Mary T. Barra, the chief executive of General Motors, faced tough questions during a Senate hearing last week. Credit Doug Mills/The New York Times

Taken together, the documents sketch a picture of a company delving repeatedly into the technical intricacies that could cause the ignition switch to, if jostled, shift into “accessory” position, disabling the air bags. During this time, regulators complained that G.M. was uncooperative, while the company was often guided by cost concerns.

Besides the House investigation, a Senate panel is conducting an inquiry, as is the chief governmental auto safety regulator, the National Highway Traffic Safety Administration. The Justice Department has also begun a preliminary criminal inquiry into the company’s actions.

The email to Ms. Barra focused on a different issue from the ignition switch — a power-steering problem that led to an earlier recall, in 2010, for the Cobalt and the Pontiac G5. But it suggests that she was at least partly involved in a continuing conversation about problems with the small cars.

The email was sent to Ms. Barra by Terry Woychowski, a senior G.M. engineer at the time. A top G.M. safety lawyer, William Kemp, was copied on the memo.

In the email, Mr. Woychowski was apparently alerting Ms. Barra, who was senior vice president for global product development at the time, that the Saturn Ion might have the same power-steering problems that the Cobalt and G5 experienced.

“Mary,” the email said, “During the initial Cobalt case, the Ion data did not justify being included. The situation has been evolving. We will meet and understand the latest data.”

There is no document reflecting a follow-up meeting.

After the documents were released on Friday, General Motors sharply rebuked efforts to tie the 2011 email to any knowledge of the ignition switch issue.

House G.M. Hearing Documents

Reference documents compiled as part of an investigation by the House Energy and Commerce Committee into why it took General Motors more than a decade to publicly disclose a dangerous flaw in a small-car ignition switch and to order a recall of the vehicles with the defective switch.

Email From Safety Regulator to General Motors

Email dated July 23, 2013, from Frank Borris of the National Highway Traffic Safety Administration…

G.M. Internal Email Chain From September 2005

Emails dated Sept. 30, 2005, related to problems with the ignition switch in Cobalts and other…

This email, from Oct. 3, 2011, was related to power-steering problems in Saturn Ions.

An email dated June 14, 2005A, from a senior project engineer at Delphi to colleagues about the..

4 documents

“The email to Mary Barra dated Oct. 3, 2011, references a Saturn Ion steering issue — an issue completely separate from the ignition-related recalls,” the company said in a statement. “The email in no way contradicts Ms. Barra’s previous statements or testimony before the House or Senate subcommittees. The email was among the thousands of documents G.M. willingly provided to the Energy and Commerce Committee of the U.S. House of Representatives upon the committee’s request.”

But some lawmakers were not appeased.

“These documents continue to raise very serious questions — which G.M. and C.E.O. Mary Barra must address immediately — about whether she knew more, and earlier, about disabling defects in G.M. cars than she has acknowledged,” said Senator Richard Blumenthal, a Connecticut Democrat who questioned Ms. Barra aggressively in a Senate hearing last week.

Representative Diana DeGette of Colorado, the ranking Democrat on the subcommittee that questioned Ms. Barra, said, “I continue to have questions about why G.M.’s problems remained hidden from the public for so long.”

Much of the discussion about G.M.’s internal deliberations has focused on midlevel employees, including two engineering supervisors, Raymond DeGiorgio and Gary Altman, whom the company suspended with pay this week as its internal investigation continued.

Some of the documents may also add to discussion about federal regulators’ failure to investigate problems with Cobalts and Ions that were since found to be connected to the switch problem. An email on Sept. 5, 2007, from Gregory E. Magno, the chief of the safety agency’s defects assessment division, recommended opening a formal investigation into the failure of air bags to deploy in frontal crashes in Cobalts and Ions.

The agency ultimately declined to open the inquiry. In the note, Mr. Magno alluded to a reluctance at G.M. to acknowledge a problem.

“Notwithstanding G.M.’s indications that they see no specific problem pattern,” he wrote, the defects division saw a pattern of the air bags’ not deploying. He added that if the air bags had deployed, they would have “reduced injury levels and saved lives.”

In an email to a G.M. employee in charge of product investigations, Frank S. Borris II, director of the safety agency’s office of defects investigations, complained about the difficulties of working with the company.

“The general perception is that G.M. is slow to communicate, slow to act and, at times, requires additional effort of O.D.I. that we do not feel is necessary with some of your peers,” he said, referring to the Office of Defects Investigation.

One of the documents released on Friday shows that G.M. began an internal inquiry after an incident, dated Oct. 29, 2004, that was discovered by Mr. Altman.

In the report, the incident is described as “while driving the vehicle the driver’s knees bumped the key in such a manner as to turn off the ignition.”

In correspondence a few months later, in February 2005, Blendi Sullaj, an engineer who was working for G.M., described the problem, but said, “After talking to Ray DeGiorgio, I found out that it is close to impossible to modify the present ignition switch.”

He added, “The switch itself is very fragile and doing any further changes will lead to mechanical and/or electrical problems.”

Mr. DeGiorgio has been identified in documents as the engineer who approved a change to improve the switch in 2006. But the change was never cataloged as a new part, and G.M. did not recall cars with the original switch.

A month later, in March 2005, the inquiry was closed, partly because of cost issues, with this explanation: “The lead time for all the solutions is too long. The tooling cost and piece price are too high. None of the solutions seems to fully countermeasure the possibility of the key being turned (ignition turn off) during driving. Thus none of the solutions represents an acceptable business case.”

But a few months later, on June 14, 2005, John Coniff, a senior project engineer at Delphi Body, Security and Mechatronics, which supplied the switch to G.M., asked colleagues who could run tests on ignition switches.

“Ray is requesting this information,” said Mr. Coniff, possibly referring to Mr. DeGiorgio. “Cobalt is blowing up in their face in regards to turning the car off with the driver’s knee.”

Reporting was contributed by Matthew L. Wald, Danielle Ivory, Rebecca R. Ruiz, Hiroko Tabuchi, Christopher Jensen and Rachel Abrams.

A version of this article appears in print on April 12, 2014, on page B1 of the New York edition with the headline: G.M. Documents Reveal Years of Talks on Defect.