dilemmas is what you do.

However, you cannot manage what you cannot name. That's why, when God gave Adam dominion over the fish of the sea and the birds of the air and every living thing that moves upon the earth, He brought the creatures before Adam, and had him name them.

A couple of years ago, an opportunity to define the dilemmas of management presented itself when Canadian Imperial Bank of Commerce decided to build an Edenic leadership center for its managers and employees on a rolling 100-acre site 45 minutes north of its Toronto headquarters. When Al Flood, the CEO of the bank (North America's eighth largest, with 1995 assets of $132 billion), okayed the plans, he did what any good banker would: He told Michele Darling, head of human resources, and Hubert Saint-Onge, who was to run the center, to make sure that this asset performed.

Thus challenged, Darling and Saint-Onge responded with an extraordinary act of in-house journalism. To create a mission and a curriculum for the center, Saint-Onge interviewed CIBC's 27 most senior executives, one on one. He asked each of them five open-ended questions: What factors in the business environment most affect CIBC's leaders? What cultural changes does the bank need to be more successful? What organizational capabilities does it need to build? What are the most pressing management-development needs? What features should leadership-center programs include?

Those questions elicited a wealth of ideas. Though some were predictable--what company's leaders don't feel they should "be more customer-driven"?--together they helped Darling and Saint-Onge define their mission. But there was more: The pages upon pages of notes made for an unusual chance to learn how leaders think and what they think about.

Poring over the interviews, Saint-Onge began highlighting phrases like "we must do a...but also b," or "in going after x, we must not lose sight of y." When he was done, he found nine "core leadership dilemmas." They fit any business and any manager, though they may be felt most keenly at the top.

Broad-based leadership vs. high-visibility leaders. The bankers acutely felt the need for top executives to get out front, to rally the troops personally and even charismatically. With equal fervor they saw a requirement to foster leadership throughout the company. If top management dominates the airwaves, it will silence others; if it is too reticent, the others will fret and wither for lack of support. For example, experts say reengineering will fail if the CEO doesn't take a strong role, but add that its success depends on a broad cadre of leaders among middle managers, team members, and others.

Independence vs. interdependence. We want entrepreneurship, a sense of ownership, and P&L responsibility, but we don't want one division badmouthing another, hogging shared resources, or refusing to take advantage of what the company has to offer. I was gored by the horns of this dilemma once: My then boss told me and a couple of peers that we had authority to make ad-budget decisions; we should consult with the marketing director but feel free to overrule her. When a few months later I felt it was necessary to do precisely that, he rapped my knuckles and said I should be more of a team player.

Long term vs. short term. The oldest dilemma in the book, but that doesn't make it easy to resolve. Several years back, I interviewed David Kearns, then CEO of Xerox. The topic: lessons learned from Xerox's failure to make successful businesses out of some of its inventions, like the desktop fax machine and the personal computer. Kearns had been wrestling with the problem of how to link labs more closely to the commercial needs of the business--while not gutting their talent for imagining and inventing a faraway future. As he showed me out at the end of our talk, he took a deep breath, shook his head, and settled his shoulders. He said: "We've been talking about five and ten years ago, and five and ten years from now. Waiting outside are a couple of security analysts, and the only thing they care about is the next quarter." What struck me most was the almost physical effort he needed to rearrange his mind.

Creativity vs. discipline. You encourage all those entrepreneurial leaders to benefit from freethinking, but they still have to make budget and adhere to company policy. It's as if Mao Zedong had said, "Let a thousand flowers bloom--in a topiary garden."

Trust vs. change. At first blush, trust and change don't seem to pull in opposite directions, but they do: Implementing organizational change--whether it's moving offices around or massive reengineering--can damage trust and commitment. Old work groups are sundered, new bosses have new standards, the world's in flux, and it's look out for No. 1. Even positive change can weaken trust: It was those wedding bells, after all, that broke up that old gang of mine. Yet without trust, change is impossible.

Bureaucracy busting vs. economies of scale. Let's centralize purchasing to leverage our size to get a better price. Let's also destroy the costly bureaucracies that clog the corporate coronary arteries when they, for example, force me to buy stuff that doesn't fit the needs of my business unit. (If you're not noticing that these dilemmas fall into a pattern, you're not reading closely enough.)

People vs. productivity. Another chestnut, but one that's really in the fire these days: The need to maximize productivity, to get everybody contributing 110%, must be balanced against the demands of personal life and the realization that, in the long run, all work and no play makes Jack a sitting duck.

Leadership vs. capability. The managerial and technical skills that enhance operations are quite different from the people-and-vision skills that produce leadership. The best strategy in the world won't work if it is poorly executed, but superb implementation of the wrong strategy simply means that Armageddon will come sooner than it otherwise would.

Revenue growth vs. cost containment. Once, in a hotel lobby, I saw a man carrying a coffee mug on which was printed BUDGETS ARE FOR WIMPS. "Where'd you get that?" I asked, hoping he would tell me a nearby shop had them.

But no: "My boss had them made for us." "He a marketing guy?" I asked.

He said: "How did you know?"

NINE DILEMMAS that describe your job. What do you do with them? First, notice that pattern. Says Saint-Onge: "These are all different, but they form a single, central dilemma." Its name: empowerment vs. alignment, the never-ending balancing act of managerial Bongo Board in which you try to give people independence and authority while making sure they use it in a way you'd approve of if they asked, which you don't want them to do except, of course, when you do want them to. Lee Iacocca sent the wrong message in that TV ad. The right one is: "Lead, follow, and get out of the way."

Second, you can make charts--always a good use of managerial time. Take each dilemma, and put one horn on a vertical axis and one on the horizontal. Draw a 45 [degree] diagonal to represent a balance between the two. Then on a scale of 1 to 10, 10 being best, locate your outfit (or yourself or your boss) on the grid. This chart shows a company that pursues revenue growth aggressively but doesn't mind costs well.

Set these up with all the "empowerment" tendencies (broad-based leadership, independence, long-termism, creativity, etc.) on the vertical axis and the "alignment" group on the horizontal. What do you see? Are you usually below the diagonal (too controlling) or above it (too loosey-goosey)?

You want to be spot on the diagonal line, and as far out toward the upper-right-hand corner as possible. This is because both sides of each dilemma are good: You want creativity and you want discipline; in fact, to get the greatest benefit of creativity, you need to temper it with discipline, and vice versa. The goal is to manage better in both directions--you want maximum empowerment and maximum alignment, just as a figure skater wants perfect scores for both artistic impression and technical merit.

For the folks at CIBC, the most important lesson of the nine dilemmas was seeing that, fundamentally, leadership is about ambiguities, not certainties. Says Michele Darling: "The dilemmas helped us come to a different understanding of the roles a leader plays."

One role she calls "polarity management." Leaders often are mesmerized by the virtues of one side of a dilemma, and ignore its worthy alternative. "Successful leaders," says Darling, "explore both ends." If you get your jollies from bureaucracy busting, force yourself to love economies of scale: Who better to reap its benefits than an honest skeptic? Polarities can also help you diagnose and deal with a group's resistance to change--chances are they're hung up on one horn of a dilemma, and you can help by showing them that they are right, but only half right.

Twinned with polarity management is ambiguity management. Too much thinking about leadership has a hortatory "set a vision and march on toward it" feel. But big strategic facts aren't always so clear--something that both leaders and their followers have to understand. Embracing ambiguities can be a powerful way to learn about a changing world.

A third role emerges from polarities and ambiguities: Making meaning. Says Darling: "The defining role of a leader is to sort out a message" from these mixed signals and cross-purposes. The new customer-satisfaction scores have just come in, and they show problems, but you also just received a market-segmentation analysis that shows that some customers demand so much service they are actually unprofitable. Your job is to take those dilemmas and make sense--and sensible plans--out of them. Otherwise it's all Greek. REPORTER ASSOCIATE Sheree R. Curry Join Tom Stewart in the Fortune Forum on Compuserve (GO FFORUM) or by E-mail: