In “Why the Arpanet Was Built” Stephen Lukasik asserts that the goal of Arpanet was to, “exploit new computer technologies to meet the needs of military command and control against nuclear threats, achieve survivable control of US nuclear forces, and improve military tactical and management decision making.”  Lukasik recognizes that even though the capabilities of a packet-switching network had non-defense applications those needs were not central to Arpanet’s decision to pursue networking.  A pivotal point in the direction of the ARPA program was in October 1962 when Jack Ruina hired J.C.R. Licklider to run the program.  Previously ARPA had been carefully watched by the Secretary of Defense, the White House, and the President’s Science Advisory Committee, but because of a shift in focus in behavioral science “beyond the narrow focus in the department beyond human factors” Licklider was left to his endeavors.  Licklider has his own vision of what the command and control problem ARPA was created to solve.  He saw it as having two parts, “the machine processing of information and the presenting of that information to humans in a form suitable for use in making decisions.”  It was this vision and room to develop that lead to the creation of a decentralized network that was the “technical solution to avoiding decapitation.”  This would allow for a network that could survive an attack.  Licklider anticipated that the general utility of networking would go far beyond the needs of the Department of Defense, but it was originally conceived as Arpanet for the purposes of command and control during the Cold War.

http://muse.jhu.edu/journals/ieee_annals_of_the_history_of_computing/summary/v033/33.3.lukasik.html