After hauling General Motors CEO Mary Barra to Capitol Hill, several members of the House and Senate told her they were outraged and mystified. How could yet another car company respond to a safety defect with obfuscation and delay instead of action and integrity?

Rep. Fred Upton (R-Mich.) was prompted to reminisce. “In this very room, I led the oversight hearings that examined the Ford-Firestone recalls,” Upton told Barra. “Yet, here we are investigating another safety failure. It is déjà vu all over again.”

Sen. Jay RockefellerJay RockefellerOvernight Tech: Trump nominates Dem to FCC | Facebook pulls suspected baseball gunman's pages | Uber board member resigns after sexist comment Trump nominates former FCC Dem for another term Obama to preserve torture report in presidential papers MORE (D-W.Va.) had much the same reaction. Four years after his hearings on Toyota’s botched response to unintended acceleration problems, Rockefeller was ready to dust off his remarks to Toyota and read them to Barra. “I can literally reread my statement from that hearing word for word,” Rockefeller said. “Once again, it seems an auto company that should be focused on building the safest fleet of vehicles disregarded a serious safety risk.”

How could GM not learn anything from the past failures of Toyota, Ford and, for that matter, itself? How could GM be incapable of making safe ignition switches, and if it was incapable, why didn’t GM just come clean about it? How could GM’s leaders be so reckless morally, or even, fiscally? Why didn’t GM do better?

Thus far, answers have eluded Upton, Rockefeller and their colleagues, as well as regulators and the general public. In fact, the matter is so baffling that Barra herself is flummoxed by the very same questions. “Sitting here today,” she told a House subcommittee, “I cannot tell you why it took years for a safety defect to be announced.” She did say she intended to find out.

The answer is not that GM succumbed to conspiracy, greed, or even incompetence. What GM fell victim to is the power of the problem itself. Like a hostage taker dictating its terms on a hapless victim, the ignition problem imposed itself on GM until there were literally no acceptable responses.

A recently released GM memo from 2005 tells the story: “The lead time for all solutions is too long. The tooling cost and piece price are too high. None of the solutions seems to fully countermeasure the possibility of…ignition turn off during driving. Thus none of the solutions represents an acceptable business case.”

It will take too long. It will cost too much. And whatever we do won’t solve the problem anyway. The problem closed in on GM’s decision makers until every direction offered only a different path to failure. This is why GM’s response to an intolerable problem was to do nothing at all.

Upton is having déjà vu all over again because companies and people routinely let problems define what’s possible. And when we do that, we fail.

More than two decades ago, researchers at Texas A&M documented the failure-inducing power of problems. Tell an engineer what’s most likely to stump him on a project and he becomes more likely to fail than if you had never mentioned the hard part. Knowing the problem means thinking problem-first, with every potential design measured on the problem’s terms. In the process, the problem stifles creativity and lures the designer into unsolvable traps. In one variation of the experiment, engineers who were told the problem they would encounter designing a spill-proof coffee mug were 17 times more likely to fail than peers who went about the same task without any advance warning. In other words, they had no problem solving the problem as long as they didn’t know what it was.

At the very same time GM was stumped by the question of how to keep its keys from accidentally turning off its cars, several automakers -- including GM itself! -- were developing keyless ignition systems in which a car can be started at the press of a button with the keyfob still in the driver’s pocket. Real innovation is possible when you are trying to come up with the best approach instead of getting yourself stuck inside the limits of a problem.

For automakers and their CEOs, there is one way to avoid being summoned to Capitol Hill to listen to the next round of déjà vu condemnations. When a safety issue comes up, don’t let the problem run the company.

Niven, Ph.D., teaches at the University of Cincinnati. His book on the problem with problems -- It’s Not About the Shark (St. Martin’s Press) -- will be published in November.